CONTENTS

DedicationIntroductionChapter1:FromCommunisttoVentureCapitalistChapter2:“IWillSurvive”Chapter3:ThisTimewithFeelingChapter4:WhenThingsFallApart

TheStruggleCEOsShouldTellItLikeItIsTheRightWaytoLayPeopleOffPreparingtoFireanExecutiveDemotingaLoyalFriendLiesThatLosersTellLeadBulletsNobodyCares

Chapter5:TakeCareofthePeople,theProducts,andtheProfits—inThatOrderAGoodPlacetoWorkWhyStartupsShouldTrainTheirPeopleIsItOkaytoHirePeoplefromYourFriend’sCompany?WhyIt’sHardtoBringBigCompanyExecsintoLittleCompaniesHiring Executives: If You’ve Never Done the Job, How Do You HireSomebodyGood?WhenEmployeesMisinterpretManagersManagementDebtManagementQualityAssurance

Chapter6:ConcerningtheGoingConcernHowtoMinimizePoliticsinYourCompanyTheRightKindofAmbition

TitlesandPromotionsWhenSmartPeopleAreBadEmployeesOldPeopleOne-on-OneProgrammingYourCultureTakingtheMysteryOutofScalingaCompanyTheScaleAnticipationFallacy

Chapter7:HowtoLeadEvenWhenYouDon’tKnowWhereYouAreGoingTheMostDifficultCEOSkillTheFineLineBetweenFearandCourageOnesandTwosFollowtheLeaderPeacetimeCEO/WartimeCEOMakingYourselfaCEOHowtoEvaluateCEOs

Chapter8:FirstRuleofEntrepreneurship:ThereAreNoRulesSolvingtheAccountabilityvs.CreativityParadoxTheFreakyFridayManagementTechniqueStayingGreatShouldYouSellYourCompany?

Chapter9:TheEndoftheBeginningAppendix:QuestionsforHeadofEnterpriseSalesForceAcknowledgmentsAbouttheAuthorCreditsCopyrightAboutthePublisher

INTRODUCTION

“Thistherealworld,homie,schoolfinishedTheydonestoleyourdreams,youdunnowhodidit.”

—KANYEWEST,“GORGEOUS”

EverytimeIreadamanagementorself-helpbook,Ifindmyselfsaying,“That’sfine, but thatwasn’t really the hard thing about the situation.”The hard thingisn’t setting a big, hairy, audacious goal. The hard thing is laying people offwhenyoumissthebiggoal.Thehardthingisn’thiringgreatpeople.Thehardthing is when those “great people” develop a sense of entitlement and startdemandingunreasonablethings.Thehardthingisn’tsettingupanorganizationalchart.Thehard thing isgettingpeople tocommunicatewithin theorganizationthat you just designed. The hard thing isn’t dreaming big. The hard thing iswakingupinthemiddleofthenightinacoldsweatwhenthedreamturnsintoanightmare.The problem with these books is that they attempt to provide a recipe for

challenges that have no recipes. There’s no recipe for really complicated,dynamicsituations.There’snorecipeforbuildingahigh-techcompany;there’sno recipe for leading a group of people out of trouble; there’s no recipe formaking a series of hit songs; there’s no recipe for playing NFL quarterback;there’snorecipeforrunningforpresident;andthere’snorecipeformotivatingteamswhenyourbusinesshas turned tocrap.That’s thehard thingabouthardthings—thereisnoformulafordealingwiththem.Nonetheless,therearemanybitsofadviceandexperiencethatcanhelpwith

thehardthings.Idonotattempttopresentaformulainthisbook.Instead,Ipresentmystory

andthedifficultiesthatIhavefaced.Asanentrepreneur,aCEO,andnowasaventurecapitalist, I still find these lessonsuseful—especiallyas Iworkwithanewgenerationoffounder-CEOs.Buildingacompanyinevitablyleadstotough

times.I’vebeenthere;I’vedonethat.Circumstancesmaydiffer,butthedeeperpatternsandthelessonskeepresonating.Forthepastseveralyears,I’veencapsulatedtheselessonsinaseriesofblog

poststhathavebeenreadbymillionsofpeople.Manyofthosehavereachedoutto me wanting to know the backstory to the lessons. This book tells thatbackstoryfor thefirst timeandincludestherelatedlessonsfromtheblog.I’vealso been inspired by many friends, advisers, and family members who havehelpedme throughoutmy career and also by hip-hop/rapmusic.Because hip-hop artists aspire to be both great and successful and see themselves asentrepreneurs, many of the themes—competing, making money, beingmisunderstood—provideinsightintothehardthings.Isharemyexperiencesinthehopeofprovidingcluesandinspirationforotherswhofindthemselvesinthestruggletobuildsomethingoutofnothing.

—CHAPTER1—

FROMCOMMUNISTTOVENTURECAPITALIST

“ThishereisallaboutMywife,mykids,thelifethatIlive

Throughthenight,Iwashis,itwasright,butIdidMyups,anddowns,myslips,myfalls

Mytrialsandtribulations,myheart,myballs.”

—DMX,“WHOWEBE”

TheotherdayIthrewabigbarbecueatmyhouseandinvitedahundredofmyclosest friends. These types of gatherings aren’t unusual. My brother-in-law,Cartheu,andIhavebeenbarbecuingforyears,andmyskillshaveearnedmethenickname from my African American friends “the Jackie Robinson ofBarbecue.”Icrossedthecolorline.At this particular barbecue, the conversation turned to the great rapperNas.

MyfriendTristanWalker,ayoungAfricanAmericanentrepreneur,commentedproudlythatNaswasfromhishomeproject,Queensbridge,NewYork—oneofthelargestpublichousingprojectsintheUnitedStates.Myseventy-three-year-oldJewishfatherinterjected,“I’vebeentoQueensbridge.”Convincedthattherewasnoway thatmyold,white fatherhadbeen toQueensbridge,Tristan said,“You must mean Queens. Queensbridge is actually a housing project in anextremelyroughneighborhood.”Myfatherinsisted:No,itwasQueensbridge.I pointed out to Tristan that my father grew up in Queens, so he couldn’t

possibly be confused. Then I asked, “Dad, what were you doing inQueensbridge?”Hereplied,“IwaspassingoutcommunistliteraturewhenIwaselevenyearsold.Irememberitwell,becausemymothergotveryupsetthattheCommunistPartysentmeintotheprojects.Shethoughtitwastoodangerousforalittlekid.”

My grandparents were actually card-carrying Communists. As an activememberintheCommunistParty,mygrandfatherPhilHorowitzlosthisjobasaschoolteacher during theMcCarthy era.My fatherwas a red-diaper baby andgrew up indoctrinated in the philosophy of the left. In 1968, he moved ourfamilywest toBerkeley,California,andbecameeditorof thefamedNewLeftmagazineRamparts.Asaresult,Igrewupinthecityaffectionatelyknownbyitsinhabitantsasthe

People’s Republic of Berkeley. As a young child, I was incredibly shy andterrified of adults.Whenmymother droppedmeoff at nursery school for thefirst time, I began to cry. The teacher told my mother to just leave, whilereassuring her that crying was common among nursery school children. ButwhenElissaHorowitzreturnedthreehourslater,shefoundmesoakingwetandstill crying. The teacher explained that I hadn’t stopped, and nowmy clotheswere drenched as a result. I got kicked out of nursery school that day. Ifmymother hadn’t been themost patient person in the world, I might never havegone to school. When everybody around her recommended psychiatrictreatment,shewaspatient,willingtowaituntilIgotcomfortablewiththeworld,nomatterhowlongittook.When Iwas five years old,wemoved from a one-bedroom house onGlen

Avenue,whichhadbecomefartoosmallforasix-personfamily,toalargeroneonBonitaAvenue.Bonitawasmiddle-classBerkeley,whichmeanssomethingabitdifferentfromwhatonefindsinmostmiddle-classneighborhoods.Theblockwas a collection of hippies, crazy people, lower-class peopleworking hard tomoveup,andupper-classpeopletakingenoughdrugstomovedown.Oneday,oneofmyolderbrotherJonathan’sfriends,Roger(nothisrealname),wasoverat our house.Roger pointed to anAfricanAmerican kid down the blockwhowasridinginaredwagon.Rogerdaredme:“Godownthestreet,tellthatkidtogive you his wagon, and if he says anything, spit in his face and call him anigger.”Afewthingsrequireclarificationhere.First,wewereinBerkeley,sothatwas

not common language. In fact, I had never heard thewordnigger before anddidn’t knowwhat itmeant, though I guessed itwasn’t a compliment. Second,Roger wasn’t racist and he wasn’t raised in a bad home. His father was aBerkeleyprofessor andbothhisparentswere someof thenicestpeople in theworld, butwe later foundout thatRoger suffered from schizophrenia, andhisdarksidewantedtoseeafight.Roger’s commandputme in adifficult situation. Iwas terrifiedofRoger. I

thought that he would surely give me a severe beating if I didn’t follow hisinstructions.Ontheotherhand,Iwasterrifiedofaskingforthewagon.Hell,Iwasterrifiedofeverything.IwasmuchtooscaredofRogertostaywhereIwas,so I began walking down the block toward the other kid. The distance wasprobably thirty yards, but it felt like thirty miles.When I finally got there, Icouldbarelymove.Ididnotknowwhattosay,soIjustopenedmymouthandstartedtalking.“CanIrideinyourwagon?”iswhatcameout.JoelClarkJr.said,“Sure.”WhenI turned toseewhatRogerwoulddo,hewasgone.Apparently,his light sidehad takenoverandhe’dmovedon to somethingelse. Joel and Iwent on to play all day that day, and we’ve been best friends ever since.Eighteenyearslater,hewouldbethebestmanatmywedding.Untilnow,I’venevertoldthatstorytoanyone,butitshapedmylife.Ittaught

methatbeingscareddidn’tmeanIwasgutless.WhatIdidmatteredandwoulddeterminewhetherIwouldbeaherooracoward.Ihaveoftenthoughtbackonthatday,realizingthatifI’ddonewhatRogerhadtoldmetodo,Iwouldhavenevermetmybestfriend.Thatexperiencealsotaughtmenottojudgethingsbytheirsurfaces.Untilyoumaketheefforttogettoknowsomeoneorsomething,you don’t know anything. There are no shortcuts to knowledge, especiallyknowledge gained from personal experience. Following conventional wisdomandrelyingonshortcutscanbeworsethanknowingnothingatall.

TURNYOURSHITINOvertheyears,Iworkedhardtoavoidbeinginfluencedbyfirstimpressionsandblindlyadheringtoconvention.GrowingupinBerkeleyasanexcellentstudentinatownthatfrowneduponfootballasbeingtoomilitaristic,Iwasn’texpectedtojointheBerkeleyHighSchoolfootballteam,butthat’swhatIdid.Thiswasabigstepforme.Ihadnotplayedinanyofthepeeweefootballleagues,soitwasmyfirstexposuretothesport.Nonetheless,thoseearlierlessonsindealingwithfearhelpedmetremendously.Inhighschoolfootball,beingabletohandlefearis75percentofthegame.Iwill never forget the first teammeetingwith head coachChicoMendoza.

CoachMendozawasatougholdguywhohadplayedcollegefootballatTexasChristianUniversity,homeofthemightyHornedFrogs.CoachMendozabeganhisopeningspeech,“Someofyouguyswillcomeouthereandyoujustwon’tbeserious.You’llgethereandstartshootingtheshit,talkingshit,bullshittin’,notdoingshit,andjustwanttolookgoodinyourfootballshit.Ifyoudothat,thenyou know what? Turn your shit in.” He went on to elaborate on what was

unacceptable:“Comelatetopractice?Turnyourshitin.Don’twanttohit?Turnyourshitin.Walkonthegrass?Turnyourshitin.CallmeChico?Turnyourshitin.”Itwas themost intense,hilarious,poetic speech I’deverheard. I loved it. I

couldn’twaittogethomeandtellmymother.Shewashorrified,butIstilllovedit. In retrospect, itwasmy first lesson in leadership.Former secretaryof stateColin Powell says that leadership is the ability to get someone to follow youevenifonlyoutofcuriosity.IwascertainlycurioustoseewhatCoachMendozawouldsaynext.Iwastheonlykidonthefootballteamwhowasalsoonthehighestacademic

trackinmath,somyteammatesandIdidn’tseeeachotherinmanyclasses.Asaresult, I endedupmoving inmultiple social circles andhangingoutwithkidswith very different outlooks on the world. It amazed me how a diverseperspectiveutterlychangedthemeaningofeverysignificanteventintheworld.For instance, when Run-D.M.C.’s Hard Times album came out, with itsrelentless bass drum, it sent an earthquake through the football team, but noteven a ripple through my calculus class. Ronald Reagan’s Strategic DefenseInitiative was considered an outrage among young scientists due to itsquestionable technical foundation,but thoseaspectswentunnoticedat footballpractice.Lookingat theworldthroughsuchdifferentprismshelpedmeseparatefacts

from perception. This ability would serve me incredibly well later when Ibecamean entrepreneur andCEO. Inparticularlydire circumstanceswhen the“facts” seemed to dictate a certain outcome, I learned to look for alternativenarratives and explanations coming from radically different perspectives toinformmyoutlook.The simple existence of an alternate, plausible scenario isoftenallthat’sneededtokeephopealiveamongaworriedworkforce.

BLINDDATEIn the summer of 1986, I had finished my sophomore year of college atColumbiaUniversity,andIwasstayingwithmyfather,whowasnowlivinginLosAngeles. I hadbeen set upon a blinddate bymy friend andhigh schoolfootballteammateClaudeShaw.ClaudeandIgotreadyforthedoubledatewithhis girlfriend, Jackie Williams, and my date, Felicia Wiley, by preparing anelaboratedinner.Wemeticulouslyplannedandcookedalldayandhadtheentiremeal, including four perfectly presented T-bone steaks, ready at 7 p.m.—datetime.Buttherewerenodates.Anhourpassed,butwedidn’tgettooworkedup.

Jackiewasknownforhertardiness,sonoworries.Thentwohourspassed,andClaudecalled forastatuscheck. I listened inshockas I lookedover thenow-cold gourmetmeal thatwe’d prepared.Mydate, Felicia, had decided that shewas“tootired”toshowupforthedate.Wow.Howobnoxious!ItoldClaudetohandmethephone.Iintroducedmyself:“Hi,thisisBen,your

blinddate.”Felicia:“Iamsorry,butIamtiredanditislate.”Me:“Well,itislate,becauseyouarelate.”Felicia:“Iknow,butIamjusttootiredtocomeover.”AtthispointIdecidedtoappealtohersenseofempathy.Me:“Well,Iunderstandyourpredicament,butthetimetocommunicatethis

messagewouldhavebeenbeforewespentalldaycookingdinner.Atthispoint,anythingshortofgetting intoyourcaranddrivinghere immediatelywouldberudeandleaveapermanentlypoorimpression.”Ifshewas totallyself-centered(assheappeared tobe),mypleawouldhave

noeffect,andIwouldbebetteroffmissingthedate.Ontheotherhand, ifshedidn’twanttogooutlikethat,thentheremightbesomethingthere.Felicia:“Okay,I’llcomeover.”Ninetyminuteslatershearrivedwearingwhiteshortsandlookingasprettyas

can be. In all my focus and anticipation about the date, I had completelyforgotten about the fistfight I’d been in the day before. During a pickupbasketball game in the San Fernando Valley, a six-foot-two-inch, crew-cut-sporting,camouflage-pants-wearing,fraternity-boy-lookingplayerthrewtheballatmybrother. Jonathanwas amusician, had longhair, and probablyweighedabout 155 pounds at the time. On the other hand, I was used to football andfightingandwasreadyforaction.Ijudgedthesituationonmyfirstimpression,and I rushed the frat boy.A scuffle ensued. I landed some good punches butcaughtarighthookundermylefteye,leavingabitofamark.It’spossiblethatmytargetplayerwassimplymadaboutahardfoulratherthantryingtobullymybrother, but that’s the price of not taking the time to understand. Iwill neverknow.Whateverthecase,whenIopenedthedoortogreetourdates,Felicia’saward-

winning green eyes immediately fixed on the welt under my eye. Her firstimpression(toldtomeyearslater):“Thisguyisathug.Comingherewasabigmistake.”Fortunately, neither of us relied on our first impressions. We have been

happilymarriedfornearlytwenty-fiveyearsandhavethreewonderfulchildren.

SILICONVALLEYDuringonesummer incollege, Igota jobasanengineeratacompanycalledSiliconGraphics(SGI).Theexperienceblewmymind.Thecompanyinventedmodern computer graphics andpowered awhole newclass set of applicationsranging from themovieTerminator2 toamazing flight simulators.Everybodytherewas so smart. The things they built were so cool. I wanted towork forSiliconGraphicsfortherestofmylife.Aftergraduatingfromcollegeandgraduateschoolincomputerscience,Iwent

backtoworkforSGI.BeingtherewasadreamcometrueandIlovedit.AftermyfirstyearatSGI,Imetaformerheadofmarketingforthecompany,RoselieBuonauro, who had a new startup. Roselie had heard about me from herdaughter, who also worked with me at SGI. Roselie recruited me hard.Eventually,shegotmeandIwenttoworkforheratNetLabs.JoiningNetLabsturnedouttobeahorribledecisionforme.Thecompanywas

run by Andre Schwager, a former Hewlett-Packard executive, and moreimportant, Roselie’s husband. Andre and Roselie had been brought in by theventurecapitalistsas the“professionalmanagement team.”Unfortunately, theyunderstood very little about the products or the technology, and they sent thecompanyoffinonecrazydirectionafterthenext.ThiswasthefirsttimethatIstartedtounderstandtheimportanceoffoundersrunningtheircompanies.Tomakemattersmore complicated,my second daughter,Mariah, had been

diagnosedwithautism,whichmadeworkingatastartupaterribleburdenforourfamily,asIneededtospendmoretimeathome.Oneveryhotdaymy father cameover for avisit.Wecouldnot afford air-

conditioning, and all three children were crying as my father and I sat theresweatinginthe105-degreeheat.Myfatherturnedtomeandsaid,“Son,doyouknowwhat’scheap?”Since I had absolutely no idea what he was talking about, I replied, “No,

what?”“Flowers.Flowersarereallycheap.Butdoyouknowwhat’sexpensive?”he

asked.Again,Ireplied,“No,what?”Hesaid,“Divorce.”Somethingaboutthatjoke,whichwasnotreallyajoke,mademerealizethatI

had run out of time. Up until that point, I had not really made any seriouschoices.IfeltlikeIhadunlimitedbandwidthandcoulddoeverythinginlifethat

I wanted to do simultaneously. But his joke made it suddenly clear that bycontinuingonthecourseIwason,Imightlosemyfamily.Bydoingeverything,Iwouldfailatthemostimportantthing.ItwasthefirsttimethatIforcedmyselfto lookat theworld throughpriorities thatwerenotpurelymyown. I thoughtthat I could pursue my career, all my interests, and build my family. Moreimportant,Ialwaysthoughtaboutmyselffirst.Whenyouarepartofafamilyorpartofagroup,thatkindofthinkingcangetyouintotrouble,andIwasindeeptrouble. Inmymind, Iwasconfident that Iwasagoodpersonandnotselfish,butmyactionssaidotherwise. Ihad tostopbeingaboyandbecomeaman. Ihadtoputfirstthingsfirst.IhadtoconsiderthepeoplewhoIcaredaboutmostbeforeconsideringmyself.Idecided toquitNetLabs thenextday. I founda jobatLotusDevelopment

thatwould allowme to getmy home life straightened out. I stopped thinkingaboutmyself and focusedonwhatwasbest formy family. I startedbeing thepersonthatIwantedtobe.

NETSCAPEOne day while working at Lotus, one of my coworkers showed me a newproductcalledMosaic,whichwasdevelopedbysomestudentsattheUniversityof Illinois. Mosaic was essentially a graphical interface to the Internet—atechnology formerly only used by scientists and researchers. It amazedme. Itwassoobviouslythefuture,andIwassoobviouslywastingmytimeworkingonanythingbuttheInternet.Severalmonthslater,IreadaboutacompanycalledNetscape,whichhadbeen

cofoundedby formerSiliconGraphics founder JimClarkandMosaic inventorMarcAndreessen. I instantlydecided that I should interview for a job there. Icalled a friend who worked at Netscape and asked if he could get me aninterviewwiththecompany.HeobligedandIwasonmyway.Duringthefirstinterviews,Imeteveryoneontheproductmanagementteam.I

thought themeetingswentwell, butwhen I arrivedhome that eveningFeliciawas in tears. TheNetscape recruiter had calledme to giveme some tips, andFeliciahadanswered.(Thiswasbeforethedaysofpervasivecellphones.)TherecruiterinformedherthatitwouldbeunlikelyI’dgetthejob,becausethegroupwas looking forcandidateswithStanfordorHarvardMBAs.Felicia suggestedthat maybe I could go back to school. Given that we had three children, sheknew thiswas unrealistic, hence the tears. I explained that recruiterswere nothiring managers, and that they might consider me despite my lack of proper

businessschooling.Thenextdaythehiringmanagercalledbacktoletmeknowthattheywanted

metointerviewwithcofounderandChiefTechnicalOfficerMarcAndreessen.Hewastwenty-twoyearsoldatthetime.In retrospect, it’s easy to thinkboth theWebbrowser and the Internetwere

inevitable,butwithoutMarc’swork,itislikelythatwewouldbelivinginaverydifferentworld.AtthetimemostpeoplebelievedonlyscientistsandresearcherswouldusetheInternet.TheInternetwasthoughttobetooarcane,insecure,andslow to meet real business needs. Even after the introduction of Mosaic, theworld’s firstbrowser, almostnobody thought the Internetwouldbe significantbeyond the scientific community—least of all the most important technologyindustry leaders, who were busy building proprietary alternatives. Theoverwhelming favorites to dominate the race to become the so-calledInformation Superhighway were competing proprietary technologies fromindustrypowerhouses suchasOracleandMicrosoft.Their storiescaptured theimagination of the business press. This was not so illogical, since mostcompaniesdidn’tevenrunTCP/IP(thesoftwarefoundationfor theInternet)—they ran proprietary networking protocols such as AppleTalk, NetBIOS, andSNA. As late as November 1995, Bill Gates wrote a book titled The RoadAhead, in which he predicted that the Information Superhighway—a networkconnectingallbusinessesandconsumersinaworldoffrictionlesscommerce—wouldbe the logicalsuccessor to theInternetandwouldrule thefuture.Gateslaterwentbackandchangedreferences fromthe InformationSuperhighway totheInternet,butthatwasnothisoriginalvision.Theimplicationsof thisproprietaryvisionwerenotgoodforbusinessorfor

consumers. In the minds of visionaries like Bill Gates and Larry Ellison, thecorporations that owned the Information Superhighway would tax everytransaction by charging a “vigorish,” as Microsoft’s then–chief technologyofficer,NathanMyhrvold,referredtoit.It’s difficult to overstate the momentum that the proprietary Information

Superhighwaycarried.AfterMosaic,evenMarcandhiscofounder, JimClark,originally planned a business for video distribution to run on top of theproprietaryInformationSuperhighway,nottheInternet.Itwasn’tuntildeepintotheplanningprocessthattheydecidedthatbyimprovingthebrowsertomakeitsecure, more functional, and easier to use, they could make the Internet thenetworkofthefuture.AndthatbecamethemissionofNetscape—amissionthattheywouldgloriouslyaccomplish.

InterviewingwithMarcwas like no other job interview I’d ever had.Gonewerequestions aboutmy résumé,mycareer progression, andmyworkhabits.Hereplacedthemwithadizzyinginquiryintothehistoryofemail,collaborationsoftware,andwhatthefuturemighthold.Iwasanexpert inthetopic,becauseI’dspentthelastseveralyearsworkingontheleadingproductsinthecategory,but I was shocked by how much a twenty-two-year-old kid knew about thehistoryofthecomputerbusiness.I’dmetmanyreallysmartyoungpeopleinmycareer, but never a young technology historian. Marc’s intellect and instinctstook me aback, but beyond Marc’s historical knowledge, his insights abouttechnologiessuchasreplicationwereincisiveandonpoint.Aftertheinterview,IphonedmybrotherandtoldhimthatI’djustinterviewedwithMarcAndreessen,andIthoughtthathemightbethesmartestpersonI’devermet.Aweeklater, Igot the job.Iwas thrilled.Ididn’treallycarewhat theoffer

was.IknewthatMarcandNetscapewouldchangetheworld,andIwantedtobepartofit.Icouldnotwaittogetstarted.OnceatNetscape,IwasputinchargeoftheirEnterpriseWebServerproduct

line.Thelineconsistedoftwoproducts:theregularWebserver,whichlistedfor$1,200,andthesecureWebserver(aWebserverthatincludedthethenbrand-newsecurityprotocolinventedbyNetscapecalledSSL,SecureSocketsLayer)for$5,000.AtthetimethatIjoined,wehadtwoengineersworkingontheWebservers: RobMcCool,who had invented theNCSAWeb server, and his twinbrother,MikeMcCool.By the timeNetscapewent public inAugust 1995,we had grown theWeb

server teamtoaboutnineengineers.TheNetscapeinitialpublicoffering(IPO)wasbothspectacularandhistoric.Thestockinitiallypricedat$14pershare,butalast-minutedecisiondoubledtheinitialofferingto$28pershare.Itspikedto$75—nearlyarecordforafirst-daygain—andclosedat$58,givingNetscapeamarketvalueofnearly$3billiononthedayoftheIPO.Morethanthat,theIPOwasanearthquake in thebusinessworld.Asmy friendand investmentbankerFrankQuattronesaidatthetime,“Noonewantedtotelltheirgrandchildrenthattheymissedoutonthisone.”Thedealchangedeverything.Microsofthadbeeninbusinessformorethana

decadebeforeitsIPO;we’dbeenaliveforsixteenmonths.Companiesbegantogetdefinedas“neweconomy”oras“oldeconomy.”Andtheneweconomywaswinning.TheNewYorkTimescalledtheNetscapeIPO“world-shaking.”But there was a crack in our armor:Microsoft announced that it would be

bundling its browser, Internet Explorer, with its upcoming breakthrough

operatingsystemrelease,Windows95—forfree.ThisposedahugeproblemtoNetscape, because nearly all of our revenue came from browser sales, andMicrosoftcontrolledmorethan90percentofoperatingsystems.Ouranswertoinvestors:WewouldmakeourmoneyonWebservers.Two months later, we got our hands on an early release of Microsoft’s

upcomingWeb server Internet InformationServer (IIS).Wedeconstructed IISand found that it had every feature thatwehad—including the security inourhigh-endproduct—andwasfivetimesfaster.Uh-oh.IfiguredthatwehadaboutfivemonthsbeforeMicrosoftreleasedIIStosolvetheproblemorelsewewouldbetoast.Inthe“oldeconomy,”productcyclestypicallytookeighteenmonthstocomplete, so this was an exceptionally short time frame even in the “neweconomy.”SoIwenttoseeourdepartmenthead,MikeHomer.With thepossible exceptionofMarc,MikeHomerwas themost significant

creative forcebehindNetscape.More important, theworseasituationbecame,the stronger Mike would get. During particularly brutal competitive attacks,mostexecutiveswouldrunfromthepress.Mike,ontheotherhand,wasalwaysfront and center. WhenMicrosoft unveiled its famous “embrace and extend”strategy—a dramatic pivot to attack Netscape—Mike took every phone call,sometimeseventalkingtotworeportersatoncewithaphoneineachhand.Hewastheultimatewarrior.MikeandIspentthenextseveralmonthsdevelopingacomprehensiveanswer

toMicrosoft’s threat. If they were going to give our products away, then wewere going to offer a dirt-cheap, open alternative to the highly expensive andproprietary Microsoft BackOffice product line. To do so, we acquired twocompanies, which provided us with a competitive alternative to MicrosoftExchange.WethencutalandmarkdealwiththedatabasecompanyInformixtoprovideusunlimitedrelationaldatabaseaccessthroughtheWebfor$50acopy,which was literally hundreds of times less than Microsoft charged. Once weassembledtheentirepackage,MikenameditNetscapeSuiteSpot,asitwouldbethe“suite”thatdisplacedMicrosoft’sBackOffice.WelinedeverythingupforamajorlaunchonMarch5,1996,inNewYork.Then, just twoweeks before the launch,Marc,without tellingMike orme,

revealed the entire strategy to the publicationComputer Reseller News. Iwaslivid.Iimmediatelysenthimashortemail:

To:MarcAndreessenCc:MikeHomer

From:BenHorowitzSubject:LaunchIguesswe’renotgoingtowaituntilthe5thtolaunchthestrategy.—Ben

Withinfifteenminutes,Ireceivedthefollowingreply.

To:BenHorowitzCc:MikeHomer,JimBarksdale(CEO),JimClark(Chairman)From:MarcAndreessenSubject:Re:LaunchApparentlyyoudonot understandhow serious the situation is.Weare

gettingkilledkilledkilledoutthere.Ourcurrentproductisradicallyworsethan the competition.We’ve had nothing to say formonths. As a result,we’ve lost over $3B in market capitalization. We are now in danger oflosingtheentirecompanyandit’sallserverproductmanagement’sfault.Nexttimedothefuckinginterviewyourself.Fuckyou,Marc

IreceivedthisemailthesamedaythatMarcappearedbarefootandsittingona throne on the cover of Time magazine. When I first saw the cover, I feltthrilled.IhadnevermetanyoneinmylifewhohadbeenonthecoverofTime.ThenIfeltsick.IbroughtboththemagazineandtheemailhometoFeliciatogetasecondopinion. Iwasveryworried. Iwas twenty-nineyearsold,hadawifeandthreechildren,andneededmyjob.Shelookedattheemailandthemagazinecoverandsaid,“Youneedtostartlookingforajobrightaway.”Intheend,Ididn’tgetfiredandoverthenexttwoyears,SuiteSpotgrewfrom

nothingtoa$400millionayearbusiness.Moreshocking,MarcandIeventuallybecamefriends;we’vebeenfriendsandbusinesspartnerseversince.People often ask me how we’ve managed to work effectively across three

companies over eighteen years.Most business relationships either become tootensetotolerateornottenseenoughtobeproductiveafterawhile.Eitherpeoplechallenge each other to the point where they don’t like each other or theybecomecomplacentabouteachother’sfeedbackandnolongerbenefitfromtherelationship.WithMarcandme,evenaftereighteenyears,heupsetsmealmostevery day by finding somethingwrong inmy thinking, and I do the same forhim.Itworks.

STARTINGACOMPANYAttheendof1998andunderimmensepressurefromMicrosoft,whichusedthefull forceof itsoperatingsystemmonopoly tosubsidizefreeproducts ineverycategoryinwhichNetscapecompeted,wesoldthecompanytoAmericaOnline(AOL).Intheshortterm,thiswasabigvictoryforMicrosoftsinceithaddrivenitsbiggest threat into thearmsofa far less threateningcompetitor. In the longterm,however,NetscapeinflictedirreparabledamageonMicrosoft’sstrongholdon the computing industry: our work moved developers from Win32 API,Microsoft’s proprietary platform, to the Internet. Someone writing newfunctionalityforcomputersnolongerwroteforMicrosoft’sproprietaryplatform.Instead, theywrote to the InternetandWorldWideWeb’s standard interfaces.Once Microsoft lost its grip on developers, it became only a matter of timebefore it lost its monopoly on operating systems. Along the way, Netscapeinvented many of the foundational technologies of the modern Internet,includingJavaScript,SSL,andcookies.OnceinsideAOL,Iwasassignedtorunthee-commerceplatformandMarc

becamethechieftechnologyofficer.Afterafewmonths,itbecameapparenttobothofusthatAOLsawitselfasmoreofamediacompanythanatechnologycompany.Technologyenabledgreatnewmediaprojects,butthestrategywasamedia strategy and the top executive, Bob Pittman, was a genius mediaexecutive. Media companies focused on things like creating great storieswhereastechnologycompaniesfocusedoncreatingabetterwayofdoingthings.Webegantothinkaboutnewideasandaboutforminganewcompany.Intheprocess,weaddedtwootherpotentialcofounderstothediscussion.Dr.

TimothyHoweswas coinventor of theLightweightDirectoryAccess Protocol(LDAP),amasterfulsimplificationofitsbyzantineX.500predecessor.WehiredTim into Netscape in 1996 and together we successfully made LDAP theInternetdirectorystandard.Tothisday,ifaprogramisinterestedininformationabout a person, it accesses that information viaLDAP.The fourthmember ofour teamwas InSikRhee,whohadcofoundedanapplicationservercompanycalledKivaSystems,whichNetscapehadacquired.HehadbeenactingasCTOofthee-commercedivisionthatIranand,inparticular,workedcloselywiththepartnercompaniesinmakingsurethattheycouldhandletheAOLscale.Aswediscussedideas,InSikcomplainedthateverytimewetriedtoconnect

an AOL partner on the AOL e-commerce platform, the partner’s site wouldcrash,becauseitcouldn’thandlethetrafficload.Deployingsoftwaretoscaleto

millionsofuserswastotallydifferentfrommakingitworkforthousands.Anditwasextremelycomplicated.Hmm,thereoughttobeacompanythatdoesallthatforthem.Asweexpandedtheidea,welandedontheconceptofacomputingcloud.The

term cloud had been used previously in the telecommunications industry todescribethesmartcloudthathandledallthecomplexityofrouting,billing,andthelike,sothatonecouldplugadumbdeviceintothesmartcloudandgetallthesmart functionality for free. We thought the same concept was needed incomputing,so thatsoftwaredeveloperswouldn’thave toworryaboutsecurity,scaling,anddisasterrecovery.Andifyouaregoingtobuildacloud,itshouldbebigandloud,andthat’showLoudcloudwasborn.Interestingly,themostlastingremnant of Loudcloud is the name itself, as the word cloud hadn’t beenpreviouslyusedtodescribeacomputingplatform.Weincorporatedthecompanyandsetouttoraisemoney.Itwas1999.

—CHAPTER2—

“IWILLSURVIVE”

“DidyouthinkI’dcrumble?DidyouthinkI’dlaydownanddie?

Ohno,notIIwillsurvive.”

—GLORIAGAYNOR,“IWILLSURVIVE”

ComingoffthesuccessofNetscape,MarcknewallthetopventurecapitalistsinSilicon Valley, so we needed no introductions. Unfortunately for us, KleinerPerkins, the firm that backed Netscape, had already funded a potentiallycompetitivecompany.Wespoketoalltheothertop-tierfirmsanddecidedtogowithAndyRachleffofBenchmarkCapital.If I had to describe Andy with one word, it would be gentleman. Smart,

refined, and gracious, Andy was a brilliant abstract thinker who couldencapsulatecomplexstrategiesintopithysentenceswithease.Benchmarkwouldinvest$15millionatapre-moneyvaluation(thevalueofthecompanybeforethecashgoes into the company treasury)of$45million. In addition,Marcwouldinvest$6million,bringingthetotalvalueof thecompanyincludingitscashto$66 million, and would serve as our “full-time chairman of the board.” TimHoweswouldbeourchieftechnologyofficer.IwouldbeCEO.Loudcloudwastwomonthsold.Thevaluationandthesizeofthefundingweresignsofthetimesandcreated

an imperative to get big and capture the market before similarly well-fundedcompetitorscould.Andysaid tome,“Ben, thinkabouthowyoumight run thebusinessifcapitalwerefree.”Two months later, we would raise an additional $45 million fromMorgan

Stanley indebtwithnocovenantsandnopayments for threeyears, soAndy’s

question was more reality-based than you might think. Nonetheless, “Whatwould you do if capital were free?” is a dangerous question to ask anentrepreneur. It’s kind of like asking a fat person, “Whatwould you do if icecream had the exact same nutritional value as broccoli?” The thinking thisquestionleadstocanbeextremelydangerous.Naturally, I took the advice and ranwith it.Wequicklybuilt outour cloud

infrastructure and began signing up customers at a rapid rate. Within sevenmonthsof founding,we’dalreadybooked$10million incontracts.Loudcloudwas taking off, but wewere in a race against time and the competition. Thismeant hiring the best people and fielding the broadest cloud service, and thatmeantspendingmoney—lotsofit.Ourninthhirewasarecruiter,andwehiredahumanresourcespersonwhen

we had a dozen employees. We were hiring thirty employees a month andsnaggingmanyoftheValley’ssmartestpeople.Oneofournewrecruitshadquithis jobatAOLtospend twomonthsmountainclimbing,but insteadhe joinedus;anotherforfeitedmillionstojoinLoudcloudwhenheresignedfromanothercompany on the day of its IPO. Six months in, we had nearly two hundredemployees.SiliconValleywasonfire,andLoudcloudwasbilledinaWiredcoverstory

as “Marc Andreessen’s second coming.” We traded our first office—whereyou’dblowacircuitifyouranthemicrowaveandcoffeemakeratthesametime—for a fifteen-thousand-square-foot warehouse in Sunnyvale, which was toosmallforusbythetimewemovedin.Wespent$5milliontomoveintoanewthree-storystuccobuildingwithjade-

coloredtileswecalled“theTaj”(asintheTajMahal).Itwasalsotoosmalltokeeppacewithourhiring frenzy, andpeoplewere sitting in thehallways.Werentedathirdparkinglotdownthestreetandranshuttlevanstotheoffice.(Theneighborshatedus.)ThekitchenwasstockedlikeCostco,andwhenwefiredthesnack contractor for making our fridge look like the one in Philip Roth’sGoodbye,Columbus,heaskedforequity.Thiswasthetime.In thenext quarter,webooked$27millionworthof newcontracts, andwe

were less than ninemonths old. It seemed likewewere building the greatestbusinessofalltime.Thencamethegreatdot-comcrash.TheNASDAQpeakedat5,048.62onMarch10,2000—morethandoubleitsvaluefromtheyearbefore—and then fell by 10 percent ten days later. A Barron’s cover story titled“Burning Up” predicted what was to come. By April, after the government

declaredMicrosoftamonopoly,theindexplummetedevenfurther.Startupslostmassivevalue,investorslostmassivewealth,anddot-coms,onceheraldedastheharbingerofaneweconomy,wentoutofbusinessalmostovernightandbecameknownasdot-bombs.TheNASDAQeventuallyfellbelow1,200,an80percentdropfromitspeak.We thoughtourbusinessmighthavebeen the fastestgrowingof all timeat

thatpoint.Thatwasthegoodnews.Thebadnewswasthatweneededtoraiseeven more money in this disastrous climate; nearly all of the $66 million inequityanddebtwehadraisedhadalreadybeendeployed inourquest tobuildthe number-one cloud service and to support our now fast-growing set ofcustomers.Thedot-comcrashhadspooked investors,soraisingmoneywasn’tgoing to

be easy, especially since most of our customers were dot-com startups. Thisbecame quite clear when we pitched the deal to the Japanese firm SoftbankCapital. My friend and Loudcloud board member Bill Campbell knew theSoftbank people well and offered to get some “back-channel” informationfollowing the pitch. When my assistant told me that Bill was on the line, Iquicklyansweredthephone.Iwaseagertohearwherewestood.I asked, “Bill, what did they say?”Bill replied in his raspy, coach’s voice,

“Ben,well,honestly,theythoughtyouweresmokingcrack.”Withnearlythreehundredemployeesandverylittlecashleft,IfeltlikeIwasgoingtodie.ItwasthefirsttimeI’dfeltthatwayasCEOofLoudcloud,butnotnearlythelast.DuringthistimeIlearnedthemostimportantruleofraisingmoneyprivately:

Lookforamarketofone.Youonlyneedoneinvestortosayyes,soit’sbesttoignoretheotherthirtywhosay“no.”WeeventuallyfoundinvestorsforaseriesCround(meaningourthirdroundoffunding)atanamazing$700millionpre-moneyvaluationandraised$120million.Thesalesforecastforthequartercameinat$100million,and thingsseemedlike theymightbeokay. I feltconfidentthat our sales forecasts would hold up given that previous forecasts hadunderestimated actual performance. And perhaps, I speculated, we couldseamlesslymigrateourcustomerbaseawayfromdot-combombstomorestable,traditionalcustomerssuchasNike,ourlargestcustomeratthetime.Andthenthewheelscameoff.Wefinishedthethirdquarterof2000with$37millioninbookings—notthe

$100million thatwehadforecast.Thedot-comimplosion turnedout tobefarmorecatastrophicthanwehadpredicted.

EUPHORIAANDTERRORI needed to raisemoney yet again. Only this time the environment was evenworse.Inthefourthquarterof2000,Imetwitheverypossiblefundingsource,includingPrinceAl-WaleedbinTalalofSaudiArabia,butnobodywaswillingto investmoneyat anyvaluation.We’dgone frombeing thehottest startup inSiliconValleytounfundableinsixmonths.With477employeesandabusinessthatresembledatickingtimebomb,Isearchedforanswers.Thinking about what might happen if we ran completely out of money—

layingoffalltheemployeesthatI’dsocarefullyselectedandhired,losingallmyinvestors’ money, jeopardizing all the customers who trusted us with theirbusiness—madeitdifficulttoconcentrateonthepossibilities.MarcAndreessenattemptedtocheermeupwithanot-so-funny-at-the-timejoke:Marc:“Doyouknowthebestthingaboutstartups?”Ben:“What?”Marc: “You only ever experience two emotions: euphoria and terror.And I

findthatlackofsleepenhancesthemboth.”With the clock ticking, one unattractive but intriguing option emerged:We

couldgopublic.Inanoddityofthetimes,theprivatefundingmarketshutdownforcompanieswithourprofile,but thewindowonthepublicmarketremainedjustslightlyopen.Thismaysoundlikeacrazyanomalyanditwas,butprivatefunds had become completely cynical while the public markets were only 80percentofthewaythere.Withnootheroptionsavailable,Ineededtoproposetotheboardthatwego

public.Inordertoprepare,ImadealistoftheprosandconsofanIPO.Iknew thatBillCampbellwouldbe thecriticalperson I’dneed topersuade

onewayoranother.BillwastheonlyoneofourboardmemberswhohadbeenapubliccompanyCEO.Heknewtheprosandconsbetterthananyoneelse.Moreimportant, everybody always seemed to defer to Bill in these kinds of stickysituations,becauseBillhadaspecialqualityabouthim.Atthetime,Billwasinhissixties,withgrayhairandagruffvoice,yethehad

theenergyofatwenty-year-old.Hebeganhiscareerasacollegefootballcoachanddidnotenterthebusinessworlduntilhewasforty.Despitethelatestart,BilleventuallybecamethechairmanandCEOofIntuit.Followingthat,hebecamealegend in high tech,mentoring great CEOs such as Steve Jobs ofApple, JeffBezosofAmazon,andEricSchmidtofGoogle.Billisextremelysmart,super-charismatic,andeliteoperationally,butthekey

to his success goes beyond those attributes. In any situation—whether it’s theboardofApple,wherehe’sservedforoveradecade; theColumbiaUniversityBoard of Trustees, where he is chairman; or the girls’ football team that hecoaches—Billisinevitablyeverybody’sfavoriteperson.People offer many complex reasons for why Bill rates so highly. In my

experience it’s pretty simple. Nomatter who you are, you need two kinds offriends in your life. The first kind is one you can call when something goodhappens, and you need someone who will be excited for you. Not a fakeexcitement veiling envy, but a real excitement. You need someone who willactuallybemoreexcitedforyouthanhewouldbeifithadhappenedtohim.Thesecondkindoffriendissomebodyyoucancallwhenthingsgohorriblywrong—when your life is on the line and you only have one phone call.Who is itgoingtobe?BillCampbellisbothofthosefriends.I presented my thinking as follows: “We have not been able to find any

investors in the private markets. Our choices are to either keep working onprivate fundingorstartpreparing togopublic.Whileourprospects for raisingmoneyprivatelyseemquitedifficult,goingpublichasalargenumberofissues:

“Oursalesprocessesarenotrobustandit’sdifficulttoforecastinanyenvironment.“Wearenotinanyenvironment;weareinarapidlydecliningenvironmentandit’snotclearwherethebottomis.“Ourcustomersaregoingbankruptatanalarmingandunpredictablerate.“Wearelosingmoneyandwillbelosingmoneyforquitesometime.“Wearenotoperationallysound.“Ingeneral,wearenotreadytobepublic.”

Theboardlistenedcarefully.TheirexpressionsshoweddeepconcernwiththeissuesI’draisedandanawkwardlylongsilenceensued.Asexpected,Billbrokethedeadair.“Ben,it’snotthemoney.”Ifeltastrangesenseofrelief.Maybewedidn’thavetogopublic.MaybeI’d

overestimatedourcashproblems.Perhapstherewasanotherway.ThenBillspokeagain,“It’sthefuckingmoney.”Okay,Iguesswe’regoingpublic.In addition to the issues I had outlined for the board, our business was

complexandhardforinvestorstounderstand.Wetypicallysignedcustomerstotwo-yearcontracts,andthenrecognizedtherevenuemonthly.Thismodelisnowcommon,butitwasquiteunusualthen.Giventhefastgrowthinourbookings,revenue lagged behind new bookings by quite a bit.As a result, our S-1 (ourregistrationwiththeSEC)statedthatwehad$1.94millionintrailingsixmonthsrevenue,andweforecast$75millionforthefollowingyear—anincrediblysteeprevenueramp.Sinceearningsaredrivenbyrevenueandnotbookings,wehadgiganticlosses.Inaddition,thestockoptionrulesatthetimemadeitseemlikeourlosseswereaboutfourtimesaslargeastheyactuallywere.ThesefactorsledtoextremelynegativepressheadingintotheIPO.AscathingstoryinRedHerring,forinstance,notedthatourlistofcustomers

was“quite thin”andthatwewere tooreliantondot-coms.ItquotedaYankeeGroupanalystpositing thatwehad“lost something like$1milliondollarsperemployeeoverthelast12months,”andconjecturingthatthewaywediditwasbyhavingabonfireintheparkinglotandgettingeveryonebusyburningdollarbills.BusinessWeek took us apart in an article that declared us “the IPO fromhell.”AWallStreetJournalcoverpiecequotedamoneymanager’sreactiontoour offering as “Wow, they were desperate.” One financier—who actuallyinvestedintheoffering—calledit“thebestoptionamongaparticularlyuglysetofoptions.”Despite the horrifying press, we prepared to hit the road. Benchmarking

ourselvesagainstcomparablecompanies,wesettledonthepriceoftheofferingat$10pershareafteranupcomingreversesplit,whichwouldvaluethecompanyat just under $700million—less than the valuation from the previous privateroundoffinancing,butmuchbetterthanbankruptcy.Itwasnotatallclearthatwewouldbesuccessfulwiththeoffering.Thestock

market was crashing, and the publicmarket investors we visitedwere visiblydistressed.Attheendofthepreparationprocessandafterthebankshadsignedoff,our

director of finance, Scott Kupor, received a call from our banker at MorganStanley.Banker:“Scott,didyouknowthat$27.6millionofyourcashisrestrictedand

tiedupinrealestatecommitments?”Scott:“Yes,ofcourse.”Banker: “So, you have just over threeweeks’worth of cash before you go

bankrupt?”Scott:“Yes.”

Scott then relayed the conversation to me, saying, “Can you believe theyunderwrotethedealanddidn’tnoticethatthecashwasrestricteduntilnow?Wegavethemallthedocuments.”RightbeforeweweretoleavefortheIPOroadshow,Icalledanall-company

meetingtosharetwopiecesofnews:First,weweregoingpublic,oratleastweweregoingto try togopublic.Second, thecompanyhadfallensofar invaluethatwewouldhavetoreversesplitthestocktwoforone.Ithoughtthefirstpartwouldgookay,butIwasworriedabouthowthesecond

piece of newswould be received.Wehad to reverse split the stock to get theprice per share high enough to go public. In theory, a reverse split shouldn’tmatter at all.Eachemployeeowneda certainpercentageof the company.Thecompany had a total number of shares of stock.Multiply the total number ofshares by the percentage, and you get the employees’ share number. Cut thenumberofshares inhalfand,whileemployeeswouldhavehalf thenumberofshares, they’d still own the exact same percentage of the company. Nothingchanged.Oh,but itdid.Aswegrewfromzero tosixhundredemployees in less than

eighteenmonths, thestagewassetforhyperboleandmomentum.Someoverlyexcitedmanagersoversoldthedream.Theyspokeonlyintermsofsharesratherthan inpercentages and spun storiesof apotential $100per share stockprice.Employees then calculated their fantasy price per share and figured out howmuchmoneytheywouldmake.Iwasawarethatthiswasgoingon,butIneverthoughtwewouldreversesplitthestock,soIneverworriedaboutit.LikemanyotherthingsthatIscrewedupduringthatperiod,Ishouldhaveworried.Mywife, Felicia, came to the all-companymeeting as she always did.This

timeherparentswereintown,sotheycame,too.Themeetingdidnotgowell.Peopledidnot realizehowclose to theedgewewere, so thenewsof the IPOdidn’tmakeanyonehappy.Thenewsof thereversesplitmade themeven lesshappy—infact,itinfuriatedthem.Ihadliterallycuttheirfantasynumberinhalf,andtheywerenotpleasedaboutit.Nobodysaidharshthingsdirectlytome.Myin-laws,however,heardeverything.And,asmyfather-in-lawput it,“itwasn’tnothin’nice.”Mymother-in-law,Loretta,askedmywife,“WhydoeseverybodyhateBenso

much?”Felicia,whoisnormallythemostelectric,outgoingpersoninanyroom,was just recovering fromhernia surgery so shewasn’thernormalbubbly self.Shewasdiscouraged.Myin-lawsweredepressed.Theemployeeswerepissed.IhadnoideaifI’dbeabletoraisethemoney.Whatawaytostartaroadshow,an

eventthat’susuallythecauseofabitmorefanfare.The road showwas brutal. The stockmarket crashed daily, and technology

stocksweretoblame.Investorslookedlikethey’dcomeoutoftorturechamberswhenwearrived.Onemutual fundmanager looked right atMarc andmeandasked,“Whyareyouhere?Doyouhaveanyideawhat’sgoingonintheworld?”Ithoughtthattherewasnowaywe’dbeabletoraisethemoney.Weweregoingto go bankrupt for sure. I did not sleepmore than two hours total during thatentirethree-weektrip.Threedaysintothetour,Ireceivedacallfrommyfather-in-law.JohnWiley

had been through a lot in his seventy-one years. As a boy, his father wasmurdered in Texas. In order to survive, he and hismothermoved in with anunkindmanandhisninechildren.There,Johnwasabused,madetostayinthebarnwiththeanimals,whiletheotherchildrenatehisdinner.Eventually,Johnand his mother left that cruelty by walking for three days down a dirt road,carrying everything theyowned. Johnwould recall that journey ingreat detailhisentire life.Asayoungman,before finishinghishigh schooleducation,helefthometofightintheKoreanWarsothathecouldsupporthismother.Asayoung father of five, he took every job imaginable to support his family,includingunloadingbananaboatsandworkingtobuildtheAlaskanpipeline.Hetragicallysawtwoofhischildrendiebeforehereachedtheageofsixty.Hehadahardlifeandwasusedtobadnews.JohnWiley did not callme for casual reasons. If he called, it was serious,

possiblyevendeadlyserious.Ben:“Hello.”John:“Ben,theofficesaidnottobotheryou,butIjustwanttoletyouknow

thatFeliciastoppedbreathing,butsheisnotgoingtodie.”Ben:“Notgoingtodie?What?!?!Whathappened?”Icouldnotbelieveit.IhadbeensofocusedonworkthatIhadlostfocusof

theonlythingthatreallymatteredtome.Onceagain,IneglectedtoworryabouttheonethingthatIshouldhaveworriedabout.Ben:“Whathappened?”John:“Theygavehersomemedicineandshehadanallergicreactionandshe

stoppedbreathing,butshe’sokaynow.”Ben:“When?”John:“Yesterday.”Ben:“What?Whydidn’tyoutellme?”John:“Iknewthatyouwerebusyandthatyouwerereallyintroubleatwork

becauseofthatmeetingthatIwentto.”Ben:“ShouldIcomehome?”John:“Ohno.We’lltakecareofher.Youjusttakecareofwhatyouneedto

do.”Iwascompletelystunned.IstartedsweatingsohardthatIhadtochangemy

clothes right after the call. I had no idea what to do. If I returned home, thecompanywouldsurelygobankrupt. If I stayed . . .howcould I stay? IcalledbackandhadhimputFeliciaonthephone.Ben:“Ifyouneedme,Iwillcomehome.”Felicia: “No. Get the IPO done. There is no tomorrow for you and the

company.I’llbefine.”I stumbled through the rest of the road show completely discombobulated.

OnedayIworeamismatchingsuitjacketandsuitpants,whichMarcpointedouttomemidway through themeeting. I had no ideawhere Iwas half the time.During the three weeks we were on the road, comparable companies in ourmarketlosthalfoftheirvalue,whichmeantthatour$10sharepricewasroughlydouble the current benchmark. The bankers recommended that we lower thepriceof theoffering to$6a share inorder to reflect thisnewreality,but theygaveusnoassurancethatthedealwouldactuallygetdone.Then,thedaybeforethe offering,Yahoo, the lighthouse company of the Internet boom, announcedTimKoogle,itsCEO,wassteppingdown.Wehadhitthenadirofthedot-comcrash.The Loudcloud offering finally sold at $6 a share, and we raised $162.5

million,buttherewasnocelebrationandnoparty.NeitherGoldmanSachsnorMorgan Stanley—the two banks that took us public—even offered us thetraditionalclosingdinner.ItmayhavebeentheleastcelebratoryIPOinhistory.ButFeliciawas feeling better, andwehadpulled it off. In a briefmoment oflightheartednessontheplaneridehome,IturnedtoScottKupor,mydirectoroffinance,andsaid,“Wedidit!”Hereplied,“Yeah,butwe’restillfucked.”Years later, in 2012, after Yahoo fired its CEO, Scott Thompson, Felicia

mused,“ShouldtheybringbackKoogle?”Ireplied,“TimKoogle?Howdoyoueven knowwhoTimKoogle is?” She then relived the conversationwe’d hadelevenyearsearlier.Itwentsomethinglikethis:Ben:“We’refucked.”Felicia:“Whatdoyoumean?Whathappened?”Ben:“YahoofiredKoogle.It’sover.Thewholethingisover.”Felicia:“WhoisKoogle?”

Ben:“HewastheCEOofYahoo.We’refucked.I’mgoingtohavetoshutthecompanydown.”Felicia:“Areyousure?”Ben:“Didn’tyouhearme?TheyfiredKoogle.We’refucked.”Shehadneverseenmethatdepressedbefore,andsheneverforgotit.Formost

CEOs, the night before their public offering is a highlight. For me, it was ahighlightofdepression.

IFYOUAREGOINGTOEATSHIT,DON’TNIBBLEDuring the road show, as a way to break the tension, Marc would say,“Remember,Ben, thingsare alwaysdarkestbefore theygocompletelyblack.”Hewas joking,but asweenteredour first quarter as apublic company, thosewords seemed prescient. Customers continued to churn, the macroeconomicenvironmentworsened,andoursalesprospectsdeclined.Aswegotclosertoourfirst earnings callwith investors, I conducted a thorough review tomake surethatwewerestillontracktomeetourguidance.Thegoodnewswasthatwewouldmeetourforecastforthequarter.Thebad

news:Therewasverylittlechancethatwewouldmeetourforecastfortheyear.Typically,investorsexpectthatcompanieswillrefrainfromgoingpubliciftheycan’t hit at least their first year’s forecast. These were exceptional times, butresettingguidanceonyourveryfirstearningscallwasstillaverybad thing todo.Aswediscussedwhere to resetguidance to investors,wewerefacedwitha

toughchoice:Shouldwetrytominimizetheinitialdamagebytakingdownthenumberaslittleaspossibleorshouldweminimizetheriskofanotherreset?Ifwereducedthenumberbyalot,thestockmightfallapart.Ontheotherhand,ifwedidn’tloweritenough,wemighthavetoresetagain,whichwouldcostusallthecredibilitywehadleft.Mycontroller,DaveConte,raisedhishandwithwhatwould be the definitive advice: “No matter what we say, we’re going to getkilled.Assoonasweresetguidance,we’llhavenocredibilitywithinvestors,sowe might as well take all the pain now, because nobody will believe anypositivityintheforecastanyway.Ifyouaregoingtoeatshit,don’tnibble.”Soweresetguidance for theyear, slashingouroriginal forecastof$75million inprojectedrevenueto$55million.Resetting revenue guidance alsomeant resetting expense guidance, and that

meantlayingpeopleoff.We’dbeenthedarlingofthestartupworld,andnowIhadtosendhome15percentofouremployees.Itwastheclearestindicationyet

that I was failing. Failing my investors, failing my employees, and failingmyself.Following the reset, Goldman Sachs and Morgan Stanley—the investment

banksthathadtakenuspublic—bothdroppedresearchcoverage,meaningtheiranalysts would no longer follow the company’s progress on behalf of theirclients.Thiswasahugeslapinthefaceandamassiverenegingofthepromisestheymadewhentheywerepitchingus,buttimesweretoughallaround,andwehadno recourse.With a vote of no confidence fromour banks and a loweredrevenueforecast,thestockpriceplummetedfrom$6ashareto$2.Despitethemammothnegativemomentum,wesoldieredon,andwereputting

togetherafairlystrongquarterinthethirdquarterof2001.Then,onSeptember11,terroristshijackedfourjetliners,flyingtwointotheWorldTradeCenterandanotherintothePentagon,andintheendthrowingthewholeworldintochaos.ItturnedoutthatourlargestdealthatquarterwaswiththeBritishgovernment.Itrepresentedone-thirdofourbookings,andwewouldmiss thequarter’s targetsbadly without it. Our champion on the deal called to inform us that PrimeMinister Tony Blair had redirected the funds for our deal to the war chest—literally. By some miracle our sales director convinced one of Tony Blair’sstafferstogetthemoneyback,sowegotthedealandmadethequarter.Nonetheless,theclosecallwasasigntomethattheentireoperationwasfar

too fragile. I got another sign when our largest competitor, Exodus, filed forbankruptcy on September 26. It was a truly incredible bankruptcy in that thecompanyhadbeenvaluedat$50billionalittlemorethanayearearlier.Itwasalso remarkable because Exodus had raised $800 million on a “fully fundedplan” justninemonthsearlier.AnExodusexecutive later joked tome:“Whenwedroveoffthecliff,weleftnoskidmarks.”IfExoduscouldlose$50billioninmarketcapitalizationand$800millionincashthatfast,Ineededabackupplan.In my first attempt at a “Plan B,” we evaluated acquiring Data Return, a

company like ours that focused more on Windows applications than Unixapplicationsaswedid.Westudied thedeal forweeks,modelingwhat the twocompanies might look like together, figuring out product offerings and costsynergies.MyCFO at the timewas extremely excited about the deal since itwouldmakeuseofhisfavoriteskillset—costcutting.Towardtheendoftheprocess,Itookatwo-dayvacationtoAshland,Oregon.

AlmostassoonasIarrived,IreceivedanurgentcallfromJohnO’Farrell,whowasinchargeofcorporateandbusinessdevelopment.John:“Ben,sorrytodisturbyouonvacation,butwejusthadameetingonthe

DataReturndealandIdon’tthinkthatweshoulddoit.”Ben:“Whynot?”John:“Quitefrankly,ourbusinessisintroubleandtheirbusinessisintrouble

andputtingthemtogetherwilljustbedoubletrouble.”Ben:“Iwasthinkingtheexactsamething.”In fact, looking at Data Return’s business made it crystal clear to me that

Loudcloudwouldprobablynotendwell.Somethingsaremucheasiertoseeinothersthaninyourself.LookingatDataReturn,IcouldseeLoudcloud’sfuture,anditwasnotpretty.IhadagreatdealoftroublesleepingasIthoughtaboutourfate. I tried tomakemyself feelbetterbyasking,“What’s theworst thing thatcouldhappen?”Theansweralwayscamebackthesame:“We’llgobankrupt,I’lllose everybody’s money including my mother’s, I’ll have to lay off all thepeople who have been working so hard in a very bad economy, all of thecustomerswho trustedmewillbescrewed,andmyreputationwillbe ruined.”Funny,askingthatquestionnevermademefeelanybetter.Then one day I askedmyself a different question: “Whatwould I do ifwe

went bankrupt?” The answer that I came up with surprisedme: “I’d buy oursoftware, Opsware, which runs in Loudcloud, out of bankruptcy and start asoftwarecompany.”Opswarewasthesoftwarethatwe’dwrittentoautomateallthetasksofrunningthecloud:provisioningserversandnetworkingequipment,deploying applications, recovering the environment in case of disaster, and soforth.ThenIaskedmyselfanotherquestion:“Isthereawaytodothatwithoutgoingbankrupt?”I ran throughdifferentscenarios inmymindwherewemightmove into the

software business and exit the cloud business. In each scenario, step onewasseparatingOpsware fromLoudcloud.Opswarehadbeenwritten to runonly inLoudcloudandhadmanyconstraintsthatpreventeditfrombeingaproductthatwouldwork in any environment. I askedmy cofounder andCTOTimHoweshowlongitwouldtaketoseparateOpswarefromLoudcloud.Hesaidaboutninemonths, which would prove to be quite optimistic. I immediately assigned ateamoftenengineerstostarttheprocessinaprojectwecalledOxide.Atthispoint,ourbusinesswasstillacloudbusiness,andIgavenoindication

to the rest of the staff that I might have other ideas. Doing so would haveinstantlydoomedtheonlybusinesswewerein,aseveryonewouldwanttoworkonthefutureandnotthepast.IsaidthatOxidewassimplyanotherproductline.This statement deeplyworried twoofmy employeeswhohad graduated fromStanford Business School. They scheduled an appointment and presented me

with a slide deck detailing why my decision to start Oxide was quixotic,misguided, and downright stupid. They argued that it would steal preciousresourcesfromourcorebusinesswhilepursuingaproductthatwouldsurelyfail.Iletthempresentallforty-fiveslideswithoutmyaskingthemasinglequestion.WhentheyfinishedIsaid,“DidIaskforthispresentation?”ThosewerethefirstwordsIspokeasImadethetransitionfromapeacetimeCEOtoawartimeCEO.Byvirtueofmypositionandthefactthatwewereapubliccompany,nobody

besides me had the complete picture. I knew we were in deep, deep trouble.Nobodybesidesmecouldgetusoutofthetrouble,andIwasthroughlisteningtoadviceaboutwhatweshoulddofrompeoplewhodidnotunderstandallthepieces.Iwantedall thedataandinformationIcouldget,butIdidn’tneedanyrecommendationsaboutthefuturedirectionofthecompany.Thiswaswartime.Thecompanywouldliveordiebythequalityofmydecisions,andtherewasnoway to hedge or soften the responsibility. If everybody I had hired—andwhogavetheir livesto thecompany—couldbesenthomewithlittle toshowfor it,then there were no excuses that would help. There would be no: “It was ahorrible economic environment”; “I got bad advice”; “Things changed soquickly.”Theonlychoicesweresurvivalor totaldestruction.Yes,most thingscould still be delegated and most managers would be empowered to makedecisionsintheirareasofexpertise,butthefundamentalquestionofwhether—andhow—Loudcloudcouldsurvivewasmineandminealonetoanswer.Wemuddled through the fourth quarter of 2001 and beat our target for the

year,delivering$57million in revenueagainstour$55million forecast.Notagreatwin,butveryfewcompaniesmetexpectations thatyear,soI tookitasasmallvictory.Thestockpriceslowlyroseto$4ashare,anditlookedasthoughwemightbeabletomakethecloudbusinesswork.Inordertodoso,weneededmorecash.Wecarefullyanalyzedourfinancial

plan and decided that we needed another $50 million to get to cash flowbreakeven—thepointatwhichwewouldnolongerneedtoraisemoney.Givenourmomentum in themarket, raisingmoneywasnowbarelypossible and theonlyway to do itwas in the formof a seldom-used construct called a privateinvestmentinpublicequity(PIPE).WeworkedwithMorganStanleytolineupinvestorswiththegoalofraising$50million.ItwasMondaymorning, andwewere all set to hit the road onTuesday to

raise thePIPEwhen I got the call. “Ben, theCEOofAtriax is on the phone;shallIputhimthrough?”Atriax,anonlineforeigncurrencyexchangebackedbyCitibankandDeutscheBank,wasourlargestcustomer.Atriaxpaidusmorethan

$1 million per month and had a two-year guaranteed contract. I was in themiddleofameetingwithDebCasados,myvicepresidentofhumanresources,butIsaid,“Puthimthrough.”HetheninformedmethatAtriaxwasbankruptandcouldnotpayanyofthe$25millionheowedus.Itwasliketheworldstoppedspinning.IsatthereinadazeuntilIheardDeb’svoicesaying,“Ben,Ben,Ben,doyouwanttohavethismeetinglater?”Isaid,“Yep.”IwalkedslowlyovertomyCFO’sofficetoassessthedamage.ItwasworsethanIthought.Giventhematerialityoflosingthecontract,wecouldnotraisemoneywithout

first disclosing thatwe’d lost our largest customer and $25million out of ourfinancial plan. We put the PIPE road show on hold and then issued a pressrelease.Thestockimmediatelyfellby50percent,andwitharapidlydecliningmarketcapof$160million,wecouldnolongerraise$50millioninaPIPE.Theplanthatwas$50millionshortofbreakevenwasnow,withthelossofAtriax,$75million short of breakevenwith noway to close the gap. Loudcloudwasdoomed.IhadtodeployOxide.Thesituationwascomplex,because440ofour450employeesworkedinthe

cloud business, which represented all of our customers and generated 100percentofourrevenue.Icouldnottelltheemployeesorevenmyexecutiveteamthat Iwas considering abandoning the cloudbusiness, becauseour stockpricewouldhave collapsed tonothing,killing anyhopeof selling the companyandavoidingbankruptcy.TheonepersonIneededandcouldtrustwasJohnO’Farrell.Johnranbusiness

and corporate development, but more than that he was the greatest big-dealpersonIhadeverknown.Toillustratemypoint, let’ssayyouwereareligiousman.We’re speaking in the hypothetical now.And let’s say you had reachedyour end of days and you faced yourmaker for final judgment. Let’s furthersupposethatasyourfatewastobedecidedforalleternity,youweregrantedasinglepersontonegotiateonyourbehalf.Whomwouldyouchoose?Well,ifitwereme,I’dtakethatIrishbrother,JohnO’Farrell.ItoldJohnthatheandIneededtoexecuteacontingencyplan,andweneeded

togetstartedimmediately.Thiswouldbeatwo-personprojecttostart,andweneededeveryoneelse focusedon the taskathand—reducingLoudcloud’scashburn.NextIcalledBillCampbelltoexplainwhyIthoughtweneededtoexitthecloudbusiness.Bill understood what a crisis looked like since he’d been CEO of GO

Corporation in the early 1990s. Essentially GO had attempted to build aniPhone-likedevicein1992andendedupbeingoneofthelargestventurecapital

losses inhistory. I tookBill throughmy logic:Theonlywayoutof thecloudbusinesswithoutgoingbankruptwas throughhigher sales,becauseeven ifwelaidoff100percentoftheemployees,theinfrastructurecostswouldstillkilluswithoutasharpersalesramp.Ifurtherexplainedthatthedwindlingcashbalancedecreased customer confidence,which in turnhurt sales,which in turn causedthecashbalancetodeclinefurther.Hesimplysaid“spiral.”AndIknewthatheunderstood.JohnandImappedouttheecosystemtofigureoutwhichcompaniesmightbe

interested in acquiring the Loudcloud business. Unfortunately, many of theprospective buyers were in dire straits themselves. Giant telecomsQwest andWorldComwereembroiled inaccounting fraudcases,andExodushadalreadygonebankrupt.Wedecidedtofocusonthethreemostlikelybuyers:IBM,Cable&Wireless,andEDS.IBM’shostingbusiness,ledbythecongenialJimCorgel,immediatelytooka

strong interest. Jim valued the Loudcloud brand and our reputation fortechnological superiority. EDS, on the other hand, showed no interest. ThisworriedmeintenselyasIstudiedall thepublicfilingsfrombothcompanies; itwas clear to me that EDS needed Loudcloud far more than IBM did. Needsalwaystrumpwantsinmergersandacquisitions.Johnsaidtome,“Ben,IthinkweneedtowalkawayfromEDS,sothatwecanfocusonthehigher-probabilitytargets.”IaskedhimtodrawtheEDSorganizationalchartonemoretimetoseeifwecould findsomeone influentialatEDSwhomwehadn’tyetapproached.Whenhedid,Iasked,“WhoisJeffKelly?”Johnpaused,thensaid,“Youknow,wehaven’tgottentoJeff,buthemaybeabletomakethisdecision.”Sure enough, Jeff was interested. Now with two potential bidders, we put

thingsinmotion.JohnandIworkedhardtocreateurgencywithbothIBMandEDS,becausetimewasagainstus.Wehostedbothcompaniesinourfacilities,sometimeswith thempassingeachother in thehallwayaspartofJohn’swell-orchestrated sales technique. The final step was to set the timeline for theendgame. John and I debated the bestway to do this as the deadlines thatweplannedtosetwereclearlyartificial.IsuggestedthatwestopbyLosAngelesonourwaytoPlano,Texas,homeofEDS,togetsomeadvicefromMichaelOvitz.MichaelwasonLoudcloud’sboard,butmoreimportant,hehadformerlybeen

knownbymanyobserversas theMostPowerfulMan inHollywood.Whenhewastwenty-eightyearsold,hestarteda talentagency,CreativeArtistsAgency(CAA),which grew to dominate the entertainment industry.CAA’s risemadeMichaelsoinfluentialthathecouldroutinelystructuredealsthathadneverbeen

donebefore.When we arrived in his offices, the place buzzed with activity. Michael

seemedtobeengagedinadozendifferentactivities,butfinallycameouttomeetwithJohnandme.Weexplainedthesituation:Wewereracingagainsttimeandhad twobidders, butno specific incentive to coax them toward the endof theprocess.Michaelpaused,thoughtforamoment,andthendeliveredhisadvice:“Gentlemen, I’vedonemanydeals inmy lifetime and through that process,

I’vedevelopedamethodology,awayofdoingthings,aphilosophyifyouwill.Withinthatphilosophy,Ihavecertainbeliefs.Ibelieveinartificialdeadlines.Ibelieve in playing one against the other. I believe in doing everything andanythingshortofillegalorimmoraltogetthedamneddealdone.”Michaelhadawayofmakingthingsextremelyclear.Wethankedhimandheadedtotheairport.WecalledbothEDSandIBMto

letthemknowthatwewouldcompletetheprocessoverthenexteightweeksandsell the Loudcloud business to someone. If they wanted to play, they had tomove on that schedule orwithdraw immediately. TheMichaelOvitz artificialdeadline was in full effect. We knew that we might have to go past it, butMichaelgaveusconfidencethatgoingpastthedeadlinewasabettermovethannothavingone.After seven weeks, we came to an agreement with EDS. They would buy

Loudcloud for $63.5 million in cash and assume its associated liabilities andcash burn.Wewould retain the intellectual property,Opsware, and become asoftwarecompany.EDSwouldthenlicenseoursoftwaretorunbothLoudcloudandthelargerEDSfor$20millionperyear.IthoughtitwasagreatdealforbothEDSandus.Itwascertainlyfarbetterthanbankruptcy.Ifelt150poundslighter.Icouldtakeadeepbreathforthefirsttimeineighteenmonths.Still,itwouldn’tbe easy. Selling Loudcloud meant selling about 150 employees to EDS andlayingoffanother140.IcalledBillCampbelltotellhimthegoodnews:Thedealwassignedandwe

would be announcing it in NewYork onMonday. He replied, “Too bad youcan’t go to NewYork and be part of the announcement; you’ll have to sendMarc.”Isaid,“Whatdoyoumean?”Hesaid,“Youneedtostayhomeandmakesureeverybodyknowswheretheystand.Youcan’twaitaday.Infact,youcan’twaitaminute.Theyneed toknowwhether theyareworking foryou,EDS,orlookingforafuckingjob.”Damn.Hewasright.IsentMarctoNewYorkandpreparedtoletpeopleknowwheretheystood.ThatsmallpieceofadvicefromBillprovedtobethefoundationweneededtorebuildthecompany.Ifwehadn’t

treatedthepeoplewhowereleavingfairly,thepeoplewhostayedwouldneverhavetrustedmeagain.OnlyaCEOwhohadbeenthroughsomeawful,horrible,devastatingcircumstanceswouldknowtogivethatadviceatthattime.

—CHAPTER3—

THISTIMEWITHFEELING

“Imoveonward,theonlydirectionCan’tbescaredtofailinsearchofperfection.”

—JAYZ,“ONTOTHENEXTONE”

OncetheEDSsalewascompleted,Ifeltlikethecompanywasingoodshape,butmy shareholders did not agree. I had sold all ofmy customers, all ofmyrevenue,and thebusiness theyunderstood.Every large shareholderbailedout,and thestockpricefell to$0.35pershare,whichrepresentedabouthalfof thecashwehadinthebank.Irealizedthatnobodybesidesmeknewhowbadthingshadbecomeandnobodybesidesmebelievedinthefuture,soIdecidedtotaketheemployeesoff-siteandresellthemontheopportunity.Irentedfortyroomsinalow-endmotelinSantaCruzandtookourremaining

eightyemployeesthereforonenightofdrinkingandonedayofexplainingtheOpswareopportunity.Attheendoftheday,Itriedtobeashonestashumanlypossible.“YouhavenowheardeverythingthatIknowandthinkabouttheopportunity

infrontofus.WallStreetdoesnotbelieveOpswareisagoodidea,butIdo.Icanunderstandifyoudon’t.Sincethisisabrand-newcompanyandabrand-newchallenge,Iamissuingeveryonenewstockgrantstoday.AllthatIaskisthatifyouhavedecidedtoquitthatyouquittoday.Iwon’twalkyououtthedoor—I’llhelpyou finda job.But,weneed toknowwherewestand.Weneed toknowwho iswithusandwhomwecancounton.Wecannotafford to slowlybleedout.Youoweittoyourteammatestobehonest.Letusknowwhereyoustand.”Thatday twoemployeesquit.Of theother seventy-eight, allbut twostayed

throughthesaletoHewlett-Packardfiveyearslater.Aftertheoff-sitegathering,thefirstthingIhadtodowasincreasethestock

price.TheNASDAQhadsentmeacurtletterstatingthatifwefailedtogetour

stockpriceoveradollar,theywould“delist”usfromtheexchangeandsendustothepurgatoryknownaspennystocks.Theboarddebatedthebestwaytodothis—reverse-split the stock, a stock buyback, or other options—but I felt wejust needed to tell our story.The storywas simple.Wehad agreat team,$60million in thebank,a$20millionayearcontractwithEDS,andsomeseriousintellectual property. Unless I was the worst CEO of all time, we should beworthmorethan$30million.Thestorytookhold,andthestockclimbedabove$1ashare.Next,Ihadtoshipaproduct.Opswarehadbeenbuilt torunLoudcloudand

Loudcloud only. Itwas not yet ready for theworld. In fact, parts of the codewere hardwired to physical machines in our building. Beyond that, the userinterfacewas far fromready forprime time.Thecomponent thatmanaged thenetworkwas called the Jive and featuredapurplepimphaton the frontpage.Project Oxide gave us a running start, but our engineers were nervous. Theybroughtmealonglistoffeaturesthattheyfeltweneededtocompletepriortoenteringthemarket.Theypointedtocompetitorswithmorefinishedproducts.AsIlistenedtotheirlengthyobjections,itbecamecleartomethatthefeatures

theengineerswantedtoaddallcamefromLoudcloudrequirements.Aspainfulas itmight be, I knew thatwe had to get into the broadermarket in order tounderstanditwellenoughtobuildtherightproduct.Paradoxically,theonlywaytodo thatwas toshipand try tosell thewrongproduct.Wewouldfallonourfaces,butwewouldlearnfastanddowhatwasneededtosurvive.Finally, I had to rebuild the executive team. I had aCFOwhodidn’t know

softwareaccounting,aheadofsaleswhohadneversoldsoftware,andaheadofmarketingwhodidnotknowourmarket.Everyoneof themwasgreatat theiroldjobs,butnotqualifiedfortheirnewjobs.Itwasmiserable,butnecessary,toseethemallgo.The strategy and the team came together, and the business startedworking.

Webegansigningcustomersataconsistentpaceandourstockpricerosefromits $0.35 low tomore than $7 a share. It felt like wewere finally out of thewoods.NaturallyIwaswrong.

SIXTYDAYSTOLIVEA few quarters into Opsware, we received very bad news from our largestcustomer,EDS.“Largestcustomer”reallyunderstatesit;EDSaccountedfor90percent of our revenue.And theywere not happy.TheirOpsware deployment

had stalled out and not met its goals as they had run into multiple difficulttechnicalissues.EDSwantedtocancelthedeployment,endthecontract,andgettheir money back. Giving EDS their money back would mean the end ofOpsware.Gettingintoabigdisputewithacustomerthataccountedforallbut10percentofourrevenuewouldalsomeantheendofOpsware.Weweredoomedagain.Icalledmytoptwolieutenantsontheaccountinforameeting.JasonRosenthalwastheveryfirstemployeeIhadhiredandthebestmanager

inthecompany.AStanfordgraduatewithanimpeccablememoryandageniusmindformanagingall thedetailsofacomplexproject,JasonwasinchargeoftheEDSdeployment.AnthonyWrightgrewupinthetoughpartofPittsburgh,thesonoflegendary

street fighter JoeWright, and had earned a black belt in several martial artshimself. Self-made, super-determined, and unwilling to fail, Anthony had anuncanny ability to quickly gain deep insight into people’s character andmotivations—“abletocharmdogsoffameattruck,”ishowanotherguyontheteamdescribedit.AnthonywastherelationshipmanagerforEDS.I beganwith an assessment:What happened? It turned out, a lot of things.

EDS’s environmentwas insane and chaotic. They had inherited networks andinfrastructure from every customer they’d ever signed and from every era inwhichtheyhadsignedthem.Theyhaddatacentersconnectedby56-kilobitlinksat a timewhen no other customer connected at speeds even twenty times thatslow.EDS ranversions of operating systems thatwere soold that theydidn’tsupportbasictechnologieslikethreads,whichmeantoursoftwarewouldn’trunon them.And thepeoplewerenotourpeople.We’d find themsleeping in thedata center at two o’clock in the afternoon; they were not motivated andgenerally not very happy. Beyond that, our product was far from perfect andevery one of the many bugs and shortcomings was a reason to stop thedeployment.I tooka longpause, rubbedmyhead, and thenbegan togive instructions. I

chosemywordscarefully:“I appreciate the difficulties andmore than that, I thank you deeply for the

effort.However,IdonotthinkthatI’vemademyselfclearonthesituationthatwe’rein.Thisisnotascenariowhereanexcusewilldo.Thisisamustwin.IfEDS drops us, we’re fucked and it’s over. The IPO, avoiding the Loudcloudbankruptcy,allthelayoffsandpainwillhavebeenfornothing—becausewe’redead.So,ouronlyoptionistowin.Wecannotlosethisone.

“Jason, thewhole company is at your command.Whateveryouneed, Iwillmake sureyouget it.Anthony, Jason isgoing towork todeliver all thevaluethatEDSexpects,buthewillfail.Hewillfailtodeliveronehundredpercentofexpectations,soyouarenowinchargeoffindingoutwhattheydon’texpect,butwant.You are in charge of finding the exciting value.When you do,wewilldeliverit.”Jason and Anthony then headed to Plano, Texas, to meet with their

counterpartsatEDS.Theydidn’tknowwhowasmakingdecisions,butafterabunchofmeetings

anddeadends, theyfound theirway to theofficeofapersonIwillcallFrankJohnson (not his real name)—a big guy who grew up in the oil fields ofOklahoma,graduatedfromWestPoint,andnowwas inchargeofanyonewhotouchedanyserversatEDS.AnthonyandJasontoutedtheOpswaretechnologyandpotentialcostsavings.Afterlisteningforabit,Frankpushedbackhischair,stoodup,andshouted,

“YoufuckingwanttoknowwhatIthinkaboutOpsware?Ithinkit’sthebiggestgoddamnpieceofshit!AllIhearaboutalldayishowmuchthisproductfuckingsucks.I’mgoingtodoeverythingIcantogetyouguysthrownoutofhere.”Frankrevealedhisplantoremoveallofoursoftwareimmediately,demanding

allfundstobereturned.Hewasdeadserious.Anthony remained calm, looked him in the eye, and said, “Frank, Iwill do

exactlyasyousay.I’veheardyouloudandclear.Thisisaterriblemomentforyouandforus.Allowmetouseyourphone,andIwillcallBenHorowitzandgive him your instructions. But before I do, can I ask you one thing? If mycompanymadethecommitment tofix theseissues,howmuchtimewouldyougiveustodothat?”He responded, “Sixty days.” Anthony told him the clock had just started

tickingandlefthisofficeimmediately.Itwasgoodnews:Wehadexactlysixtydays tofixall theproblemsandmake thedeploymentwork.Ifwedidnot,weweredone.Wehadsixtydaystolive.AnearlylessonIlearnedinmycareerwasthatwheneveralargeorganization

attemptstodoanything,italwayscomesdowntoasinglepersonwhocandelaythe entire project. An engineer might get stuck waiting for a decision or amanagermaythinkshedoesn’thaveauthoritytomakeacriticalpurchase.Thesesmall,seeminglyminorhesitationscancausefataldelays.Icouldnotaffordanyhesitation,soIscheduledadailymeetingwithAnthony,Jason,andtheteam—though they were now based in Plano. The purpose was to remove all

roadblocks. If anyonewas stuck on anything for any reason, it could not lastmorethantwenty-fourhours—thetimebetweenmeetings.Meanwhile, Anthony worked furiously to find the exciting value we could

offerEDS.Westartedwithlittlethingsthatdidnotchangeourfate,butrevealedimportantclues.WeflewourmainEDSexecutive,Frank,outtomeetwithourtop engineers and architects. In booking the trip,Anthony reported that Frankrequestedthelongestlayoverpossibleintheconnectingairport.IthoughtthatImisheardhim.“What,hewantsalonglayover?”Anthony:“Yep.”Ben:“Whywouldanybodywantalonglayoverinanairport?”Anthony:“Apparently,helikestohangoutintheairportbarbetweenflights.”Ben:“Whydoesheliketodothat?”Anthony:“Iaskedhimthesamequestion.Franksaid:‘BecauseIhatemyjob

andIhatemyfamily.’”Wow. Ihadno ideawho Iwasdealingwithuntil thatpoint.Understanding

how differently Frank viewed the world than the people at Opsware helpedclarify my thoughts. Frank expected to get screwed by us. It’s what alwayshappened to him in his job and presumably in his personal life. We neededsomethingdramatic tobreakhispsychology.Weneeded tobeassociatedwiththeairportbar,notwithhisjoborhisfamily.At the same time, Jason marched the team through the deployment with

unrelentingprecision.Amonth into theplan, theSouthwestAirlinescrew thatworkedtheSanJose–DallasflightknewJasonandhisteambyname.Theymadesteadyprogress,butitwasn’tgoingtobeenough.WewouldnotgetEDSfullydeployed in sixtydays—sonowwe really neededAnthony to deliver excitingvalue.AsIsat inmyofficehopingforabreakthrough,mycellphonerang. Itwas

Anthony.Anthony:“Ben,IthinkI’vegotit.”Ben:“Gotwhat?”Anthony:“TheexcitingvalueisTangram.”Ben:“What?”Anthony:“Tangram.EDSusesaproductfromacompanycalledTangramthat

inventories their hardware and software. Frank absolutely loves it, but thepurchasing guys are going to force him to switch to an equivalent ComputerAssociatesproduct,becauseit’sfreeaspartofEDS’ssettlementwithCA.FrankhatestheCAproduct.Frankisgettingscrewedagain.”

Ben:“Sowhatcanwedo?”Anthony:“IfTangramcancomefreewithOpsware,thenFrankwillloveus.”Ben: “That sounds economically impossible. If we buy the licenses from

TangramandgivethemtoEDS,thatwillbeacolossalexpense.We’llneverbeabletodescribeittoWallStreet.”Anthony: “You asked me what EDS really wanted. They really want

Tangram.”Ben:“Gotit.”IhadneverheardofTangram,soIquicklylookedthemup.Theywereasmall

company inCary,NorthCarolina, but they traded on theNASDAQmarket. Ilookeduptheirmarketcapitalization.Thiscouldn’tberight.TangramEnterpriseSolutions,accordingtoYahooFinance,wasworthonly$6million.Ihadneverheardofapubliccompanybeingthatcheap.I immediately calledmyheadofbusinessdevelopment, JohnO’Farrell, and

told him that Iwanted to buyTangram, and I needed the entire process to beextremelyquick—asin,IwantedtheTangramacquisitiondonebeforeoursixty-daywindowwithEDSclosed.TangramwasrunbyNormPhelps,an interimCEO,whichwasagreatsign

that they’d bewilling to sell the company, becausemost boardswouldmuchrathersellacompanythanrollthedicebyhiringanewCEO.JohngotintouchwithTangramandtheywereimmediatelyinterested,soweassembledateamtoconduct due diligencewhilewe negotiated amerger agreement in parallel.Attheendofduediligence,Ibroughtmyteambacktogether.TheypromptlyandunanimouslyagreedthatbuyingTangramwouldbeabadidea:Thetechnologywouldbedifficulttointegrateandnotthatvaluable.ThecompanywasinNorthCarolina.Itwasfifteenyearsoldandthetechnologywasold, too.Thefinanceteamthoughttheacquisitionwasamoneyloser.Ilistened,andthenItoldthemallthatIdidn’tcareaboutanyofthat.WeweregoingtobuyTangram.Theteamseemedshocked,butdidnotarguewithme.JohnandInegotiatedadealtobuyTangramfor$10millionincashandstock.

Wesigned thedealprior to theendof the sixty-dayplan. I calledFrank fromEDStotellhimthatoncethetransactionclosed,wewouldincludeallTangramsoftwareforfreeaspartofhisOpswarecontract.Frankwasecstatic.NowthatwehadsolvedFrank’sTangramproblem,heviewedtheworkthatJason’steamcompletedinatotallydifferentlight.Attheendofthesixtydays,Frankgatheredourteamandmadethefollowingspeech:“I’vegiventhespeechthatIgavetoyouguysatthebeginningofthisprocess

to at least a dozen other vendors. They all promised things, but none everdelivered.YouguysreallydeliveredandIamshocked.YouarethebestvendorthatIhaveandIamhappytobeworkingwithyou.”We’ddone it.Wesaved theaccountandsaved thecompany.Whata relief!

Butwestillhadthesmallmatterofthecompanythatwe’djustpurchasedanditsfifty-sevenemployees.Somedecisionsweresimple—wedidn’tneednineoutofthe ten salespeople, because they weren’t selling anything. Some were morecomplex:ShouldwekeeptheNorthCarolinalocation?Intheend,wedecidedtokeepitandlocatecustomersupportthere.Itturnedoutthatwhenyouaccountedforturnoverratesandthecostofrecruitingandtraining,Cary,NorthCarolina,engineerswere cheaper to hire thanBangalore, India, engineers.As the yearswentby,Tangramprovedtobeahighlyprofitableacquisition—wellbeyondthecriticalroleitplayedinsavingtheEDSaccount.During acquisition talks, both sides had agreed that Tangram’s CFO, John

Nelli,wouldnotbecomepartofOpsware.Butduringthetimebetweensigningandclose, Johnbegan toget severeheadaches.Hisdoctorsdiscovered that hehadbraincancer.BecausehewouldnotbeanOpswareemployeeanditwasapreexisting condition, hewould not be eligible for health insurance under ourplan.Thecostof the treatmentwithouthealth insurancewould likelybankrupthisfamily.IaskedmyheadofHRwhatitwouldcosttokeephimonthepayrolllong enough to qualify for COBRA and what COBRAwould cost. It wasn’tcheap—about$200,000.Thiswasasignificantamountofmoneyforacompanyinoursituation.Ontopofthat,webarelyknewJohnandtechnicallywedidn’t“owe”himanything.Thiswasn’tourproblem.Wewerefightingforourlives.Wewerefightingforourlives,buthewasabouttolosehis.Idecidedtopay

forhishealthcostsandfindthemoneyelsewhereinthebudget.Ineverexpectedtohear anything else about that decision, but fifteenmonths later I received ahandwritten letter from John’s wife letting me know that John had died. ShewrotethatshewasabsolutelyshockedthatIwouldhelpatotalstrangerandhisfamily and that I had saved her from total despair. She went on for severalparagraphs saying that she didn’t know why I did it, but it enabled her tocontinuelivingandshewaseternallygrateful.IguessIdiditbecauseIknewwhatdesperationfeltlike.

SURVIVALOFTHEFITTESTAlmostassoonastheEDScrisiswasresolved,Igotnewsthatthreenewclientswehadexpected tosignwerenowfadingaway.Anexcellentnewcompetitor,

BladeLogic, had arisen and was beating us in key accounts. We lost severaldeals to them and missed our quarterly numbers as a result. The stock pricedroppedbackdownto$2.90.Herewegoagain.Witha losingproduct,adwindlingstockprice,anda tired team,Iknewwe

were in trouble. To make matters worse, Marc, who had been workingexclusivelywithmeonLoudcloudandOpsware as “full-timechairmanof theboard,”haddecidedtofoundanothercompany,Ning.ThesuccessorfailureofOpswarewasreallyuptotheteamandmeatthispoint,butthetimingsucked.Not onlywas the company circling the drain, but ourmost visible spokesmanwasgoingtoworkonsomethingelse.Damn.Afterallthatwehadbeenthrough,howcouldIask the teamtochargeupyetanother impossiblemountain?HowcouldImusterthestrengthtodoitmyself?IfeltlikeIhadnomorestories,nomorespeeches,andnomore“rah-rah”in

me. I decided to level with the team and see what happened. I called an allengineeringmeetingandgavethefollowingspeech:“Ihavesomebadnews.WearegettingourasseskickedbyBladeLogicand

it’saproductproblem.Ifthiscontinues,Iamgoingtohavetosellthecompanyfor cheap. There is no way for us to survive if we don’t have the winningproduct.So,Iamgoingtoneedeveryoneofyoutodosomething.Ineedyoutogo home tonight and have a serious conversation with your wife, husband,significantother,orwhoevercaresmostaboutyouandtellthem,‘Benneedsmeforthenextsixmonths.’Ineedyoutocomeinearlyandstaylate.Iwillbuyyoudinner,andIwillstayherewithyou.Makenomistake,wehaveonebulletleftinthegunandwemusthitthetarget.”Atthetime,Ifelthorribleaskingtheteamtomakeyetanotherbigsacrifice.

Amazingly,IfoundoutwhilewritingthisbookthatIprobablyshouldhavefeltgoodaboutit.Here’swhatTedCrossman,oneofmybestengineers,saidaboutthattimeandthelaunchoftheaptlynamedDarwinProjectmanyyearslater:

OfallthetimesIthinkofatLoudcloudandOpsware,theDarwinProjectwasthemostfunandthemosthard.Iworkedsevendaysaweek8a.m.–10p.m.forsixmonthsstraight.Itwasfullon.OnceaweekIhadadatenightwith my wife where I gave her my undivided attention from 6 p.m. untilmidnight.Andthenextday,evenifitwasSaturday,I’dbebackintheofficeat8a.m.andstaythroughdinner.Iwouldcomehomebetween10–11p.m.Everynight.Anditwasn’tjustme.Itwaseverybodyintheoffice.

Thetechnicalthingsaskedofusweregreat.Wehadtobrainstormhowtodothingsandtranslatethosethingsintoanactualproduct.Itwashard,butfun.Idon’trememberlosinganyoneduringthattime.It

waslike,“Hey,wegottagetthisdone,orwewillnotbehere,we’llhavetoget another job.” It was a tight-knit group of people. A lot of the reallyjuniorpeoplereallysteppedup.Itwasagreatgrowingexperienceforthemtobethrownintothemiddleoftheoceanandtold,“Okay,swim.”Six months later we suddenly started winning proofs of concepts we

hadn’tbefore.Bendidagreatjob,he’dgiveusfeedback,andpatpeopleonthebackwhenweweredone.

Eightyearslater,whenIreadwhatTedhadwritten,Icried.IcriedbecauseIdidn’tknow.IthoughtIdid,butIreallydidn’t.IthoughtthatIwasaskingtoomuchofeverybody.IthoughtthatafterbarelysurvivingLoudcloud,nobodywasreadyforanotherdo-or-diemission.IwishIknewthenwhatIknownow.Afterthespeechcamethehardworkofdefiningtheproduct.Theproductplan

wasweigheddownwithhundredsofrequirementsfromourexistingcustomers.Theproductmanagementteamhadanallergicreactiontoprioritizingpotentiallygood features above features thatmight hypothetically beat BladeLogic. Theywouldsay,“Howcanwewalkawayfromrequirementsthatweknowtobetruetopursuesomethingthatwethinkwillhelp?”Itturnsoutthatisexactlywhatproductstrategyisallabout—figuringoutthe

rightproduct is the innovator’s job,not thecustomer’s job.Thecustomeronlyknows what she thinks she wants based on her experience with the currentproduct. The innovator can take into account everything that’s possible, butoftenmustgoagainstwhatsheknowstobetrue.Asaresult,innovationrequiresacombinationofknowledge,skill,andcourage.Sometimesonlythefounderhasthecouragetoignorethedata;wewererunningoutoftime,soIhadtostepin:“Idon’tcareaboutanyoftheexistingrequirements;Ineedyoutoreinventthe

product and we need to win.” Nine months later, when we released our newproduct we could now win any deal. Armed with the new product, MarkCranney,headofsales,wenttowar.After assembling a top-end sales force, he completely revamped the sales

processandsenteverysalesperson througharigorousandunforgiving trainingprogram.Hedemandedmastery.Anyslip-up in technique, skill,orknowledgewouldbemetwithtotalintolerancefromMark.WeheldaweeklyforecastcallwhereMarkreviewedeverydealinfrontofthe

entire 150 person sales force. On one such call, a salesperson described anaccountthathe’dforecastindetail:“Ihavebuy-infrommychampion,thevicepresident thathe reports to, and theheadofpurchasing.Mychampionassuresmethatthey’llbeabletocompletethedealbytheendofthefiscalquarter.”Mark quickly replied, “Have you spoken to the vice president’s peer in the

networkinggroup?”Salesrep:“Um,noIhaven’t.”Mark:“Haveyouspokentothevicepresidentyourself?”Salesrep:“No.”Mark:“Okay,listencarefully.Here’swhatI’dlikeyoutodo.First,reachup

toyourfaceandtakeoffyourrose-coloredglasses.ThengetaQ-tipandcleanthewaxoutofyourears.Finally,takeoffyourpinkpantiesandcallthefuckingvicepresidentrightnow,becauseyoudonothaveadeal.”Mark was right. It turned out that we did not have a deal, as the vice

president’s peer in networking was blocking it.We eventually got a meetingwith him and won the deal. More important, Mark set the tone: Sloppinesswouldnotbetolerated.Nowthatwe’dimprovedourcompetitiveposition,wewentontheoffensive.

Inmyweeklystaffmeeting,Iinsertedanagendaitemtitled“WhatAreWeNotDoing?” Ordinarily in a staff meeting, you spend lots of time reviewing,evaluating, and improving all of the things that you do: build products, sellproducts,supportcustomers,hireemployees,andthelike.Sometimes,however,thethingsyou’renotdoingarethethingsyoushouldactuallybefocusedon.In one such meeting, after asking the question, every person on my staff

agreed:“Wearenotautomatingthenetwork.”AlthoughtheoriginalversionofOpsware thatweused inLoudcloud automatedournetwork, the softwarewasnot robust and, of course, featured the purple-pimp-hat user interface. As aresult,whenweswitchedover tobeinga softwarecompany,wenarrowedourfocustoserverautomationandneverrevisitedthedecision.ThisworkedwellforthefirstseveralyearsofOpsware,butnowwehadanopportunitytobringbackournetworkautomationproduct.Unfortunately,theJivewasnotagoodcodebaseandcouldnotbeturnedinto

acommercialproduct.Mychoiceswere:(a)startanewprojector(b)buyoneofthe four existing network automation companies. Early in my career as anengineer,I’dlearnedthatalldecisionswereobjectiveuntilthefirstlineofcodewas written. After that, all decisions were emotional. In addition, I had JohnO’Farrell, the industry’sgreatestM&Anegotiator,onmyteamsoIdecided to

investigatetheothercompaniesbeforesizingtheinternaleffort.Surprisingly, among the four existing network automation players, the

companythatwethoughthadthebestproductarchitecture,RenditionNetworks,hadthelowestrevenues.Thismadesomeofourbusinesspeopleskepticalofourtechnicalevaluation.However, if I’d learnedanything itwas thatconventionalwisdomhadnothingtodowiththetruthandtheefficientmarkethypothesiswasdeceptive.HowelsecouldoneexplainOpsware tradingathalfof thecashwehad in the bank when we had a $20 million a year contract and fifty of thesmartestengineers in theworld?No,marketsweren’t“efficient”at finding thetruth; they were just very efficient at converging on a conclusion—often thewrongconclusion.Afterconfirmingthatacquiringwouldbesuperiortobuilding,wenegotiateda

deal to buy Rendition Networks for $33 million. Within three months ofcompleting the acquisition, John negotiated a deal with Cisco Systems—theworld’s largestnetworkingcompany—to resell ourproduct.Thedeal includedan agreement to prepay us $30million for advanced licenses.As a result, theCiscodealalonepaidmorethan90percentoftheacquisitioncosts.Notetoself:It’sagoodideatoask,“WhatamInotdoing?”

THEULTIMATEDECISIONAswefieldedthebroaderproductline,ourmomentumsteadilygrew.Fromtheashes,we’dbuiltasoftwarebusinessthatapproacheda$150millionrevenuerunrate.Alongwith our revenue, our stock price rose from its floor of $0.35 pershare aswell aswe tradedbetween$6per share and$8per share, sometimestradingatamarketcapitalizationofmorethan$800million.Still,everythingwasnot rosy.Everyquarterwas tough,and thecompetitive

and the technology landscapes changed rapidly. A technology calledvirtualizationwas taking themarketbystormandchangingthewaycustomersthought about automating their environments. In fact, it looked to me likevirtualizationmight be the technological breakthrough that finally enabled thecloudcomputingbusinessmodeltowork.Beyondthat,beingapubliccompanywas still never going to get easy. At one point, a shareholder activist namedRachelHymandecidedthatmyegowasoutofcontrol,andshedemandedthattheboard removemeand sell the company immediately.Thiswasdespite thefactthatweweretradingat$7pershare,whichwastentimestheoriginalpriceofhershares.Nonetheless, Iwas not looking for the exits.Whenever a potential acquirer

approached us, I would always reply, “We are not for sale.” It was a greatanswerinthatIwasn’treadytosellanditconveyedthat,butitalsoleftthedooropen to a particularly aggressive buyer. “Not for sale” didn’t mean that wewouldn’t listen to offers—it just meant that we weren’t trying to sell thecompany.So,whenEMCimpliedthatitwantedtobuyus,Ithoughtnothingofit.Wewere trading at about $6.50 per share and I wasn’t planning to sell atanything close to that price.But this time the news of the offer leaked to thepressandthestockshotupto$9.50pershare,changingtheeconomicequation,especiallysincethestockwasgoingupforallthewrongreasons.Ironically,thehigherthestockpricewentup,themorecompanieswantedto

buyus.Overthecourseofthenextmonth,elevencompaniesexpressedinterest.Given the uncertainty in the business and the implied earningsmultiple, theirinterestwastoomuchtoignore.Togetthingsstarted,JohnandIcalledMichaelOvitztogetsomeadvice.We

felt oneof thepotential bidders,Oracle,wouldbe the least likely tobidhigh,becauseitwasextremelydisciplinedinitsfinancialanalysis.WeconveyedthistoMichaelandquestionedwhetherweshouldpursueOracleatall.Hisreplywaspriceless:“Well,boys,ifyouaregoingtohaveadograce,thenyouaregoingtoneedarabbit.AndOraclewillbeonehellofarabbit.”Withthatstrategyinhand,wegeneratedabroadsetofbids,allbetween$10

and$11pershare,withthehighestbidsrepresentinga38percentpremiumoverthecurrentstockprice.Althoughthiswasconsideredagoodpremium,Ididnotfeelrightsellingthecompanyfor$11pershare.Theteamhadworkedtoohard,we’daccomplished toomuch, andwewere toogooda company.The risksofstaying stand-alone were substantial, but I still wanted to bet on the team. Irecommendedtotheboardthatwenotsell.The board was surprised, but supportive. Still, they had a fiduciary

responsibilitytoshareholderstoaskthetoughquestions.“Ifyou’reunwillingtosellatelevendollarspershare,isthereapriceatwhichyouwouldsell?”Ihadtothinkaboutthatone.Ihadpromisedtheteamthatifwegottobethenumber-onecompanyinabigmarket,wewouldnotsell.Wewerenumberone,buthowbigwas themarket?Did the team reallywant to continue orwas it justmewhowanted to continue?Howcould I knowwithout panicking the company?Andthusbeganaseriesofverylongtalkswithmyself.Itwasanargumenttothedeath,anditwasmeagainstme.Ontheonehand,I

argued that virtualization created an explosion of virtual server instances,makingwhatwedidmoreessentialthanever.Inthenextbreath,Iretortedthat

whilethatmayhavebeentrue,thearchitecturalchangeswouldmakeourmarketposition vulnerable. I battled myself for weeks before concluding that thingswerechangingfastenoughthatwe’dneedtomakemajorchangestoourproductarchitectureinordertostayontop.Thekeytoansweringtheultimatequestionwasknowingthestateoftheteam.Weretheyupforyetanothergiantchallengeorweretheyattheendofaverylongroad?Idecidedtobringmydirectreportsinto the loop and ask themwhat they thought. The answers came back clear:Everyone,withtheexceptionofonepersonwhofeltthattheopportunityinfrontofuswasstillquitelarge,optedforthesale.Nowitwasjustamatterofprice.Butwhatprice?Aftera longdiscussionwithJohnO’Farrell, Idecided that therightprice to

sell the company would be $14 per share, or about $1.6 billion. I took thatnumberbacktotheboard.Theythoughtthenumberwasextremelyhighandthatit was unlikely we’d be able to generate a bid at that level, but they weresupportive nonetheless. I called back all the potential acquirers and let themknowthatwewouldonlyentertainbidsof$14ormore.Therewerenotakers.Morethanamonthpassedwithoutaword,andIfiguredtheM&Atalkshad

ended. I began refocusing on how tomake the necessary changes to keep uscompetitive.AndthenIreceivedacallfromBobBeauchamp,theCEOofBMCSoftware.Heoffered$13.25per share. I held firm: “Bob, that’sgreat, but thenumberisfourteendollarspershare.”Bobsaidthathe’dhavetothinkaboutit.He called back two days later and offered $14 per share.Wow. The dog hadcaughtthebus.JohnandIimmediatelycalledbackalltheothersuitorstoletthemknowthat

wehadanofferthatweplannedtotake.Hewlett-Packardwasstillinterestedandoffered $13.50 per share in an effort to make sure that I wasn’t bluffing. Iresponded that as a public company CEO, I couldn’t take a lower offer. HPeventuallyoffered$14.25or$1.65billionincash.Wehadadeal.Whenitfinallyended—thelongroadfromLoudcloudtoOpsware—Icouldn’t

believe that I’d soldwhat it tookeightyears andall ofmy life force tobuild.Howcould Ihavedone that? Iwas sick. I couldn’t sleep, Ihadcold sweats, Ithrewup,andIcried.AndthenIrealizedthatitwasthesmartestthingthatI’deverdone inmycareer.We’dbuiltsomethingfromnothing,sawitgoback tonothingagain,andthenrebuiltitintoa$1.65billionfranchise.Atthatpoint,itfeltlikemybusinesslifewaskindofover.Ihadhiredallthe

bestpeople that Ikneworcouldfind,andIhadgone througheverystepfromfoundingtogoingpublictosale.Idefinitelydidnotfeellikedoinganyofthat

again.ButIhadlearnedsomuch.Itseemedlikesuchawastetodosomethingcompletely different. And then I got an idea to build a new kind of venturecapitalfirm.Wewillexplorethisideainchapter9,butfirst,chapters4through8willtake

youthroughmosteverythingI learnedtothispointplusafewnewwarstoriesfrommyexperiencesrunningLoudcloudandOpsware.

—CHAPTER4—

WHENTHINGSFALLAPART

“Thereare severaldifferent frameworksonecoulduse toget ahandleontheindeterminatevs.determinatequestion.Themathversioniscalculusvs.statistics. In a determinate world, calculus dominates. You can calculatespecific thingspreciselyanddeterministically.Whenyousenda rocket tothemoon,youhave tocalculatepreciselywhere it isatall times. It’snotlikesomeiterativestartupwhereyoulaunchtherocketandfigurethingsoutstepbystep.Doyoumakeittothemoon?ToJupiter?Doyoujustgetlostinspace?Therewerelotsofcompaniesinthe’90sthathadlaunchpartiesbutnolandingparties.“But the indeterminate future issomehowone inwhichprobabilityand

statistics are the dominant modality for making sense of the world. Bellcurvesandrandomwalksdefinewhatthefutureisgoingtolooklike.Thestandard pedagogical argument is that high schools should get rid ofcalculusandreplaceitwithstatistics,whichisreallyimportantandactuallyuseful.Therehasbeenapowerfulshifttowardtheideathatstatisticalwaysofthinkingaregoingtodrivethefuture.”

—PETERTHIEL

When I was attempting to sell the cloud computing services part of theLoudcloud business, Imetwith Bill Campbell to update him onwhere Iwaswith the deal. The deal was critical, because without it, the company wouldalmostcertainlygobankrupt.After Icarefullybriefedhimonwherewewerewithboth interestedparties,

IBMandEDS,Billpaused foramoment.He lookedme in theeyesandsaid,“Ben,youneedtodosomethinginadditiontoworkingonthisdeal.Youneedtodo it alone with your general counsel. You need to prepare the company for

bankruptcy.”Toanobjectiveobserver,thismightsoundlikeBillwasprudentlyadvisingme tobuildmycontingencyplan.Butsomething inhisvoiceandhiseyessaidsomethingdifferent.Theysaid thathebelieved thecontingencyplanwasgoingtobetheplan.The conversation brought to mind a story that a friend told me about his

brother, a young doctor. A thirty-five-year-old man came to see my friend’sbrother.Theman lookedawful.Hiseyeswerehollowandhisskinwasashen.Theyoungdoctorknewsomethingwaswrong,buthecouldnotfigureoutwhat,so he brought in an elder colleague to help with the diagnosis. The moreexperienced doctor examined theman and then sent him on hisway. The olddoctorthenturnedtotheyoungdoctorandsaid,“He’sdead.”Theyoungdoctorwas flabbergasted: “What are you talking about? He just walked out of herealive!”Theolderdoctor replied, “Hedoesn’t know it yet, but he’sdead.He’shadaheartattackandwhenpeoplethatyounghaveheartattackstheirbodiesarenotyetpliableenough to recover.Hewon’t recover.He’sdead.”Threeweekslaterthepatientdied.IfeltthatBillwastellingme,althoughIwaswalkingaroundtryingtogetthe

dealdone,thatIwasalreadydeadandthatIdidnotknowit.Itwasaveryhardthingforhimtosayandonlythebestoffriendswillmusterthecouragetobreaknewsthathorrible.Itwasanevenharderthingformetohear.HetoldmesothatIcouldemotionallypreparemyselfandfinanciallypreparethecompanyfortheinevitable funeral. The odds of landing a company-saving deal during thetechnology industry’s nuclear winter were close to nil. Chances were, I wasdead.I never built that contingency plan. Through the seemingly impossible

LoudcloudseriesCandIPOprocesses, I learnedone important lesson:StartupCEOs shouldnot play theodds.Whenyou arebuilding a company, youmustbelievethereisananswerandyoucannotpayattentiontoyouroddsoffindingit.Youjusthavetofindit.Itmattersnotwhetheryourchancesarenineintenoroneinathousand;yourtaskisthesame.In theend, Idid find theanswer,wecompleted thedealwithEDS,and the

company did not go bankrupt. Iwas notmad at Bill. To this day, I sincerelyappreciatehistellingmethetruthabouttheodds.ButIdon’tbelieveinstatistics.Ibelieveincalculus.Peoplealwaysaskme,“What’s thesecret tobeingasuccessfulCEO?”Sadly,thereisnosecret,butifthereisoneskillthatstandsout,it’stheabilitytofocus

andmakethebestmovewhentherearenogoodmoves.It’sthemomentswhereyoufeelmostlikehidingordyingthatyoucanmakethebiggestdifferenceasaCEO.Intherestofthischapter,Ioffersomelessonsonhowtomakeitthroughthestrugglewithoutquittingorthrowinguptoomuch.Whilemostmanagementbooks focusonhow todo thingscorrectly, soyou

don’t screwup, these lessonsprovide insight intowhatyoumustdo after youhavescrewedup.Thegoodnewsis, Ihaveplentyofexperienceat thatandsodoeseveryotherCEO.Iputthissectionfirsteventhoughitdealswithsomeseriousendgameissues

suchashowtofireanexecutiveandhowtolaypeopleoff.Indoingso,IfollowthefirstprincipleoftheBushido—thewayofthewarrior:keepdeathinmindatalltimes.Ifawarriorkeepsdeathinmindatalltimesandlivesasthougheachday might be his last, he will conduct himself properly in all his actions.Similarly, ifaCEOkeeps thefollowing lessons inmind,shewillmaintain theproperfocuswhenhiring,training,andbuildingherculture.

THESTRUGGLE

Everyentrepreneurstartshercompanywithaclearvisionforsuccess.Youwillcreate an amazing environment and hire the smartest people to join you.Togetheryouwillbuildabeautifulproduct thatdelights customersandmakestheworldjustalittlebitbetter.It’sgoingtobeabsolutelyawesome.Then,afterworkingnightanddaytomakeyourvisionareality,youwakeup

tofindthatthingsdidnotgoasplanned.YourcompanydidnotunfoldliketheJackDorsey keynote that you listened towhen you started.Your product hasissuesthatwillbeveryhardtofix.Themarketisn’tquitewhereitwassupposedtobe.Youremployeesarelosingconfidenceandsomeofthemhavequit.Someoftheoneswhoquitwerequitesmartandhavetheremainingoneswonderingifstayingmakes sense.Youare running lowoncash andyourventure capitalisttellsyou that itwillbedifficult to raisemoneygiven the impendingEuropeaneconomiccatastrophe.Youloseacompetitivebattle.Youlosealoyalcustomer.Youloseagreatemployee.Thewallsstartclosingin.Wheredidyougowrong?Whydidn’tyourcompanyperformasenvisioned?Areyougoodenoughtodothis?Asyourdreamsturnintonightmares,youfindyourselfintheStruggle.

ABOUTTHESTRUGGLE

“Lifeisstruggle.”

—KARLMARX

The Struggle is when you wonder why you started the company in the firstplace.TheStruggleiswhenpeopleaskyouwhyyoudon’tquitandyoudon’tknow

theanswer.TheStruggleiswhenyouremployeesthinkyouarelyingandyouthinkthey

mayberight.TheStruggleiswhenfoodlosesitstaste.TheStruggleiswhenyoudon’tbelieveyoushouldbeCEOofyourcompany.

TheStruggle iswhenyouknow thatyouare inoveryourheadandyouknowthatyoucannotbereplaced.TheStruggleiswheneverybodythinksyouareanidiot,butnobodywillfireyou.TheStruggleiswhereself-doubtbecomesself-hatred.TheStruggle iswhenyouarehaving a conversationwith someone andyou

can’thearawordthattheyaresayingbecauseallyoucanhearistheStruggle.TheStruggleiswhenyouwantthepaintostop.TheStruggleisunhappiness.TheStruggleiswhenyougoonvacationtofeelbetterandyoufeelworse.TheStruggleiswhenyouaresurroundedbypeopleandyouareallalone.The

Strugglehasnomercy.TheStruggleisthelandofbrokenpromisesandcrusheddreams.TheStruggle

isacoldsweat.TheStruggleiswhereyourgutsboilsomuchthatyoufeellikeyouaregoingtospitblood.TheStruggle isnot failure,but itcauses failure.Especially ifyouareweak.

Alwaysifyouareweak.Mostpeoplearenotstrongenough.Everygreatentrepreneur fromSteveJobs toMarkZuckerbergwent through

theStruggleandstruggletheydid,soyouarenotalone.Butthatdoesnotmeanthatyouwillmakeit.Youmaynotmakeit.ThatiswhyitistheStruggle.TheStruggleiswheregreatnesscomesfrom.

SOMESTUFFTHATMAYORMAYNOTHELP

ThereisnoanswertotheStruggle,butherearesomethingsthathelpedme:

Don’tputitallonyourshoulders.Itiseasytothinkthatthethingsthatbotheryouwillupsetyourpeoplemore.That’snottrue.Theoppositeistrue.Nobodytakesthelossesharderthanthepersonmostresponsible.Nobodyfeelsitmorethanyou.Youwon’tbeabletoshareeveryburden,butshareeveryburdenthatyoucan.Getthemaximumnumberofbrainsontheproblemseveniftheproblemsrepresentexistentialthreats.WhenIranOpswareandwewerelosingtoomanycompetitivedeals,Icalledanallhandsandtoldthewholecompanythatweweregettingourasseskicked,andifwedidn’tstopthebleeding,weweregoingtodie.Nobodyblinked.Theteamrallied,builtawinningproduct,andsavedmysorryass.Thisisnotcheckers;thisismotherfuckin’chess.Technologybusinessestendtobeextremelycomplex.Theunderlyingtechnologymoves,the

competitionmoves,themarketmoves,thepeoplemove.Asaresult,likeplayingthree-dimensionalchessonStarTrek,thereisalwaysamove.Youthinkyouhavenomoves?Howabouttakingyourcompanypublicwith$2millionintrailingrevenueand340employees,withaplantodo$75millioninrevenuethenextyear?Imadethatmove.Imadeitin2001,widelyregardedastheworsttimeeverforatechnologycompanytogopublic.Imadeitwithsixweeksofcashleft.Thereisalwaysamove.Playlongenoughandyoumightgetlucky.Inthetechnologygame,tomorrowlooksnothingliketoday.Ifyousurvivelongenoughtoseetomorrow,itmaybringyoutheanswerthatseemssoimpossibletoday.Don’ttakeitpersonally.Thepredicamentthatyouareinisprobablyallyourfault.Youhiredthepeople.Youmadethedecisions.Butyouknewthejobwasdangerouswhenyoutookit.Everybodymakesmistakes.EveryCEOmakesthousandsofmistakes.EvaluatingyourselfandgivingyourselfanFdoesn’thelp.Rememberthatthisiswhatseparatesthewomenfromthegirls.Ifyouwanttobegreat,thisisthechallenge.Ifyoudon’twanttobegreat,thenyounevershouldhavestartedacompany.

THEEND

WhenyouareintheStruggle,nothingiseasyandnothingfeelsright.Youhavedropped into theabyssandyoumaynevergetout. Inmyownexperience,butforsomeunexpectedluckandhelp,Iwouldhavebeenlost.Sotoallofyouinit,mayyoufindstrengthandmayyoufindpeace.

CEOSSHOULDTELLITLIKEITIS

One of the most important management lessons for a founder/CEO is totallyunintuitive.My single biggest personal improvement as CEO occurred on thedaywhenIstoppedbeingtoopositive.AsayoungCEO,Ifeltthepressure—thepressureofemployeesdependingon

me,thepressureofnotreallyknowingwhatIwasdoing,thepressureofbeingresponsible for tens of millions of dollars of other people’s money. As aconsequenceofthispressure,Itooklossesextremelyhard.Ifwefailedtowinacustomer or slipped a date or shipped a product that wasn’t quite right, itweighed heavily on me. I thought that I would make the problem worse bytransferring thatburden tomyemployees. Instead, I thought I shouldprojectapositive, sunny demeanor and rally the unburdened troops to victory. I wascompletelywrong.Irealizedmyerrorduringaconversationwithmybrotherin-law,Cartheu.At

thetime,CartheuworkedforAT&Tasatelephonelineman(heisoneofthoseguyswhoclimbthepoles).IhadjustmetaseniorexecutiveatAT&T,whomI’llcall Fred, and I was excited to find out if Cartheu knew him. Cartheu said,“Yeah,IknowFred.Hecomesbyaboutonceaquartertoblowalittlesunshineupmyass.”Atthatmoment,IknewthatI’dbeenscrewingupmycompanybybeingtoopositive.In my mind, I was keeping everyone in high spirits by accentuating the

positive and ignoring the negative. But my team knew that reality was morenuancedthanIwasdescribingit.Andnotonlydidtheyseeforthemselvestheworldwasn’tasrosyasIwasdescribingit;theystillhadtolistentomeblowingsunshineuptheirbuttsateverycompanymeeting.HowdidImakesuchamistakeandwhywasitsuchabigmistake?

THEPOSITIVITYDELUSION

As the highest-ranking person in the company, I thought that Iwould be bestabletohandlebadnews.Theoppositewastrue:Nobodytookbadnewsharderthan I did. Engineers easily brushed off things that kept me awake all night.

Afterall,IwasthefoundingCEO.Iwastheone“married”tothecompany.Ifthings went horribly wrong, they could walk away, but I could not. As aconsequence,theemployeeshandledlossesmuchbetter.Evenmore stupidly, I thought that itwasmy jobandmy jobonly toworry

aboutthecompany’sproblems.HadIbeenthinkingmoreclearly,Iwouldhaverealizedthatitdidn’tmakesenseformetobetheonlyonetoworryabout,forexample, theproductnotbeingquite right—because Iwasn’twriting the codethatwouldfixit.Amuchbetter ideawouldhavebeen togive theproblemto thepeoplewho

couldnotonlyfixit,butwhowouldalsobepersonallyexcitedandmotivatedtodoso.Anotherexample:Ifwelostabigprospect,thewholeorganizationneededtounderstandwhy,sothatwecouldtogetherfixthethingsthatwerebrokeninourproducts,marketing,andsalesprocess.IfIinsistedonkeepingthesetbackstomyself,therewasnowaytojump-startthatprocess.

WHYIT’SIMPERATIVETOTELLITLIKEITIS

There are three key reasons why being transparent about your company’sproblemsmakessense:

1.Trust.Withouttrust,communicationbreaks.Morespecifically:Inanyhumaninteraction,therequiredamountofcommunicationisinversely

proportionaltotheleveloftrust.Consider the following: If I trust you completely, then I require no

explanationorcommunicationofyouractionswhatsoever,becauseIknowthatwhateveryouaredoingisinmybestinterests.Ontheotherhand,ifIdon’ttrustyou at all, then no amount of talking, explaining, or reasoning will have anyeffectonme,becauseIdonottrustthatyouaretellingmethetruth.In a company context, this is a critical point. As a company grows,

communicationbecomes itsbiggest challenge. If the employees fundamentallytrust the CEO, then communication will be vastly more efficient than if theydon’t.Tellingthingsastheyareisacriticalpartofbuildingthistrust.ACEO’sability to build this trust over time is often the difference between companiesthatexecutewellandcompaniesthatarechaotic.

2.Themorebrainsworkingonthehardproblems,thebetter.

Inordertobuildagreattechnologycompany,youhavetohirelotsofincrediblysmartpeople.It’satotalwastetohavelotsofbigbrainsbutnotletthemworkonyour biggest problems.Abrain, nomatter howbig, cannot solve a problem itdoesn’t know about.As the open-source communitywould explain it, “Givenenougheyeballs,allbugsareshallow.”

3.AgoodcultureisliketheoldRIProutingprotocol:Badnewstravelsfast;goodnewstravelsslow.Ifyouinvestigatecompaniesthathavefailed,youwillfindthatmanyemployeesknewabout the fatal issues longbefore those issueskilled thecompany. If theemployees knew about the deadly problems, why didn’t they say something?Toooftentheansweristhatthecompanyculturediscouragedthespreadofbadnews,sotheknowledgelaydormantuntilitwastoolatetoact.Ahealthycompanycultureencouragespeopletosharebadnews.Acompany

that discusses its problems freely and openly can quickly solve them. Acompanythatcoversupitsproblemsfrustrateseveryoneinvolved.Theresultingaction item forCEOs:Build a culture that rewards—notpunishes—people forgettingproblemsintotheopenwheretheycanbesolved.As a corollary, beware of management maxims that stop information from

flowing freely in your company. For example, consider the old managementstandard:“Don’tbringmeaproblemwithoutbringingmeasolution.”What ifthe employee cannot solve an important problem? For example, what if anengineeridentifiesaseriousflawinthewaytheproductisbeingmarketed?Doyou reallywant him to bury that information?Management truisms like thesemaybegoodforemployeestoaspiretointheabstract,buttheycanalsobetheenemyoffree-flowinginformation—whichmaybecriticalforthehealthofthecompany.

FINALTHOUGHT

Ifyourunacompany,youwillexperienceoverwhelmingpsychologicalpressuretobeoverlypositive.Standuptothepressure,faceyourfear,andtellitlikeitis.

THERIGHTWAYTOLAYPEOPLEOFF

ShortlyafterwesoldOpswaretoHewlett-Packard,IhadaconversationwiththelegendaryventurecapitalistDougLeoneofSequoiaCapital.Hewantedme torecountthestoryofhowwewentfromdoomedintheeyesoftheworldtoa$1.6billionoutcomewithnorecapitalization.After I tookhim through thedetails—including severalnearbankruptcies, a

stock price of $0.35 per share, unlimited bad press, and three separate layoffswherewelosta totaloffourhundredemployees—hewasmostamazedbythelayoffs. During more than twenty years in the venture capital business, he’dnever seen a company recover fromconsecutive layoffs andachieve abillion-dollar-plusoutcome.Heconfessed that he’dbet against that every time.Sincemyonlyexperiencewasthegreatexception,Ineededmoreinformation.Iaskedhimwhyalltheotherstartupsfailed.Herepliedthatthelayoffsinevitablybrokethe company’s culture. After seeing their friends laid off, employeeswere nolongerwilling tomake the requisite sacrifices needed to build a company.Hesaid that although it was possible to survive an isolated layoff, it was hugelyunlikely that a company would experience great success. Building a highlyvaluablebusiness,he added, after three consecutivegiant layoffs accompaniedbyhorribleprominentpresscoverage (wegot takenapartwithcoverstories inboththeWallStreetJournalandBusinessWeek),wasacompleteviolationofthelaws of venture capital physics. He wanted to know how we did it. Afterthinkingaboutthisquestion,here’smyanswer.In retrospect, we were able to keep cultural continuity and retain our best

employeesdespitemultiplemassivelayoffsbecausewelaidpeopleofftherightway.Thismay sound nutty—how can you do something that’s fundamentallywrongin“therightway”?Here’show.

STEP1:GETYOURHEADRIGHT

Whenacompany fails tohit its financialplan so severely that itmust fire theemployees itwent to great time and expense to hire, itweighs heavily on thechief executive. During the first layoff at our company, I remember being

forwardedanemailexchangeamongagroupofemployees.Intheexchange,oneof our smarter employees wrote, “Ben is either lying or stupid or both.” Irememberreadingthatandthinking,“definitelystupid.”Duringatimelikethis,itisdifficulttofocusonthefuture,becausethepastoverwhelmsyou—butthat’sexactlywhatyoumustdo.

STEP2:DON’TDELAY

Onceyoudecidethatyouwillhavetolaypeopleoff,thetimeelapsedbetweenmakingthatdecisionandexecutingthatdecisionshouldbeasshortaspossible.Ifwordleaks(whichitwillinevitablyifyoudelay),thenyouwillbefacedwithanadditionalsetofissues.Employeeswillquestionmanagersandaskwhetheralayoff is coming. If the managers don’t know, they will look stupid. If themanagersdoknow,theywilleitherhavetolietotheiremployees,contributetothe leak, or remain silent, which will create additional agitation. AtLoudcloud/Opsware,webadlymismanagedthisdynamicwithourfirstroundoflayoffs,butsharplycorrectedthingsonthenexttwo.

STEP3:BECLEARINYOUROWNMINDABOUTWHYYOUARELAYINGPEOPLEOFF

Goingintoalayoff,boardmemberswillsometimestrytomakeyoufeelbetterby putting a positive spin on things. They might say, “This gives us a greatopportunity to dealwith some performance issues and simplify the business.”Thatmaybetrue,butdonotletthatcloudyourthinkingoryourmessagetothecompany.Youarelayingpeopleoffbecausethecompanyfailedtohititsplan.Ifindividual performance were the only issue, then you’d be taking a differentmeasure.Company performance failed.This distinction is critical, because themessagetothecompanyandthelaid-offindividualsshouldnotbe“Thisisgreat,we are cleaning up performance.”Themessagemust be “The company failedand in order to move forward, we will have to lose some excellent people.”Admittingtothefailuremaynotseemlikeabigdeal,buttrustme,itis.“Trustme.”That’swhataCEOsayseverydaytoheremployees.Trustme:Thiswillbeagoodcompany.Trustme:Thiswill begood foryour career.Trustme:Thiswillbegoodforyourlife.Alayoffbreaksthattrust.Inordertorebuildtrust,youhavetocomeclean.

STEP4:TRAINYOURMANAGERS

Themostimportantstepinthewholeexerciseistrainingthemanagementteam.If you sendmanagers into this super-uncomfortable situationwith no training,mostofthemwillfail.Training startswith a golden rule:Managersmust layoff their ownpeople.

TheycannotpassthetasktoHRortoamoresadisticpeer.Youcannothireanoutsourcingfirmlike theone in themovieUpin theAir.Everymanagermustlayoffhisownpeople.Whysostrict?Whycan’tthemoreconfrontationalmanagersjusthandlethis

taskforeveryone?Becausepeoplewon’tremembereverydaytheyworkedforyourcompany,but theywill surely remember thedayyou laid themoff.Theywillremembereverylastdetailaboutthatdayandthedetailswillmattergreatly.The reputations of your company and yourmanagers depend on you standingtall,facingtheemployeeswhotrustedyouandworkedhardforyou.IfyouhiredmeandIbustedmyassworkingforyou,Iexpectyoutohavethecouragetolaymeoffyourself.Onceyoumakeitclearthatmanagersmustlayofftheirownpeople,besure

topreparethemforthetask:

1.Theyshouldexplainbrieflywhathappenedandthatitisacompanyratherthanapersonalfailure.2.Theyshouldbeclearthattheemployeeisimpactedandthatthedecisionisnonnegotiable.3.Theyshouldbefullypreparedwithallofthedetailsaboutthebenefitsandsupportthecompanyplanstoprovide.

STEP5:ADDRESSTHEENTIRECOMPANY

Prior to executing the layoff, the CEOmust address the company. The CEOmustdelivertheoverallmessagethatprovidesthepropercontextandaircoverforthemanagers.Ifyoudoyourjobright,themanagerswillhaveamucheasiertimedoingtheirjobs.KeepinmindwhatformerIntuitCEOBillCampbelltoldme—Themessageis for thepeoplewhoarestaying.Thepeoplewhostaywillcaredeeplyabouthowyoutreattheircolleagues.Manyofthepeoplewhomyoulayoffwillhavecloserrelationshipswith thepeoplewhostay thanyoudo,sotreat themwith theappropriate levelof respect.Still, thecompanymustmoveforward,sobecarefulnottoapologizetoomuch.

STEP6:BEVISIBLE,BEPRESENT

Afteryoumakethespeechtellingyourcompanythatyouwillbelettingmanyofthem go, you will not feel like hanging out and talking to people. You willprobablyfeellikegoingtoabaranddrinkingafifthoftequila.Donotdothis.Bepresent.Bevisible.Beengaging.Peoplewanttoseeyou.Theywanttoseewhetheryoucare.Thepeoplewhomyoulaidoffwillwanttoknowiftheystillhavearelationshipwithyouandthecompany.Talktopeople.Helpthemcarrytheirthingstotheircars.Letthemknowthatyouappreciatetheirefforts.

PREPARINGTOFIREANEXECUTIVE

Whenyourecruitanexecutive,youpaintabeautifulpictureofherfutureinyourcompany.Youdescribeingreatdepthandinvibrantcolorhowawesomeitwillbeforhertoacceptyourofferandhowmuchbetteritwillbethanjoiningthatothercompany.Thenonedayyourealizeyoumustfireher.Reconcilethat,Ms.CEO.It turns out that the actual act of firing an executive can be relatively easy

comparedwithanyotherfiring.Executiveshaveexperiencebeingontheothersideof the conversation and tend tobequite professional.Firing an executivecorrectlyisabitmorecomplicatedandextremelyimportant.Ifyoudonotlearntherightlessons,youwillbedoingitagainsoon.Like somany things, thekey to correctly firinganexecutive ispreparation.

Hereisafour-stepprocessthatwilltreattheexecutivefairlyandimproveyourcompany.

STEP1:ROOTCAUSEANALYSIS

While it’s possible to fire an executive for bad behavior, incompetence, orlaziness,thosecasesarerareandrelativelyeasy.Unfortunately,unlessyouhaveahorriblydeficienthiringprocess, thoseareprobablynot thereasonswhyyougottothispoint.Atthislevel,almosteverycompanyscreensfortheproperskillset,motivation,andtrackrecord.Yes,thereasonthatyouhavetofireyourheadofmarketingisnotbecausehesucks;it’sbecauseyousuck.Inotherwords,thewrongwaytoviewanexecutivefiringisasanexecutive

failure;thecorrectwaytoviewanexecutivefiringisasaninterview/integrationprocesssystemfailure.Therefore,thefirststeptoproperlyfiringanexecutiveisfiguringoutwhyyouhiredthewrongpersonforyourcompany.Youmayhaveblownitforavarietyofreasons:

Youdidapoorjobdefiningthepositioninthefirstplace.Ifyoudon’tknowwhatyouwant,youwillbeunlikelytogetit.Fartoooften,CEOshireexecutivesbasedonanabstractnotionofwhattheythinkandfeelthe

executiveshouldbelike.Thiserroroftenleadstotheexecutivenotbringingthekey,necessaryqualitiestothetable.Youhiredforlackofweaknessratherthanforstrengths.Thisisespeciallycommonwhenyourunaconsensus-basedhiringprocess.Thegroupwilloftenfindthecandidate’sweaknesses,buttheywon’tplaceahighenoughvalueontheareaswhereyouneedtheexecutivetobeaworld-classperformer.Asaresult,youhireanexecutivewithnosharpweaknesses,butwhoismediocrewhereyouneedhertobegreat.Ifyoudon’thaveworld-classstrengthswhereyouneedthem,youwon’tbeaworld-classcompany.Youhiredforscaletoosoon.ThemostconsistentlywrongadvicethatventurecapitalistsandexecutiverecruitersgiveCEOsistohiresomeone“bigger”thanrequired.“Thinkaboutthenextthreetofiveyearsandhowyouwillbealargecompany”ishowthebadadviceusuallysounds.It’sgreattohirepeoplewhocanrunalarge-scaleorganizationifyouhaveone.It’salsogreattohirepeoplewhoknowhowtogrowanorganizationveryfastifyouarereadytogrowyourorganizationveryfast.However,ifyoudonotoryouarenot,thenyouneedsomeonewhocandothejobforthenexteighteenmonths.Ifyouhiresomeonewhowillbegreatineighteenmonthsbutwillbepoorforthenexteighteenmonths,thecompanywillrejectherbeforesheevergetsachancetoshowherstuff.Yourotheremployeeswillwonder:Whydidwegiveherallthosestockoptionswhenshe’snotcontributinganything?Thosekindsofquestionsareimpossibletorecoverfrom.Itturnsoutthatventurecapitalistsandexecutiverecruitersarenotstupid;theyjustlearnedthewronglessonsfrompreviousfailures.Tolearntherightlesson,seethespecialcaseofscalingandthespecialcaseoffastgrowthasexplainedbelow.Youhiredforthegenericposition.ThereisnosuchthingasagreatCEO,agreatheadofmarketing,oragreatheadofsales.Thereisonlyagreatheadofsalesforyourcompanyforthenexttwelvetotwenty-fourmonths.ThatpositionisnotthesameasthesamepositionatMicrosoftorFacebook.Don’tlookforthecandidateoutofcentralcasting.Thisisnotamovie.Theexecutivehadthewrongkindofambition.Inchapter6,Iwilldescribethedifferencebetweenambitionforthecompanyandambitionforoneself.Ifanexecutivehasthewrongkindofambition,thendespiteherskills,thecompanymayrejecther.

Youfailedtointegratetheexecutive.Bringinganewpersonintoyourcompanyinanimportantroleisdifficult.Otheremployeeswillbequicktojudge,herexpectationsmaybedifferentfromyours,andthejobmaybelargelyundefined.Besuretoreviewandimproveyourintegrationplanafteryoufireanexecutive.

ThespecialcaseofscalingA fairly common reason for firing an executive is that when the companyquadruples in size, theexecutiveno longerdoes the jobeffectivelyat thenewsize.Thereasonisthatwhenacompanymultipliesinsize,themanagementjobsbecomebrand-newjobs.Asa result,everybodyneeds to requalify for thenewjob, because the new job and the old job are not the same. Running a two-hundred-person global sales organization is not the same job as running atwenty-five-personlocalsalesteam.Ifyougetlucky,thepersonyouhiredtorunthe twenty-five-person team will have learned how to run the two-hundred-person team. If not, youneed tohire the rightperson for thenew job.This isneitheranexecutivefailurenorasystemfailure;itislifeinthebigcity.Donotattempttoavoidthisphenomenon,asyouwillonlymakethingsworse.

ThespecialcaseoffastgrowthIf you build a great product and the market wants it, you will find yourselfneeding to grow your company extremely quickly. Nothing will ensure yoursuccesslikehiringtherightexecutivewhohasgrownanorganizationlikeyoursveryquicklyandsuccessfullybefore.Notethatthisisnotthesameasinheritinga very large organization or working your way up to running a very largeorganization.Makesureyouhiretherightkindoffast-growthexecutive.Also,donothirethispersonifyouarenotreadytogivethemlotsofbudgettogrowtheirorganization;expect themtodowhat theydo.Thesuccessful fast-growthexecutive is so important to building successful startups that recruiters andventure capitalists often adviseCEOs to bring them in before the company isready.Onceyouidentifytheproblem,thenyoucreatethebasisforthenextsteps.

STEP2:INFORMINGTHEBOARD

Informingtheboardistrickyandmanyissuescanfurthercomplicatethetask:

Thisisthefifthorsixthexecutivethatyouhadtofire.Thisisthethirdexecutivethatyoufiredinthisrole.Thecandidatewasreferredbyaboardmemberwhorecommendedtheexecutiveasasuperstar.

Realizethatinanyofthesecasestheboardwillbeatleastsomewhatalarmedand there is nothing that you can do about that. But keep in mind that yourchoicesare:(a)alarmtheboardor(b)enableanineffectiveexecutivetoremainin her position.While choice (a) is not great, it’s a heck of a lot better thanchoice(b).Leavinga failing leader inplacewillcauseanentiredepartment inyour company to slowly rot.Let that happen and the boardwill bemore thanalarmed.Youshouldhavethreegoalswiththeboard:

Gettheirsupportandunderstandingforthedifficulttaskthatyouwillexecute.Youshouldstartbymakingsurethattheyunderstandtherootcauseandyourplantoremedythesituation.Thiswillgivethemconfidenceinyourabilitytohireandmanageoutsideexecutivesinthefuture.Gettheirinputandapprovalfortheseparationpackage.Thiswillbecriticalforthenextstep.Executivepackagesarelargerthanregularseverancepackagesandrightlyso.Ittakesabouttentimeslongerforanexecutivetofindanewjobthanitdoesforanindividualcontributor.Preservethereputationofthefiredexecutive.Thefailurewasverylikelyateameffort,andit’sbesttoportrayitthatway.Youdon’tmakeyourselflookgoodbytrashingsomeonewhoworkedforyou.AmatureapproachtothisissuewillhelpkeeptheboardconfidentinyourabilitytobeCEO.It’salsothefairanddecentthingtodo.

Finally, firing an executive turns out to be a piece of news that’s handledbetter with individual phone calls than in dramatic fashion during a boardmeeting. It takes a bit longer, but it’s well worth the effort. Individualconversations will be particularly important if one of the board membersintroduced the executive to the company. Once everyone agrees individually,youcanfinalizethedetailsinaboardmeetingorcall.

STEP3:PREPARINGFORTHECONVERSATION

Afteryouknowwhatwentwrongandhaveinformedtheboard,youshouldtellthe executive as quickly as possible. In preparation for that meeting, Irecommend scripting or rehearsing what you plan to say so that you do notmisspeak.Theexecutivewillremembertheconversationforaverylongtime,soyouneedtogetitright.

Aspartofyourpreparation,you should reviewanyperformance reviewsorwritten performance conversations to understand any inconsistencies in yourpriorcommunication.Threekeystogettingitright:

1.Beclearonthereasons.Youhavethoughtaboutthislongandhard;don’tequivocateorsugarcoatit.Youoweittothemtobeclearaboutwhatyouthinkhappened.2.Usedecisivelanguage.Donotleavethediscussionopen-ended.Thisisnotaperformancereview;it’safiring.Usewordsandphraseslike“Ihavedecided”ratherthan“Ithink.”3.Havetheseverancepackageapprovedandready.Oncetheexecutivehearsthenews,shewillstopcaringaboutthecompanyanditsissues;shewillbehighlyfocusedonherselfandherfamily.Bereadytoprovidespecificdetailsoftheseverancepackage.

Finally, the executive will be keenly interested in how the news will becommunicated to the company and to the outside world. It is best to let herdecide.BillCampbelloncegavemeacriticalbitofadvicewhenIwaspreparingto fire an executive.He said, “Ben, you cannot let him keep his job, but youabsolutelycanlethimkeephisrespect.”

STEP4:PREPARINGTHECOMPANYCOMMUNICATION

Afteryouhave informed the executive,youmustquicklyupdate the companyandyourstaffonthechange.Thecorrectorderforinformingthecompanyis(1)theexecutive’sdirectreports—becausetheywillbemostimpacted;(2)theothermembersofyourstaff—becausetheywillneedtoanswerquestionsaboutit;and(3)therestofthecompany.Allofthesecommunicationsshouldhappenonthesamedayandpreferablywithinacoupleofhours.Whendisclosingthefiringtothedirectreports,makesurethatyouhaveaplanforwhomtheywillreporttointhe meantime and what’s next (executive search, reorganization, internalpromotion, or something else).Generally, it’s smart for theCEO to act in theexecutiveroleinthemeanwhile.Ifyoudoactintherole,youmustreallyact—staff meetings, one-on-ones, objective setting, etc. Doing so will provideexcellentcontinuityfor the teamandgreatly informyour thinkingonwhomtohirenext.

Asyoudidwhenyouupdatedtheboard,keepthemessagepositiveandrefrainfrom throwing the executive under the bus. The best employees in theorganizationwilllikelybetheonesclosesttotheexecutive.Ifyoutrashher,youwillputallherbestemployeesonnoticethattheyarenext.Isthatamessageyouwanttosend?When you update the company, you might worry about employees

misinterpreting thenewsand thinking thecompany is in trouble.Donot try tomaneuveraroundsuchareaction.Whenyouexpectyouremployeestoactlikeadults,theygenerallydo.Ifyoutreatthemlikechildren,thengetreadyforyourcompanytoturnintoonebigBarneyepisode.

INTHEEND

EveryCEOlikestosaysherunsagreatcompany.It’shardtotellwhether theclaimistrueuntilthecompanyortheCEOhastodosomethingreallydifficult.Firinganexecutiveisagoodtest.

DEMOTINGALOYALFRIEND

When I started Loudcloud, I hired the best people I knew—people whom Irespected, trusted, and liked. Like me, many of them did not have deepexperienceinthejobsthatIgavethem,buttheyworkednightanddaytolearn,andtheymadegreatcontributions to thecompany.Nevertheless, thedaycamewhenIneededtohiresomeoneelse,someonewithmoreexperience,torunthefunctionthatIhadpreviouslyentrustedtomyloyalfriend.Damn.Howdoyoudothat?

SHOULDYOUDOITATALL?

The first question that always comes tomind is “Do I really need to do this?WhocouldIpossiblyhirewhowillworkthishardandbleedthecompanycolorslikethis?”Sadly,ifyou’reaskingthequestion,youverylikelyalreadyknowtheanswer.Ifyouneedtobuildaworldwidesalesorganization,yourbuddywhodidthefirstfewdealsisalmostcertainlynotthebestchoice.Ashardasitmaybe,youneedtotakeaConfucianapproach.Youmustconsiderfirstalloftheotheremployeesandsecondyourfriend.Thegoodoftheindividualmustbesacrificedforthegoodofthewhole.

HOWDOYOUBREAKTHENEWS?

Onceyoumakethedecision,breakingthenewswillnotbeeasy.It’simportanttoconsidertwodeepemotionsyourfriendwillfeel:

EmbarrassmentDonotunderestimatewhatalargefactorthiswillbeinhisthinking.Allofhisfriends,relatives,andcolleaguesknowhiscurrentposition.Theyknowhowhardhe’sworkedandhowmuchhe’ssacrificedforthecompany.Howwillhepossiblyexplaintothemthathewillnolongerbepartoftheexecutiveteam?BetrayalYourfriendwillundoubtedlyfeelsomethinglikethis:I’vebeentherefromthebeginning,I’veworkedsidebysidewithyou.Howcould

youdothis?It’snotlikeyou’reperfectinyourjob,either.Howcanyoubesocomfortablesellingmeout?

Those are some powerful emotions, so get ready for an intense discussion.Ironically,thekeytoanemotionaldiscussionistotaketheemotionoutofit.Todo that, youmust be very clear in yourmind aboutwhat you’ve decided andwhatyouwanttodo.Themost important thingtodecide is thatyoureallywant todothis. Ifyou

walkintoademotiondiscussionwithanopendecision,youwillwalkoutwithamess:amessofasituationandamessofarelationship.Aspartofthedecision,you must get comfortable with the thought that the employee may quit thecompany.Giventheintenseemotionshewillfeel,thereisnoguaranteethathewillwanttostay.Ifyoucannotaffordtolosehim,youcannotmakethischange.Finally,youmustdecidethebestroleforhiminyourcompany.Theobvious

thingis tohavehimcontinueunderhisnewboss,but thismaynotbe thebestthingforhim,hisboss,orhiscareer.Yourloyalemployeewillcontinuetohavelotsofknowledgeaboutyourcompany,competition,customers,andmarketthathisnewbosslacks.Ontheonehand,thiscanbeagoodthing—hecanhelpgetthe new boss up to speed. On the other hand, when mixed with the intenseemotions of embarrassment and betrayal, you might end up with a sabotagecocktail.Another problem with this approach is that from a career path perspective

thereisnowaytopaintapictureofhimreportingtohisnewbossasanythingbutademotion.Analternative,ifappropriate,wouldbetomovehimtoanotherareaofthecompanywherehisskills,talent,andknowledgewillhelp.Thiskindof move will give him a chance to develop a new set of skills and help thecompany while he’s doing it. For young employees, getting experience indifferentareascanbehighlyvaluable.Sadly, thisoptionmaynotbe a silverbullet.Hemightnotwant towork in

anotherjob;hemightbehell-bentonkeepinghiscurrentjob,soprepareforthataswell.Onceyou’vedecided tohiresomeoneaboveyour friendanddecidedon the

alternativesthatyou’dliketoofferhim,youcanhavetheconversation.Keepinmindthatyoucannotlethimkeephisoldjob,butyoucanbefairandyoucanbehonest.Somekeystodoingthat:

Useappropriatelanguage.Makeclearwithyourlanguagethatyou’ve

decided.Aspreviouslydiscussed,usephraseslike“Ihavedecided”ratherthan“Ithink”or“I’dlike.”Bydoingthis,youwillavoidputtingtheemployeeintheawkwardpositionofwonderingwhetherheshouldlobbyforhisoldjob.Youcan’ttellhimwhathewantstohear,butyoucanbehonest.Admitreality.Ifyouareafounder-CEOlikeIwas,itprobablywon’tbelostontheemployeethatyouarejustasunderskilledforyourjobasheisforhis.Don’tdodgethisfact.Infact,admitthatifyouwereamoreexperiencedCEO,youmightbeabletodevelophimintotherole,buttwopeoplewhodon’tknowwhattheyaredoingisarecipeforfailure.Acknowledgethecontributions.Ifyouwanthimtostayinthecompany,youshouldsaythatandmakeitcrystalclearthatyouwanttohelphimdevelophiscareerandcontributetothecompany.Lethimknowthatyouappreciatewhathe’sdoneandthatyourdecisionresultsfromaforward-lookingexaminationofwhatthecompanyneeds,notareviewofhispastperformance.Thebestwaytodothis,ifappropriate,istocouplethedemotionwithanincreaseincompensation.Doingsowilllethimknowthathe’sbothappreciatedandvaluedgoingforward.

Throughallofthis,keepinmindthatitiswhatitisandnothingyoucansaywillchangethatorstopitfrombeingdeeplyupsetting.Yourgoalshouldnotbetotakethestingoutofit,buttobehonest,clear,andeffective.Yourfriendmaynotappreciatethatinthemoment,buthewillappreciateitovertime.

LIESTHATLOSERSTELL

Whenacompanystartstoloseitsmajorbattles,thetruthoftenbecomesthefirstcasualty.CEOsandemployeesworktirelesslytodevelopcreativenarrativesthathelpthemavoiddealingwiththeobviousfacts.Despitetheirintensecreativity,manycompaniesoftenendupwiththesamefalseexplanations.

SOMEFAMILIARLIES

“Sheleft,butweweregoingtofireher,orgiveherabadperformancereview.”High-techcompaniestendtotrackemployeeattritioninthreecategories:

1.Peoplewhoquit2.Peoplewhogotfired3.Peoplewhoquit,butit’sokaybecausethecompanydidn’twantthemanyway

Fascinatingly,ascompaniesbegintostruggle,thethirdcategoryalwaysseemsto grow much faster than the first. In addition, the sudden wave of “semi-performance-relatedattrition”usuallyhappensincompaniesthatclaimtohavea“super-hightalentbar.”Howdoallthesesuperstaremployeessuddenlygofromgreattocrap?Howisitpossiblethatwhenyouloseatop-ratedemployeebeforeyou can say “unwanted attrition,” the manager carefully explains how herperformancefelloff?“Wewouldhavewon,buttheotherguysgavethedealaway.”“Thecustomer

selectedustechnicallyandthinkswearethebettercompany,butourcompetitorjustgavetheproductaway.Wewouldneversellsocheaplyasitwouldhurtourreputation.”Anybodywhohaseverrunanenterprisesalesforcehasheardthisliebefore.Yougointoanaccount,youfighthard,andyoulose.Thesalesrep,notwantingtoshinethelightonhimself,blamesthe“usedcardealer”repfromtheothercompany.TheCEO,notwanting tobelieve that she’s losingproductcompetitiveness,believes the rep. Ifyouhear this lie, try tovalidate theclaimwiththeactualcustomer.I’llbetyoucan’t.

“Justbecausewemissedtheintermediatemilestonesdoesn’tmeanwewon’thitourproductschedule.”Inengineeringmeetingswherethereisgreatpressureto ship on time—a customer commitment, a quarter that depends on it, or acompetitiveimperative—everybodyhopesforgoodnews.Whenthefactsdon’talign with the good news, a clever manager will find the narrative to makeeverybodyfeelbetter—untilthenextmeeting.“Wehaveaveryhighchurnrate,butassoonasweturnonemailmarketingto

ouruserbase,peoplewill comeback.”Yes,ofcourse.The reason thatpeopleleaveour serviceanddon’t comeback is thatwehavenotbeen sending themenoughspam.Thatmakestotalsensetome,too.Wheredoliescomefrom?Toanswerthatquestion,IthoughtbacktoaconversationIhadyearsagowith

theincomparableAndyGrove.Back at the tail end of the Great Internet Bubble in 2001, as all the big

technology companies beganmissing their quarters by giant amounts, I foundmyselfwonderinghownoneofthemsawitcoming.Onewouldthinkthatafterthedot-comcrashofApril2000, companies likeCisco,Siebel, andHPwouldrealizethattheywouldsoonfaceaslowdownasmanyoftheircustomershitthewall. But despite perhaps the most massive and public early warning systemever, each CEO reiterated strong guidance right up to the point where theydramaticallywhiffedtheirquarters.IaskedAndywhythesegreatCEOswouldlieabouttheirimpendingfate.He said they were not lying to investors, but rather, they were lying to

themselves.Andyexplainedthathumans,particularlythosewhobuildthings,onlylisten

to leading indicators of good news. For example, if a CEO hears thatengagementforherapplicationincreasedanincremental25percentbeyondthenormalgrowthrateonemonth,shewillbeofftotheraceshiringmoreengineersto keep up with the impending tidal wave of demand. On the other hand, ifengagement decreases 25 percent, she will be equally intense and urgent inexplainingitaway:“Thesitewasslowthatmonth,therewerefourholidays,andwemade a UI change that caused all the problems. For gosh sakes, let’s notpanic!”Bothleadingindicatorsmayhavebeenwrong,orbothmayhavebeenright,

butourhypotheticalCEO—likealmosteveryotherCEO—only tookactiononthe positive indicator and only looked for alternative explanations on thenegativeleadingindicator.

Ifthisadvicesoundstoofamiliar.andyoufindyourselfwonderingwhyyourhonestemployeesarelyingtoyou,theansweristheyarenot.Theyarelyingtothemselves.Andifyoubelievethem,youarelyingtoyourself.

LEADBULLETS

Early inmy tenure at Netscape, whenwe realized thatMicrosoft’s newWebserverhadeveryfeaturethatourshad,butwasalsofivetimesfasterandwastobegivenaway for free, I immediatelywent towork trying topivotour serverproduct line to something thatwe could sell formoney. The late, greatMikeHomer and I worked furiously on a set of partnerships and acquisitions thatwould broaden the product line and surround the Web server with enoughfunctionalitythatwewouldbeabletosurvivetheattack.I excitedly reviewed the planwithmy engineering counterpart,Bill Turpin,

wholookedatmeasthoughIwerealittlekidwhohadmuchtolearn.BillwasalongtimeveteranofbattlingMicrosoftfromhistimeatBorlandandunderstoodwhat Iwas trying todo,buthewasnotpersuaded.Hesaid,“Ben, thosesilverbulletsthatyouandMikearelookingforarefineandgood,butourWebserverisfivetimesslower.Thereisnosilverbulletthat’sgoingtofixthat.No,wearegoingtohavetousealotofleadbullets.”Ohsnap.As a result ofBill’s advice,we focused our engineering teamon fixing the

performance issueswhileworking on the other things in the background.WeeventuallybeatMicrosoft’sperformanceandgrew the server line tobecomea$400 million business, and we would never have done it without those leadbullets.Icarriedthatlessonwithmeformanyyears.Sixyearslater,whenIwasCEO

ofOpsware,ourtoughestcompetitor,BladeLogic,startedtoconsistentlybeatusinlargedeals.Wewereapubliccompany,andthelosseswerealltoovisible.Tomakemattersworse,weneededtowinthosedealsinordertobeatWallStreet’sprojections, so the company felt tremendous pressure. Many of our smartestpeoplecametomewithideasforavoidingthebattle:

“Let’sbuildalightweightversionoftheproductandgodown-market.”“Let’sacquireacompanywithasimplerarchitecture.”“Let’sfocusonserviceproviders.”

All these approaches reinforced tomewas thatweweren’t facing amarketproblem.Thecustomerswerebuying;theyjustweren’tbuyingourproduct.Thiswasnotatimetopivot.SoIsaidthesamethingtoeveryoneofthem:“Therearenosilverbulletsforthis,onlyleadbullets.”Theydidnotwanttohearthat,but itmade thingsclear:Wehad tobuildabetterproduct.Therewasnootherway out.Nowindow, no hole, no escape hatch, no back door.We had to gothroughthefrontdooranddealwiththebig,uglyguyblockingit.Leadbullets.After ninemonths of hardwork on an extremely rugged product cycle,we

regainedourproduct leadandeventuallybuilt acompany thatwasworth$1.6billion.Without the leadbullets, I suspectwewouldhaveendedataboutone-tenththatvalue.Theremaybenothingscarierinbusinessthanfacinganexistentialthreat.So

scarythatmanyintheorganizationwilldoanythingtoavoidfacingit.Theywilllookforanyalternative,anywayout,anyexcusenot to liveordie inasinglebattle. I see this often in startup pitches. The conversations go something likethis:Entrepreneur: “We have the best product in the market by far. All the

customersloveitandpreferittocompetitorX.”Me:“WhydoescompetitorXhavefivetimesyourrevenue?”Entrepreneur: “We are using partners andOEMs, becausewe can’t build a

directchannellikecompetitorX.”Me:“Whynot?Ifyouhavethebetterproduct,whynotknuckleupandgoto

war?”Entrepreneur:“Ummm.”Me:“Stoplookingforthesilverbullet.”Therecomesatimeineverycompany’slifewhereitmustfightforitslife.If

youfindyourselfrunningwhenyoushouldbefighting,youneedtoaskyourself,“Ifourcompanyisn’tgoodenoughtowin,thendoweneedtoexistatall?”

NOBODYCARES

“Justwin,baby.”

—ALDAVIS

BackinthosebadolddaysatLoudcloud,Ioftenthoughttomyself:HowcouldIhave possibly prepared for this? How could I know that half our customerswouldgooutofbusiness?HowcouldIknowthatitwouldbecomeimpossibletoraise money in the private markets? How could I have figured out that therewould be 221 IPOs in 2000 and 19 in 2001? Could anybody expect me toachieveareasonableoutcomegiventhosecircumstances?As I was feeling sorry for myself, I randomly watched an interview with

famous football coachBill Parcells.Hewas telling the story of howhe had asimilar dilemma when he began his head coaching career. In his very firstseason,Parcells’steam,theNewYorkGiants,washitwitharashofinjuries.Heworriedincessantlyabouttheimpactoftheinjuriesonhisteam’sfortunes,asitisdifficultenoughtowinwithyourbestplayers,letaloneabunchofsubstitutes.Whenhis friendandmentor,RaidersownerAlDavis,calledParcells tocheckin,Parcellsrelayedhisinjuryissues.Parcells:“Al,Iamjustnotsurehowwecanwinwithoutsomanyofourbest

players.WhatshouldIdo?”Davis:“Bill,nobodycares,justcoachyourteam.”That might be the best CEO advice ever. Because, you see, nobody cares.

When thingsgowrong inyourcompany,nobodycares.Themediadon’tcare,your investorsdon’tcare,yourboarddoesn’tcare,youremployeesdon’tcare,andevenyourmamadoesn’tcare.Nobodycares.Andtheyarerightnottocare.Agreatreasonforfailingwon’tpreserveone

dollar for your investors,won’t save one employee’s job, or get you one newcustomer.Itespeciallywon’tmakeyoufeelonebitbetterwhenyoushutdownyourcompanyanddeclarebankruptcy.

All themental energyyouuse to elaborate yourmiserywould be far betterusedtryingtofindtheoneseeminglyimpossiblewayoutofyourcurrentmess.Spendzero timeonwhatyoucouldhavedone,anddevoteallofyour timeonwhatyoumightdo.Becauseintheend,nobodycares;justrunyourcompany.

—CHAPTER5—

TAKECAREOFTHEPEOPLE,THEPRODUCTS,ANDTHEPROFITS—INTHATORDER

“Irollwiththehardestniggas,makemoneywiththesmartestniggasIain’tgottimeforyoufuckinartistniggas

BettershutyourtrapbeforeyoubecomeatargetniggaY’allarmybratsI’mthemotherfuckinsergeantnigga.”

—THEGAME,“SCREAMON’EM”

OncewepushedtheOpswarestockpricebackabove$1,thenextproblemwasto rebuild the executive team.We had cloud services executives, but nowwewereasoftwarecompanyandneededsoftwareexecutives.Inenterprisesoftwarecompanies, thetwomost importantpositionstendtobeVPofsalesandVPofengineering.InitiallyI’dattemptedtotaketheVPofprofessionalservicesfromLoudcloudandmakehim theVPofsales.Thatdidn’twork.Thenextheadofsaleswould be the fourth one hired sincewe had founded the company threeyearsearlier—notagreattrackrecord.Moreimportant,thenextmistakeImadeonasalesleaderwouldbemylast.Themarketplace,nottomentionWallStreetinvestors,didnotleavemewithmuchrope.Tobetterprepare for thehire this time, Idecided in the interim to runsales

myself. I managed the team, ran the forecast calls, and was the one personresponsiblefor therevenuenumberforOpsware.I’d learnedthehardwaythatwhenhiringexecutives,oneshouldfollowColinPowell’s instructionsandhirefor strength rather than lackofweakness.By running sales, I understoodveryclearlythestrengthsweneeded.ImadeacarefullistandsetouttofindthesalesexecutiveswiththerightskillsandtalentsforOpsware.After interviewing about two dozen candidates—none of whom had the

strengthsIsought—IinterviewedMarkCranney.Hewasn’twhatIexpected;he

didn’t fit the stereotype of a hard-charging sales executive. For starters,Markwasaverageheight,whereasmostsalesexecutivestendtoberathertall.Next,hewasasquareguy—thatis,hewasaswideashewastall.Notfat,justsquare.Hissquare body seemed to fit rather uncomfortably into what must have been acustom-tailored suit—there is nowayanoff-the-rackbusiness suitwould fit asquareguylikeMark.AndthenIlookedathisrésumé.ThefirstthingInoticedwasthathewenttoa

schoolthatI’dneverheardof,SouthernUtahUniversity.Iaskedhimwhatkindofschool itwas.He replied,“Itwas theMITofsouthernUtah.”Thatwas thelastjokehetold.Mark’sseriousnesswassointensethatitseemedtomakehimuncomfortableinhisownskin.Hemademeuncomfortable,too.Ordinarily,thatvibewouldruleoutacandidateforme,butthestrengthsthatIneededweresocritical to the business that I was willing to overlook every weakness. Oneinterview technique that I’d used to sort the good from the badwas to ask aseriesofquestionsabouthiring,training,andmanagingsalesreps.Typically,itwouldgolikethis:Ben:“Whatdoyoulookforinasalesrep?”Candidate:“Theyneedtobesmart,aggressive,andcompetitive.Theyneedto

knowhowtodocomplexdealsandnavigateorganizations.”Ben:“Howdoyoutestforthosethingsinaninterview?”Candidate:“Umm,well,Ihireeverybodyoutofmynetwork.”Ben:“Okay,onceyougetthemonboard,whatdoyouexpectfromthem?”Candidate:“Iexpectthemtounderstandandfollowthesalesprocess,Iexpect

themtomastertheproduct,Iexpectthemtobeaccurateintheirforecasting....”Ben:“Tellmeaboutthetrainingprogramthatyoudesignedtoachievethis.”Candidate:“Umm.”Theywouldthenproceedtomakesomethingupasthey

wentalong.Mark aced the profile and interview questions, and then I asked him the

trainingquestion.I’llneverforgetthepainedlookthatcameacrosshisface.Helooked likehewanted togetupand leave the interviewright thenand there. Ifelt likeofferinghimanaspirinormaybesomeAbilify.His reactionsurprisedme,becausehe’ddonesuchanexcellentjobuptothatpoint.IlaterrealizedthatformetoaskMarkCranneytodescribetheproperwaytotrainsalesrepswaslike a layman asking Isaac Newton to explain the laws of physics.Where tobegin?Afterwhatseemedlikefiveminutesofsilence,Markreachedintohisbagand

pulledoutagianttrainingmanualhehaddesigned.Hesaidhecouldn’tpossiblyexplainwhat I needed toknowabout training in the timewehad left, but if I

wanted to schedule a follow-up meeting, he would explain the nuances oftraining salespeople to be elite across a broad set of disciplines includingprocess,products,andorganizationalselling.Heexplainedfurtherthatevenwithallthosethings,asuccessfulsalesleaderstillmustinspirecourageintheteam.HesoundedlikeGeneralPatton.IknewIhadmyguy.Unfortunately, nobody else knew that. Everymember of the executive staff

(withoneexception)andeverymemberoftheboardofdirectorsvotedthumbs-downonMarkCranney.WhenIaskedBillCampbellwhathethought,hesaid,“Iwon’tlaydownontherailroadtrackstostopyoufromhiringCranney.”Thatwasn’t theringingendorsementIwaslookingfor.Thereasonsforvoting“no”never referred to Mark’s lack of strength, but rather to his abundance ofweakness:MarkwenttoSouthernUtah.Markmadepeoplefeeluncomfortable.Markdidnotlooklikeaheadofsales.Still, themore time I spent with him, themore I knew hewas the one. In

talkingwithhimforanhour,I’dlearnmoreaboutsalesthanIhadinsixmonthsrunningsales.Hewouldevencallmewithdetailsaboutdealsmysalesteamwascompetingfor—dealsmyownsalesrepsdidn’tseemtoknowabout.ItwaslikehehadhisownsalesFBI.Idecidedtotakeastand.ItoldmyteamandtheboardthatIunderstoodtheir

concerns,butIstillwantedtomoveforwardwithMarkandplannedtoproceedwithreferencechecks.WhenIaskedMarkforhisreferences,hesurprisedmeagain.Hegavemea

list of seventy-five references.He saidhehadmore if I needed them. I calledeveryreferenceonthelist,andeveryonecalledmebackwithinonehour.Markranatightnetwork.MaybethesereferenceswerethesalesFBI.Then,justasIwasgettingreadytomakethehire,anotherexecutiveonmyteamcalledtosaythat a friend of hers knew Mark Cranney and wanted to give a negativereference.Icalledthefriend—I’llcallhimJoe—andproceededtohavethemostunusual

referencecallofmycareer:Ben:“Thanksverymuchforreachingout.”Joe:“Mypleasure.”Ben:“HowdoyouknowMarkCranney?”Joe: “Mark was an area vice president when I taught sales training at my

previous employer. Iwant to tell you that under no circumstances should youhireMarkCranney.”Ben:“Wow,that’sastrongstatement.Isheacriminal?”Joe:“No,I’veneverknownMarktodoanythingunethical.”

Ben:“Ishebadathiring?”Joe:“No,hebroughtsomeofthebestsalespeopleintothecompany.”Ben:“Canhedobigdeals?”Joe:“Yes,definitely.Markdidsomeofthelargestdealswehad.”Ben:“Isheabadmanager?”Joe:“No,hewasveryeffectiveatrunninghisteam.”Ben:“Well,thenwhyshouldn’tIhirehim?”Joe:“He’llbeaterribleculturalfit.”Ben:“Pleaseexplain.”Joe: “Well, when I was teaching new-hire sales training at Parametric

Technology Corporation, I brought inMark as a guest speaker to fire up thetroops. We had fifty new hires and I had them all excited about selling andenthusiastic about working for the company. Mark Cranney walks up to thepodium,looksatthecrowdoffreshnewrecruits,andsays,‘Idon’tgiveafuckhowwelltrainedyouare.Ifyoudon’tbringmefivehundredthousanddollarsaquarter,I’mputtingabulletinyourhead.’”Ben:“Thankyouverymuch.”Theworldlooksonewayinpeacetimebutverydifferentwhenyoumustfight

foryourlifeeveryday.Intimesofpeace,onehastimetocareaboutthingslikeappropriateness, long-term cultural consequences, and people’s feelings. Intimes ofwar, killing the enemy and getting the troops safely home is all thatcounts.IwasatwarandIneededawartimegeneral.IneededMarkCranney.Asafinalstepinmakingthehire,IneededtoexplainittoMarcAndreessen.

Ascofounderandchairmanoftheboard,Marc’sopinionmattereddeeplytotheboardandMarcwasstilluncomfortablewithCranney.Marctrustedmeenoughthathewouldletmemakethehirewhetherhelikedthecandidateornot,butitwasimportanttomethatMarcbefullyonboard.I let Marc open the conversation, because despite consistently being the

smartestpersonintheroomandpossiblytheworld,Marcissohumblethatheneverbelievesthatotherpeoplethinkheissmart,whichmakeshimsensitivetobeing ignored.Heopened the conversationby listinghis issueswithCranney:doesn’t lookor sound likeaheadof sales,went toaweak school,makeshimuncomfortable. I listenedverycarefullyandreplied,“Iagreewitheverysingleoneofthoseissues.However,MarkCranneyisasalessavant.Hehasmasteredsales to a level that far exceeds anybody that I have ever known. If he didn’thavethethingswrongwithhimthatyouenumerated,hewouldn’tbewillingtojoin a company that just traded at thirty-five cents per share; he’d beCEOof

IBM.”Marc’sreplycamequickly:“Gotit.Let’shirehim!”Andthat’showItookthekeystepinbuildingaworld-classsoftwareteamout

oftheLoudcloudrubble.AsIgottoknowMarkovertheyears,everythingthatIlearnedintheinterviewandthereferencecheckprovedout.Hewasn’taneasyculturalfit,buthewasagenius.Ineededhisgeniusandworkedwithhimonthefit.Idon’tknowthateverymemberoftheteameverbecametotallycomfortablewithMark,butintheendtheyallagreedthathewasthebestpersonpossibleforthejob.MyoldbossJimBarksdalewasfondofsaying,“Wetakecareofthepeople,theproducts, and the profits—in that order.” It’s a simple saying, but it’s deep.“Takingcareof thepeople” is themostdifficultof the threebyfarand ifyoudon’t do it, the other twowon’tmatter. Taking care of the peoplemeans thatyourcompanyisagoodplacetowork.Mostworkplacesarefarfromgood.Asorganizationsgrowlarge,importantworkcangounnoticed,thehardestworkerscangetpassedoverbythebestpoliticians,andbureaucraticprocessescanchokeoutthecreativityandremoveallthejoy.When everything went wrong from the dot-com crash to NASDAQ

threatening to delist the company, the thing that saveduswere the techniquesdevelopedinthischapter.Ifyourcompanyisagoodplacetowork,youtoomaylivelongenoughtofindyourglory.

AGOODPLACETOWORK

AtOpswareIusedtoteachamanagementexpectationscoursebecauseIdeeplybelievedintraining.ImadeitclearthatIexpectedeverymanagertomeetwithherpeopleonaregularbasis.Ievengaveinstructionsonhowtoconductaone-on-onemeetingsotherecouldbenoexcuses.Thenoneday,whileIhappilywentaboutmyjob,itcametomyattentionthat

oneofmymanagershadn’thadaone-on-onemeetingwithanyofhisemployeesinmore than sixmonths.While I knew to “expect what I inspect,” I did notexpectthis.Noone-on-oneinmorethansixmonths?Howwasitpossibleformeto invest so much time thinking about management, preparing materials, andpersonally trainingmymanagers and thengetnoone-on-ones for sixmonths?Wow,somuchforCEOauthority.Ifthat’showthemanagerslistentome,thenwhydoIevenbothercomingtowork?Ithoughtthatleadingbyexamplewouldbethesurewaytogetthecompany

todowhatIwanted.Lordknowsthecompanypickedupallofmybadhabits,sowhy didn’t they pick up my good habits? Had I lost the team? I recalled aconversation I’d had with my father many years ago regarding TommyHeinsohn, theBostonCelticsbasketball coachat the time.Heinsohnhadbeenoneofthemostsuccessfulcoachesintheworld,includingbeingnamedcoachoftheyearandwinningtwoNBAchampionships.However, he had gone downhill fast and now had the worst record in the

league.Iaskedmyfatherwhathappened.Hesaid,“Theplayersstoppedpayingattention tohis temper tantrums.Heinsohnused toyell at the teamand they’drespond. Now they just ignore him.”Was the team now ignoring me? Had Iyelledatthemonetimetoomany?ThemoreIthoughtaboutit,themoreIrealizedthatwhileIhadtoldtheteam

“what”todo,Ihadnotbeenclearabout“why”Iwantedthemtodoit.Clearly,myauthorityalonewasnotenoughtogetthemtodowhatIwanted.Giventhelargenumberofthingsthatweweretryingtoaccomplish,managerscouldn’tgetto everything andcameupwith their ownpriorities.Apparently, thismanagerdidn’t think that meeting with his people was all that important and I hadn’texplainedtohimwhyitwassoimportant.

SowhydidIforceeverymanagerthroughmanagementtraining?WhydidIdemand that managers have one-on-ones with employees? After muchdeliberation withmyself, I settled on an articulation of the core reason and Icalleduptheoffendingmanager’sboss—I’llcallhimSteve—andtoldhimthatIneededtoseehimrightaway.WhenStevecameintomyofficeIaskedhimaquestion:“Steve,doyouknow

whyIcametoworktoday?”Steve:“Whatdoyoumean,Ben?”Me: “Why did I botherwaking up?Why did I bother coming in? If itwas

about themoney,couldn’tIsell thecompanytomorrowandhavemoremoneythanIeverwanted?Idon’twanttobefamous,infactjusttheopposite.”Steve:“Iguess.”Me:“Well,thenwhydidIcometowork?”Steve:“Idon’tknow.”Me: “Well, let me explain. I came to work because it’s personally very

importanttomethatOpswarebeagoodcompany.It’simportanttomethatthepeoplewho spend twelve to sixteen hours a day here, which ismost of theirwakinglife,haveagoodlife.It’swhyIcometowork.”Steve:“Okay.”Me: “Doyouknow thedifferencebetweenagoodplace towork andabad

placetowork?”Steve:“Umm,Ithinkso.”Me:“Whatisthedifference?”Steve:“Umm,well...”Me:“Letmebreakitdownforyou.Ingoodorganizations,peoplecanfocus

ontheirworkandhaveconfidencethatiftheygettheirworkdone,goodthingswillhappen forboth thecompanyand thempersonally. It isa truepleasure toworkinanorganizationsuchasthis.Everypersoncanwakeupknowingthatthework they do will be efficient, effective, and make a difference for theorganizationandthemselves.Thesethingsmaketheirjobsbothmotivatingandfulfilling.“Inapoororganization,on theotherhand,peoplespendmuchof their time

fighting organizational boundaries, infighting, and broken processes. They arenot even clear onwhat their jobs are, so there is noway to know if they aregettingthejobdoneornot.Inthemiraclecasethattheyworkridiculoushoursandgetthejobdone,theyhavenoideawhatitmeansforthecompanyortheircareers.Tomakeitallmuchworseandrubsaltinthewound,whentheyfinally

work up the courage to tell management how fucked-up their situation is,managementdeniesthereisaproblem,thendefendsthestatusquo,thenignorestheproblem.”Steve:“Okay.”Me: “Are you aware that your manager Tim has not met with any of his

employeesinthepastsixmonths?”Steve:“No.”Me:“Nowthatyouareaware,doyourealizethatthereisnopossiblewayfor

himtoevenbeinformedastowhetherornothisorganizationisgoodorbad?”Steve:“Yes.”Me:“Insummary,youandTimarepreventingmefromachievingmyoneand

only goal. You have become a barrier blocking me from achieving my mostimportantgoal.Asaresult,ifTimdoesn’tmeetwitheachoneofhisemployeesin thenext twenty-fourhours, Iwillhavenochoicebut to firehimand to fireyou.Areweclear?”Steve:“Crystal.”

WASTHATREALLYNECESSARY?

Youmight argue that nomatter howwellmanaged a company is, it will failwithout product/market fit. You might argue further that horribly managedcompanies that achievemassive product/market fit succeed just fine.And youwouldberightonbothaccounts.Sowasitreallynecessaryformetomakesuchadramaticspeechandthreatenoneofmyexecutives?Ithinkitwas,forthefollowingthreereasons:

Beingagoodcompanydoesn’tmatterwhenthingsgowell,butitcanbethedifferencebetweenlifeanddeathwhenthingsgowrong.Thingsalwaysgowrong.Beingagoodcompanyisanendinitself.

THEDIFFERENCEBETWEENLIFEANDDEATH

Whenthingsgowell,thereasonstostayatacompanyaremany:

Yourcareerpathiswideopenbecauseasthecompanygrowslotsofinterestingjobsnaturallyopenup.

Yourfriendsandfamilythinkyouareageniusforchoosingtoworkatthe“it”companybeforeanyoneelseknewitwas“it.”Yourrésumégetsstrongerbyworkingatablue-chipcompanyinitsheyday.Oh,andyouaregettingrich.

Whenthingsgopoorly,allthosereasonsbecomereasonstoleave.Infact,theonlythingthatkeepsanemployeeatacompanywhenthingsgohorriblywrong—otherthanneedingajob—isthatshelikesherjob.

THINGSALWAYSGOWRONG

There has never been a company in the history of the world that had amonotonously increasing stock price. In bad companies, when the economicsdisappear,sodotheemployees.Intechnologycompanies,whentheemployeesdisappear,thespiralbegins:Thecompanydeclinesinvalue,thebestemployeesleave, the company declines in value, the best employees leave. Spirals areextremelydifficulttoreverse.

BEINGAGOODCOMPANYISANENDINITSELF

When I first met Bill Campbell, he was chairman of Intuit, on the board ofApple,andamentor tomanyof the topCEOsin the industry.However, thosethingsdidnot impressmenearlyasmuchashis timerunningGOCorporationbackin1992.Thecompanyraisedmoremoneythanalmostanyotherventure-capital-backedstartupinhistoryandlostnearlyallofitbeforesellingitselffornearlynothingtoAT&Tin1994.Nowthatprobablydoesn’tsoundimpressive.Infact,itprobablysoundslikea

horrible failure.But I’dmet dozens ofGOemployees inmy career, includinggreatpeoplelikeMikeHomer,DannyShader,FrankChen,andStrattonSclavos.Theamazing thingwas thateveryoneof thoseGOemployeescountedGOasone of the greatestwork experiences of their lives. The bestwork experienceeverdespitethefactthattheircareersstoodstill,theymadenomoney,andtheywerefront-pagefailures.GOwasagoodplacetowork.ThismademerealizewhatanamazinglyeffectiveCEOBillwas.Apparently,

JohnDoerrthoughtthat,too,becausewhenScottCookneededaCEOforIntuit,JohnrecommendedBilleventhoughBilllostatonofJohn’smoneyatGO.And

foryears,everyonewhoevercameintocontactwithGOemployeesknewwhatBillwasabout.Hewasaboutbuildinggoodcompanies.Ifyoudonothingelse,belikeBillandbuildagoodcompany.

WHYSTARTUPSSHOULDTRAINTHEIRPEOPLE

I learned about why startups should train their people when I worked atNetscape.PeopleatMcDonald’sgettrainedfortheirpositions,butpeoplewithfarmorecomplicatedjobsdon’t.Itmakesnosense.Wouldyouwanttostandonthe line of the untrained person at McDonald’s?Would you want to use thesoftwarewrittenby the engineerwhowasnever toldhow the rest of the codeworked?Alotofcompaniesthinktheiremployeesaresosmartthattheyrequirenotraining.That’ssilly.When I first became a manager, I had mixed feelings about training.

Logically, training for high-tech companies made sense, but my personalexperience with training programs at the companies where I hadworkedwasunderwhelming. The courses were taught by outside firms who didn’t reallyunderstandourbusinessandwere teachingthings thatweren’trelevant.ThenIread chapter 16 of Andy Grove’s management classic, High OutputManagement,titled“WhyTrainingIstheBoss’sJob,”anditchangedmycareer.Grovewrote,“Mostmanagersseemtofeelthattrainingemployeesisajobthatshouldbeleft toothers.I,ontheotherhand,stronglybelievethatthemanagershoulddoithimself.”When I was director of product management at Netscape, I was feeling

frustrated by how little value most product managers added to the business.Based on Andy’s guidance, I wrote a short document called “Good ProductManager/Bad ProductManager,” which I used to train the team onmy basicexpectations. (That document follows on page 111.) I was shocked by whathappened next. The performance of my team instantly improved. Productmanagerswhom I had almostwritten off as hopeless became effective. PrettysoonIwasmanagingthehighest-performingteaminthecompany.Basedonthisexperience,afterstartingLoudcloud,Iheavilyinvestedintraining.Icreditthatinvestmentwithmuchofoureventualsuccess.Andthewholethingstartedwithasimpledecisiontotrainmypeopleandanevensimplertrainingdocument.So,IwillnowpayforwardmydebttoAndyGroveandexplainwhy,what,andhow

youshoulddothesameinyourcompany.

WHYYOUSHOULDTRAINYOURPEOPLE

Almosteveryonewhobuildsa technologycompanyknows thatpeopleare themost important asset. Properly run startups place a great deal of emphasis onrecruitingandtheinterviewprocessinordertobuildtheirtalentbase.Toooftenthe investment in people stops there. There are four core reasons why itshouldn’t:

1.Productivity

I often see startups keep careful statistics of how many candidates they’vescreened,howmanyhavemadeit tothefull interviewprocess,andhowmanypeople they’ve hired. All of these statistics are interesting, but the mostimportantstatisticismissing:Howmanyfullyproductiveemployeeshavetheyadded?Byfailingtomeasureprogresstowardtheactualgoal,theylosesightofthevalueof training. If theymeasuredproductivity, theymightbehorrified tofindthatallthoseinvestmentsinrecruiting,hiring,andintegrationweregoingtowaste. Even if they were made aware of low productivity among newemployees, most CEOs think that they don’t have time to invest in training.AndyGrovedoesthemathandshowsthattheoppositeistrue:

Training is, quite simply, one of the highest-leverage activities amanagercanperform.Considerforamomentthepossibilityofyourputtingonaseriesoffourlecturesformembersofyourdepartment.Let’scountonthreehourspreparationforeachhourofcoursetime—twelvehoursofworkintotal.Saythatyouhavetenstudentsinyourclass.Nextyeartheywillworkatotalofabouttwentythousandhoursforyour

organization. Ifyour trainingeffortsresult ina1percent improvement inyoursubordinates’performance,yourcompanywillgaintheequivalentoftwohundredhoursofworkastheresultof theexpenditureofyourtwelvehours.

2.Performancemanagement

Whenpeopleinterviewmanagers,theyoftenliketoask,Haveyoufiredanyone?

Or how many people have you fired? Or how would you go about firingsomeone?Theseareallfinequestions,butoftentherightquestionistheonethatisn’tasked:Whenyoufiredtheperson,howdidyouknowwithcertaintythattheemployee both understood the expectations of the job and was still missingthem? The best answer is that themanager clearly set expectations when shetrainedtheemployeeforthejob.Ifyoudon’ttrainyourpeople,youestablishnobasis for performance management. As a result, performance management inyourcompanywillbesloppyandinconsistent.

3.Productquality

Often founders start companies with visions of elegant, beautiful productarchitecturesthatwillsolvesomanyofthenastyissuesthattheywereforcedtodeal with in their previous jobs. Then, as their company becomes successful,theyfindthattheirbeautifulproductarchitecturehasturnedintoaFrankenstein.Howdoes this happen?As success drives the need to hire newengineers at arapid rate, companies neglect to train the new engineers properly. As theengineersareassignedtasks,theyfigureouthowtocompletethemasbesttheycan. Often this means replicating existing facilities in the architecture, whichleads to inconsistencies in the user experience, performance problems, and ageneralmess.Andyouthoughttrainingwasexpensive.

4.Employeeretention

DuringatimeofparticularlyhighattritionatNetscape,Idecidedtoreadalloftheexitinterviewsfortheentirecompanytobetterunderstandwhypeoplequithigh-techcompanies.Afterputtingeconomicsaside,Ifoundthatthereweretwoprimaryreasonswhypeoplequit:

Theyhatedtheirmanager;generallytheemployeeswereappalledbythelackofguidance,careerdevelopment,andfeedbacktheywerereceiving.Theyweren’tlearninganything:Thecompanywasn’tinvestingresourcesinhelpingemployeesdevelopnewskills.

Anoutstandingtrainingprogramcanaddressbothissueshead-on.

WHATSHOULDYOUDOFIRST?

Thebestplacetostartiswiththetopicthatismostrelevanttoyouremployees:the knowledge and skill that they need to do their job. I call this functionaltraining. Functional training can be as simple as training a new employee onyour expectations for them (see “Good Product Manager/Bad ProductManager”)andascomplexasamultiweekengineeringbootcamptobringnewrecruitscompletelyup tospeedonallof thehistoricalarchitecturalnuancesofyourproduct.Thetrainingcoursesshouldbetailoredtothespecificjob.Ifyouattemptthemorecomplex-stylecourse,besuretoenlistthebestexpertsontheteamaswellasthemanager.Asahappysideeffect, thistypeofeffortwilldomore to build a powerful, positive company culture than a hundred culture-buildingstrategicoff-sitemeetings.The other essential component of a company’s training program is

management training. Management training is the best place to start settingexpectations for yourmanagement team.Do you expect them to hold regularone-on-one meetings with their employees? Do you expect them to giveperformancefeedback?Doyouexpectthemtotraintheirpeople?Doyouexpectthem to agree on objectiveswith their team? If you do, then you’d better tellthem, because the management state of the art in technology companies isextremely poor. Once you’ve set expectations, the next set of managementcourseshasalreadybeendefined;theyarethecoursesthatteachyourmanagershowtodothethingsyouexpect(howtowriteaperformancerevieworhowtoconductaone-on-one).Onceyouhavemanagementtrainingandfunctionaltraininginplace,thereare

other opportunities as well. One of the great things about building a techcompany is the amazing people that you can hire. Take your best people andencouragethemtosharetheirmostdevelopedskills.Traininginsuchtopicsasnegotiating,interviewing,andfinancewillenhanceyourcompany’scompetencyinthoseareasaswellasimproveemployeemorale.Teachingcanalsobecomeabadgeofhonorforemployeeswhoachieveanelitelevelofcompetence.

IMPLEMENTINGYOURTRAININGPROGRAM

Nowthatweunderstandthevalueofthetrainingandwhattotrainon,howdowegetourorganizationtodowhatwewant?Thefirstthingtorecognizeisthatnostartuphastimetodooptionalthings.Therefore,trainingmustbemandatory.The first two types of training (functional and management) can be easily

enforcedasfollows:

Enforcefunctionaltrainingbywithholdingnewemployeerequisitions.AsAndyGrovewrites,thereareonlytwowaysforamanagertoimprovetheoutputofanemployee:motivationandtraining.Therefore,trainingshouldbethemostbasicrequirementforallmanagersinyourorganization.Aneffectivewaytoenforcethisrequirementisbywithholdingnewemployeerequisitionsfrommanagersuntilthey’vedevelopedatrainingprogramfortheTBH,“ToBeHired.”Enforcemanagementtrainingbyteachingityourself.ManagingthecompanyistheCEO’sjob.Whileyouwon’thavetimetoteachallofthemanagementcoursesyourself,youshouldteachthecourseonmanagementexpectations,becausetheyare,afterall,yourexpectations.Makeitanhonortoparticipateinthesesessionsbyselectingthebestmanagersonyourteamtoteachtheothercourses.Andmakethatmandatory,too.

Ironically, the biggest obstacle to putting a training program in place is theperception that it will take too much time. Keep in mind that there is noinvestmentthatyoucanmakethatwilldomoretoimproveproductivityinyourcompany.Therefore,beingtoobusytotrainisthemoralequivalentofbeingtoohungrytoeat.Furthermore,it’snotthathardtocreatebasictrainingcourses.When I ran the server product management group at Netscape, I becameextremely frustrated that everybody on the team I inherited had a completelyuniqueanddifferent interpretationof their job.Finally, Ihadanepiphany thatnobody in the industry had ever defined the product management job. Whatfollows was my attempt to do that and bring down my blood pressure.Amazingly,peoplestillreadittoday.Thistaughtmetheimportanceoftraining.

GOODPRODUCTMANAGER/BADPRODUCTMANAGER

Goodproductmanagersknowthemarket,theproduct,theproductline,andthecompetition extremelywell andoperate froma strongbasisofknowledge andconfidence.Agoodproductmanager is theCEOof theproduct.Goodproductmanagerstakefullresponsibilityandmeasurethemselvesintermsofthesuccessoftheproduct.Theyareresponsibleforrightproduct/right timeandall thatentails.Agood

productmanagerknowsthecontextgoingin(thecompany,ourrevenuefunding,competition, etc.), and they take responsibility for devising and executing awinningplan(noexcuses).Bad product managers have lots of excuses. Not enough funding, the

engineering manager is an idiot, Microsoft has ten times as many engineersworkingonit,I’moverworked,Idon’tgetenoughdirection.OurCEOdoesn’tmakethesekindsofexcusesandneithershouldtheCEOofaproduct.Goodproductmanagersdon’tgetallof their timesuckedupby thevarious

organizations thatmustwork together to deliver the right product at the righttime.Theydon’t takeall theproduct teamminutes; theydon’tprojectmanagethevariousfunctions; theyarenotgofers forengineering.Theyarenotpartofthe product team; they manage the product team. Engineering teams don’tconsider good product managers a “marketing resource.” Good productmanagersarethemarketingcounterpartstotheengineeringmanager.Goodproductmanagers crisplydefine the target, the “what” (asopposed to

the“how”),andmanagethedeliveryofthe“what.”Badproductmanagersfeelbest about themselves when they figure out “how.” Good product managerscommunicatecrisplytoengineeringinwritingaswellasverbally.Goodproductmanagers don’t give direction informally. Good product managers gatherinformationinformally.Good product managers create collateral, FAQs, presentations, and white

papers thatcanbe leveragedbysalespeople,marketingpeople,andexecutives.Badproductmanagerscomplainthattheyspendalldayansweringquestionsforthesalesforceandareswamped.Goodproductmanagersanticipatetheseriousproduct flaws and build real solutions.Bad productmanagers put out fires allday.Good product managers take written positions on important issues

(competitivesilverbullets,tougharchitecturalchoices,toughproductdecisions,and markets to attack or yield). Bad product managers voice their opinionsverbally and lament that the “powers that be” won’t let it happen. Once badproductmanagersfail,theypointoutthattheypredictedtheywouldfail.Good product managers focus the team on revenue and customers. Bad

product managers focus the team on how many features competitors arebuilding. Good product managers define good products that can be executedwith a strongeffort.Badproductmanagersdefinegoodproducts that can’t beexecutedorletengineeringbuildwhatevertheywant(that is,solvethehardestproblem).

Good product managers think in terms of delivering superior value to themarketplace during product planning and achievingmarket share and revenuegoalsduring thego-to-marketphase.Badproductmanagersget very confusedabout the differences among delivering value, matching competitive features,pricing, and ubiquity. Good product managers decompose problems. Badproductmanagerscombineallproblemsintoone.Goodproductmanagersthinkaboutthestorytheywantwrittenbythepress.

Badproductmanagers thinkaboutcoveringeveryfeatureandbeingabsolutelytechnically accurate with the press. Good product managers ask the pressquestions. Bad product managers answer any press question. Good productmanagers assumemembers of the press and the analyst community are reallysmart. Bad product managers assume that journalists and analysts are dumbbecausetheydon’tunderstandthesubtlenuancesoftheirparticulartechnology.Goodproductmanagerserronthesideofclarity.Badproductmanagersnever

even explain the obvious. Good product managers define their job and theirsuccess.Badproductmanagersconstantlywanttobetoldwhattodo.Good product managers send their status reports in on time every week,

becausetheyaredisciplined.Badproductmanagersforgettosendintheirstatusreportsontime,becausetheydon’tvaluediscipline.

ISITOKAYTOHIREPEOPLEFROMYOURFRIEND’SCOMPANY?

Everygoodtechnologycompanyneedsgreatpeople.Thebestcompaniesinvesttime,money, and sweatequity intobecomingworld-class recruitingmachines.Buthowfarshouldyoutakeyourquesttobuildtheworld’sgreatestteam?Isitfair game to hire employees from your friend’s company? Will you still befriends?First,whatdoImeanby“friends”?Therearetworelevantcategories:

ImportantbusinesspartnersFriends

For this discussion, friends and important business partners are roughly thesame.MostCEOswouldnever targeta friend’scompanyasasourceof talent.As

CEO,onegenerallydoesn’thavemanytruefriendsinbusiness,andraidingyourfriend’scompanyisasurewaytoloseone.Nevertheless,almosteveryCEOwillbe facedwith the decision ofwhether to hire an employee out of her friend’scompany.Howdoesithappen?Whenisitokay?Whenwillitcostyouafriend?

BUTTHEYWEREALREADYLOOKING

Italwaysstartsinthesameway.YourfriendCathyhasagreatengineerworkingfor her namedMitchell.Mitchell happens to be friends with one of your topengineers. Your engineer bringsMitchell in for an interview, unbeknownst toyou, andhenaturally sails through theprocess.The final step is the interviewwith you, the CEO. You immediately notice thatMitchell currently works atyourgood friendCathy’scompany.Youcheckwithyourpeople tomakesurethattheydidnotapproachMitchellfirst,andtheyassureyouthatMitchellwasalreadylookingandwillgotoanothercompanyifnotyours.Nowwhat?Atthispoint,youmightbethinking,“IfMitchellisleaving,thenlogicallymy

friendCathyshouldwanthimtogotomycompanyratherthantoacompetitor

oracompanywithaCEOwhomshedoesn’tlike.”MaybeCathywillseeitthatway,butprobablynot.People generally leave companies when things are not going well, so you

should assume that Cathy is fighting for her company’s life. In this situation,nothingwillcutherdeeperthanlosingagreatemployee,becausesheknowsthattheotheremployeeswillseethatasaleadingindicatorofthecompany’sdemise.EvenmoredamagingforCathyisthefactthatheremployeeswillperceiveyourmove as an act of betrayal—Cathy’s so-called friend is raiding her company.Theywill think, “Cathy is suchan ineffectiveCEO that shecannot evenkeepherfriendsfromhiringherpeople.”Inthisway,alogicalissuequicklybecomesanemotionalone.Youdon’twanttoloseCathyasafriend,soyouassureherthatMitchellisthe

exceptionandthathecametoyouandthathewillbethefirstandonlyoneofher employees that joins your company.Generally, this explanationwillworkandCathywillunderstandandappreciatethegesture.Shewillforgive,butrestassured,shewillnotforget.HermemoryofMitchellwillbe important,becauseMitchellwillbe just the

first step in the demise of your relationship. Since Mitchell is a stellar hire,Cathy’sother strongemployeeswill likely callMitchell tounderstandwhyheleftandwhereheisgoing.Hewillexplainhisreasoningandhisreasonswillbecompelling.AndsuddenlytheywillwanttofollowMitchell’spathandjoinyourcompany, too. By the time you become aware of the situation, promises willhavebeenmadetoprospectiveemployeeswhoapproachedMitchellandoffersmaybeout.In eachcase, your employeeswill assureyou that theywere approachedby

Cathy’s employees andnot viceversa.Theywill point out that the candidateshaveoffersfromothercompaniesaswell,sotheywilldefinitelyleaveandyoumight as well benefit from their restlessness. Cathy’s managers will almostcertainlytelladifferentstory.Theywillpleadwithhertogetherfriendtostopraiding their stable of employees or else theywill never be able tomeet theircommitments.ThiswillembarrassandenrageCathy.Intheend,socialpressurewilltrumpallyourbrilliantcountervailinglogic.Here’s an easy way to think about the dynamic. If your husband left you,

wouldyouwantyourbestfriendtodatehim?He’sgoingtodatesomebody,sowouldn’tyouwantyourfriendtohavehim?Itseemslogical,butthissituationisfarfromlogicalandyoujustlostonefriend.

SOWHATSHOULDYOUDO?

First,keepinmindthattheemployeesareeitherextremelygoodoryouprobablywon’twanttheminyourcompanyanyway.So,youwilleitherberecruitingtop-notch employees fromyour friend’s companyor youwill be addingmediocrepeople.Donotassumethepeopleyouaretakingwillnotbemissed.AgoodruleofthumbismyReflexivePrincipleofEmployeeRaiding,which

states, “If youwouldbe shockedandhorrified ifCompanyXhired several ofyour employees, then you should not hire any of theirs.”The number of suchcompaniesshouldbesmallandmayverywellbezero.Inorder toavoid thesestickysituations,manycompaniesemploywrittenor

unwrittenpoliciesthatnamecompanieswhereitisnotokaytohirewithoutCEO(orseniorexecutive)approval.Withsuchapolicyinplace,youwillbeabletogiveyourfriendonelastchancetosavetheiremployeeortoobjectpriortoyouhiringthem.With that inmind, the bestway to dealwith these situations is openly and

transparently. Once you become aware of the conflict between hiring thesuperstaremployeeanddouble-crossingyourvaluedfriend,youshouldget theissue onto the table by informing the employee that you have an importantbusinessrelationshipwithhisexistingcompanyandyouwillhavetocompleteareferencecheckwiththeCEOpriortoextendingtheoffer.Lethimknowthatifhedoesnotwantthattohappen,thenyouwillstoptheprocessnowandkeeptheprocess to date confidential. By speaking with your friend beforemaking thehire, you will be able to better judge the relationship impact of hiring heremployee. In addition, youmay avoidmaking a bad hire, as often candidateswhodowellininterviewsturnouttobebademployees.

CLOSINGTHOUGHTS

In the classic movie The Good, the Bad and the Ugly, Clint Eastwood “TheGood” and EliWallach “The Ugly” are partners in crime.Wallach, a knowncriminal,hasabountyonhisheadandthetwoofthemrunascamtocollectthereward money. Eastwood turns Wallach in and collects the reward. ThenWallachissentencedtodeathbyhanging.AsWallachsitsonahorse,handstiedbehindhisbackandabouttobehangedwitharopearoundhisneck,Eastwoodshoots the rope from a distance and freesWallach and they split the rewardmoney.Thisschemeworksbrilliantly,untilonedayEastwoodfreesWallachbutinformsWallach:“Idon’t thinkyou’lleverbeworthmorethanthreethousand

dollars.” Wallach retorts, “What do you mean?” Eastwood informs him, “Imean,ourpartnershipisuntied.Ohno,notyou.Youremaintied.I’llkeepthemoney, andyoucanhave the rope.”What follows isoneof thegreat revengepursuitsinmotionpicturehistory.So,whenyoutellyourCEOfriendthatyoudon’tthinkshe’lleverbeworth

morethanthisemployee,don’texpecttostayfriends.

WHYIT’SHARDTOBRINGBIGCOMPANYEXECSINTOLITTLECOMPANIES

Soyou’ve achieved productmarket fit and you are ready to start building thecompany.Theboard encourages you to bring in some “been there, done that”executiveswhowillprovidetherightfinancial,sales,andmarketingexpertisetohelp you transition from aworld-class product to aworld-class business.Youseeafewcandidatesthatyoulike,but theventurecapitalistontheboardsays,“You are undershooting. This is going to be a huge company.We can attractbettertalent.”Soyouaimhighandbringinasuper-accomplishedheadofsales.Thisguyhasrunhugeorganizationswiththousandsofemployees.Hehasstellarreferences and even looks the part. Your VC loves him, because he has anawesomerésumé.

SIXMONTHSLATER...

Fast-forward sixmonths and everyone in the company is wondering why thesales (ormarketing or finance or product) guywho has produced nothing gotsuchamonsterstockoptionpackage.Meanwhile,thepeopledoingalltheworkhavemuchfeweroptions.Evenworsethannotgettingyourmoney’sworth,nowthe company is in trouble, because you’ve beenmissing the numbers as yoursuper-expensiveexecutivesitsonhisbutt.Whatthefrakjusthappened?The most important thing to understand is that the job of a big company

executiveisverydifferentfromthejobofasmallcompanyexecutive.WhenIwasmanagingthousandsofpeopleatHewlett-PackardafterthesaleofOpsware,there was an incredible number of incoming demands on my time. Everyonewanted a piece of me. Little companies wanted to partner with me or sellthemselves tome,people inmyorganizationneededapprovals, otherbusinessunitsneededmyhelp,customerswantedmyattention,andsoforth.Asaresult,Ispentmostofmytimeoptimizingandtuningtheexistingbusiness.Mostofthework that I didwas “incoming.” In fact,most skilled big company executiveswilltellyouthatifyouhavemorethanthreenewinitiativesinaquarter,youaretryingtodotoomuch.Asaresult,bigcompanyexecutivestendtobeinterrupt-

driven.In contrast, when you are a startup executive, nothing happens unless you

make ithappen. In theearlydaysofacompany,youhave to takeeight to tennew initiativesadayor thecompanywill stand still.There isno inertia that’sputting thecompanyinmotion.Withoutmassive inputfromyou, thecompanywillstayatrest.

SOWHATHAPPENS?

Once you hire one of these big company executives, there are two dangerousmismatchesthatyouwillface:

1.RhythmmismatchYourexecutivehasbeenconditionedtowaitfortheemailstocomein,waitforthephonetoring,andwaitforthemeetingstogetscheduled.Inyourcompany,hewillbewaitingalongtime.Ifyournewexecutivewaits(asperhistraining),yourotheremployeeswillbecomesuspicious.You’llhearthingslike“Whatdoesthatguydoalldaylong?”and“Whydidhegetsomanyoptions?”2.SkillsetmismatchRunningalargeorganizationrequiresverydifferentskillsthancreatingandbuildinganorganization.Whenyourunalargeorganization,youtendtobecomeverygoodattaskssuchascomplexdecision-making,prioritization,organizationaldesign,processimprovement,andorganizationalcommunication.Whenyouarebuildinganorganization,thereisnoorganizationtodesign,therearenoprocessestoimprove,andcommunicatingwiththeorganizationissimple.Ontheotherhand,youhavetobeveryadeptatrunningahigh-qualityhiringprocess,haveterrificdomainexpertise(youarepersonallyresponsibleforqualitycontrol),knowhowtocreateprocessfromscratch,andbeextremelycreativeaboutinitiatingnewdirectionsandtasks.

HOWCANYOUSTOPTHINGSFROMGOINGHORRIBLYWRONG?

Therearetwokeystepstoavoidingdisaster:

1.Screenfordevastatingmismatchesintheinterviewprocess.2.Takeintegrationasseriouslyasinterviewing.

SCREENFORMISMATCHES

Howdoyou tell if the rhythmmismatchor the skill setmismatchwill be toomuchtoovercome?HerearesomeinterviewquestionsthatIfoundveryhelpful:

Whatwillyoudoinyourfirstmonthonthejob?Beware of answers that overemphasize learning. This may indicate that thecandidate thinks there is more to learn about your organization than thereactuallyis.Morespecifically,hemaythinkthatyourorganizationisascomplexashiscurrentorganization.Bewareofanyindicationthatthecandidateneedstobeinterrupt-drivenrather

thansettingthepacepersonally.Theinterruptswillnevercome.Lookforcandidateswhocomeinwithmorenewinitiativesthanyouthinkare

possible.Thisisagoodsign.

Howwillyournewjobdifferfromyourcurrentjob?Lookfor self-awarenessof thedifferenceshere. If theyhave theexperience inwhatyouneed,theywillbearticulateonthispoint.Beware of candidates who think that too much of their experience is

immediatelytransferable.Itmaypayoffdowntheline,butlikelynottomorrow.

Whydoyouwanttojoinasmallcompany?Beware of equity being the primary motivation. One percent of nothing isnothing.That’ssomethingthatbigcompanyexecutivessometimeshaveahardtimeunderstanding.It’s much better if they want to be more creative. The most important

difference between big and small companies is the amount of time runningversuscreating.Adesiretodomorecreatingistherightreasontowanttojoinyourcompany.

AGGRESSIVELYINTEGRATETHECANDIDATEONCEONBOARD

Perhaps themost critical step is integration.You should plan to spend a hugeamountoftimeintegratinganynewexecutive.Herearesomethingstokeepinmind:

Forcethemtocreate.Givethemmonthly,weekly,andevendaily

objectivestomakesurethattheyproduceimmediately.Therestofthecompanywillbewatchingandthiswillbecriticaltotheirassimilation.Makesurethatthey“getit.”Content-freeexecutiveshavenovalueinstartups.Everyexecutivemustunderstandtheproduct,thetechnology,thecustomers,andthemarket.Forceyournewbietolearnthesethings.Considerschedulingadailymeetingwithyournewexecutive.Requirethemtobringacomprehensivesetofquestionsabouteverythingtheyheardthatdaybutdidnotcompletelyunderstand.Answerthosequestionsindepth;startwithfirstprinciples.Bringthemuptospeedfast.Iftheydon’thaveanyquestions,considerfiringthem.Ifinthirtydaysyoudon’tfeelthattheyarecominguptospeed,definitelyfirethem.Puttheminthemix.Makesurethattheyinitiatecontactandinteractionwiththeirpeersandotherkeypeopleintheorganization.Givethemalistofpeopletheyneedtoknowandlearnfrom.Oncethey’vedonethat,requireareportfromthemonwhattheylearnedfromeachperson.

FINALTHOUGHTS

Nothingwillaccelerateyourcompany’sdevelopment likehiringsomeonewhohasexperiencebuildingaverysimilarcompanyatlargerscale.However,doingsocanbefraughtwithperil.Makesuretopayattentiontotheimportantleadingindicatorsofsuccessandfailure.

HIRINGEXECUTIVES:IFYOU’VENEVERDONETHEJOB,HOWDOYOUHIRE

SOMEBODYGOOD?

The biggest difference between being a great functionalmanager and being agreat general manager—and particularly a great CEO—is that as a generalmanager,youmusthireandmanagepeoplewhoarefarmorecompetentattheirjobs than youwould be at their jobs. In fact, often youwill have to hire andmanagepeopletodojobsthatyouhaveneverdone.HowmanyCEOshavebeenheadofHR,engineering,sales,marketing,finance,andlegal?Probablynone.So,withnoexperience,howdoyouhiresomeonegood?

STEP1:KNOWWHATYOUWANT

Step1isdefinitelythemostimportantstepintheprocessandalsotheonethatgetsskippedmostoften.Asthegreatself-helpcoachTonyRobbinssays,“Ifyoudon’tknowwhatyouwant,thechancesthatyou’llgetitareextremelylow.”Ifyouhaveneverdonethejob,howdoyouknowwhattowant?First, you must realize how ignorant you are and resist the temptation to

educateyourselfsimplybyinterviewingcandidates.Whiletheinterviewprocesscanbehighlyeducational,usingthatasthesoleinformationsourceisdangerous.Doingsowillmakeyoususceptibletothefollowingtraps:

HiringonlookandfeelItmayseemsillytothinkthatanyonewouldhireanexecutivebasedonthewaytheylookandsoundinaninterview,butinmyexperience,lookandfeelarethetopcriteriaformostexecutivesearches.WhenyoucombineaCEOwhodoesn’tknowwhatshewantsandaboardofdirectorsthathasn’tthoughtmuchaboutthehire,whatdoyouthinkthecriteriawillbe?LookingforsomeoneoutofcentralcastingIfIhadfollowedthispath,IwouldneverhavehiredMarkCranneyandyouprobablywouldnotbereadingthisnow.Thiswrongheadedapproachisthemoralequivalentof

lookingforthePlatonicidealforaheadofsales.Youimaginewhattheperfectsalesexecutivemightbelike,andthenyouattempttomatchreal-worldcandidatestoyourmodel.Thisisabadideaforseveralreasons.First,youarenothiringanabstractexecutivetoworkatanarbitrarycompany.Youmusthiretherightpersonforyourcompanyatthisparticularpointintime.TheheadofsalesatOraclein2010wouldlikelyhavefailedin1989.TheVPofengineeringatApplemightbeexactlythewrongchoiceforFoursquare.Thedetailsandthespecificsmatter.Second,yourimaginarymodelisalmostcertainlywrong.Whatisyourbasisforcreatingthismodel?Finally,itwillbeincrediblydifficulttoeducateaninterviewteamonsuchanabstractsetofcriteria.Asaresult,everybodywillbelookingforsomethingdifferent.ValuinglackofweaknessratherthanstrengthThemoreexperienceyouhave,themoreyourealizethatthereissomethingseriouslywrongwitheveryemployeeinyourcompany(includingyou).Nobodyisperfect.

Theverybestwaytoknowwhatyouwantistoactintherole.Notjustintitle,but in real action. Inmy career, I’ve been actingVPofHR,CFO, andVPofsales.OftenCEOsresistactinginfunctionalroles,becausetheyworrythattheylacktheappropriateknowledge.Thisworryispreciselywhyyoushouldact—togettheappropriateknowledge.Indeed,actingisreallytheonlywaytogetalltheknowledgethatyouneedtomakethehire,becauseyouarelookingfortherightexecutiveforyourcompanytoday,notagenericexecutive.Inadditiontoactingintherole,ithelpsgreatlytobringindomainexperts.If

youknowagreatheadofsales, interviewthemfirstand learnwhat they thinkmade themgreat. Figure outwhich of those strengthsmost directlymatch theneedsofyourcompany.Ifpossible, includethedomainexpertintheinterviewprocess. However, be aware that the domain expert only has part of theknowledge necessary to make the hire. Specifically, she has very littleknowledgeofyourcompany,how itworks,andwhat itsneedsare.Therefore,youcannotdeferthedecisiontothedomainexpert.Finally, be clear in your ownmind about your expectations for this person

upon joining your company.Whatwill this person do in the first thirty days?What do you expect theirmotivation to be for joining?Do youwant them tobuild a large organization right away or hire only one or two people over thenextyear?

STEP2.RUNAPROCESSTHATFIGURESOUTTHERIGHTMATCH

Inordertofindtherightexecutive,youmustnowtaketheknowledgethatyouhavegatheredandtranslateitintoaprocessthatyieldstherightcandidate.HereistheprocessthatIliketouse.

Writedownthestrengthsyouwantandtheweaknessesthatyouarewillingtotolerate.In order to ensure completeness, I find it useful to include criteria from thefollowingsubdivisionswhenhiringexecutives:

Willtheexecutivebeworld-classatrunningthefunction?Istheexecutiveoutstandingoperationally?Willtheexecutivemakeamajorcontributiontothestrategicdirectionofthecompany?Thisisthe“aretheysmartenough?”criterion.Willtheexecutivebeaneffectivememberoftheteam?Effectiveisthekeyword.It’spossibleforanexecutivetobewelllikedandtotallyineffectivewithrespecttotheothermembersoftheteam.It’salsopossibleforanexecutivetobehighlyeffectiveandprofoundlyinfluentialwhilebeingtotallydespised.Thelatterisfarbetter.

These functions do not carry equalweight for all positions.Make sure thatyou balance them appropriately. Generally, operational excellence is farmoreimportantforaVPofengineeringoraVPofsalesthanforaVPofmarketingoraCFO.

Developquestionsthattestforthecriteria(seetheappendix).This effort is important even if you never ask the candidate any of the pre-preparedquestions.Bywritingdownquestionsthattestforwhatyouwant,youwill get to a level of specificity that will be extremely difficult to achieveotherwise. (See the appendix for an example of questions that I wrote forrunning the enterprise sales function andoperational excellence.)Assemble anappropriateinterviewteamandconducttheinterviews.

Assembletheinterviewteam.

Inassemblingtheteam,youshouldkeeptwoquestionsinmind:

1.Whowillbesthelpyoufigureoutwhetherthecandidatemeetsthecriteria?Thesemaybeinternalorexternalpeople.Theycanbeboardmembers,otherexecutives,orjustexperts.2.Whodoyouneedtosupportthedecisiononcetheexecutiveisonboard?Thisgroupisjustasimportantasthefirst.Nomatterhowgreatanexecutiveis,theywillhavetroublesucceedingifthepeoplearoundthemsabotageeverythingtheydo.Thebestwaytoavoidthatistounderstandanypotentialissuesbeforethepersonishired.

Clearly,somepeoplewillbeinbothgroupsoneandtwo.Theopinionsofbothgroups will be very important: Group one will help you determine the bestcandidate and group twowill help you gauge how easily each candidate willintegrate into your company.Generally, it’s best to have group two interviewfinalistcandidatesonly.Next, assign questions to interviewers based on their talents. Specifically,

makesurethattheinterviewerwhoasksthequestionsdeeplyunderstandswhatagoodanswerwillsoundlike.As you conduct the interviews, be sure to discuss each interview with the

interviewer.Usethistimetodrivetoacommonunderstandingofthecriteria,sothatyouwillgetthebestinformationpossible.

Backdoorandfront-doorreferences.For the final candidates, it’s critically important that the CEO conduct thereference checks herself. The references need to be checked against the samehiring criteria that you tested for during the interview process. Backdoorreference checks (checks from people who know the candidate, but were notreferred by the candidate) can be an extremely usefulway to get an unbiasedview. However, do not discount the front-door references.While they clearlyhavecommittedtogivingapositivereference(ortheywouldn’tbeonthelist),youarenot lookingforpositiveornegativewith them.Youare lookingforfitwithyourcriteria.Often,thefront-doorreferenceswillknowthecandidatebestandwillbequitehelpfulinthisrespect.

STEP3:MAKEALONELYDECISION

Despitemanypeoplebeinginvolvedintheprocess,theultimatedecisionshouldbemadesolo.OnlytheCEOhascomprehensiveknowledgeof thecriteria, therationale for the criteria, all of the feedback from interviewers and references,and the relative importance of the various stakeholders. Consensus decisionsaboutexecutivesalmostalwaysswaytheprocessawayfromstrengthandtowardlackofweakness.It’salonelyjob,butsomebodyhastodoit.

WHENEMPLOYEESMISINTERPRETMANAGERS

EarlyonatLoudcloud,manypeoplewoulddocrazythingsbackedupby“Bensaid.”OftenIdidn’tsayanyofit,butIdefinitelydidn’tsayit inthewaytheyusedit.ThemanagementprinciplesIsharehereareconnectedtomanyofthoseexperiences.When I ran Opsware, we had the nonlinear quarter problem also known

affectionately as the hockey stick.The hockey stick refers to the shape of therevenuegraphoverthecourseofaquarter.Ourhockeystickwassobadthatonequarter, we booked 90 percent of our new bookings on the last day of thequarter. Sales patterns like this make it difficult to plan the business and areparticularlyharrowingwhenyouare,aswewere,apubliccompany.Naturally,Iwasdeterminedtostraightenoutthehockeystickandbringsome

sanity to thebusiness. Idesignedan incentiveforsalespeople toclosedeals inthefirsttwomonthsofthequarterbyissuingbonusesfordealsinthosemonths.As a result, the next quarter became slightlymore linear, and slightly smallerthananticipated—dealsjustmovedfromthethirdmonthtothefirst twoof thefollowingquarter.When I ran a large engineering group at Netscape, I measured one of our

engineering products on schedule, quality, and features. The team shipped aproduct with all the required features, on time and with very few bugs.Unfortunately,theproductwasmediocre,becausenoneofthefeatureswerethatgreat.WhenIwasatHP,weranallthebusinessesbythenumberswithextremely

strictrevenueandmargintargets.Somedivisionsmadetheirnumbers,butdidsobyunderfundingR&D.Theydramaticallyweakenedtheirlong-termcompetitivepositionandsetthemselvesupforfuturedisaster.Inallthreecases,managersgotwhatweaskedfor,butnotwhatwewanted.

Howdidthishappen?Let’stakealook.

FLATTENINGOUTTHEHOCKEYSTICK:THEWRONGGOAL

Inretrospect,Ishouldneverhaveaskedtheteamtoflattenthequarters.Ifthatiswhat I wanted, I had to be willing to—at least temporarily—accept smallerquarters.Wehadafixednumberofsalespeoplewhoweremaximizingthesizeof each quarter. In order to deliver linear quarters, they had to modify theirbehavior and adjust their priorities. Unfortunately, I liked the old priority ofmaximizingrevenuebetter.Giventhesituation,Iwasactuallyprettylucky.SunTzu,inhisclassicwork

TheArtofWar,warnsthatgivingtheteamataskthatitcannotpossiblyperformiscalledcripplingthearmy.Inmycase,Ididnotcrippletheteam,butIscrewedup my priorities. The right thing to do would have been to make the harddecisionupfront,aboutwhatwasmore important,maximizingeachquarterorincreasingpredictability.Theinstructiononlymadesenseiftheanswerwasthesecondone.

FOCUSINGTOOMUCHONTHENUMBERS

Inthesecondexample,ImanagedtheteamtoasetofnumbersthatdidnotfullycapturewhatIwanted.Iwantedagreatproductthatcustomerswouldlovewithhighqualityandontime—inthatorder.Unfortunately,themetricsthatIsetdidnotcapturethosepriorities.Atabasic

level,metrics are incentives.Bymeasuringquality, features, and schedule anddiscussing them at every staffmeeting,my people focused intensely on thosemetrics to the exclusion of other goals. Themetrics did not describe the realgoalsandIdistractedtheteamasaresult.Interestingly, I see this same problem play out in many consumer Internet

startups. I often see teams that maniacally focus on their metrics aroundcustomer acquisition and retention. This usually works well for customeracquisition,butnotsowellforretention.Why?Formany products, metrics often describe the customer acquisition goal in

enough detail to provide sufficient management guidance. In contrast, themetrics for customer retention do not provide enough color to be a completemanagementtool.Asaresult,manyyoungcompaniesoveremphasizeretentionmetrics and do not spend enough time going deep enough on the actual userexperience.Thisgenerallyresultsinafranticnumberschasethatdoesnotendinagreatproduct.It’simportanttosupplementagreatproductvisionwithastrongdiscipline around themetrics, but if you substitutemetrics for product vision,youwillnotgetwhatyouwant.

MANAGINGSTRICTLYBYNUMBERSISLIKEPAINTINGBYNUMBERS

Somethingsthatyouwanttoencouragewillbequantifiable,andsomewillnot.Ifyoureportonthequantitativegoalsandignorethequalitativeones,youwon’tget thequalitativegoals,whichmaybe themost importantones.Managementpurelybynumbersissortoflikepaintingbynumbers—it’sstrictlyforamateurs.AtHP,thecompanywantedhighearningsnowandinthefuture.Byfocusing

entirelyonthenumbers,HPgotthemnowbysacrificingthefuture.Note that there were many numbers as well as more qualitative goals that

wouldhavehelped:

Wasourcompetitivewinrateincreasingordeclining?Wascustomersatisfactionrisingorfalling?Whatdidourownengineersthinkoftheproducts?

Bymanagingtheorganizationasthoughitwereablackbox,somedivisionsatHPoptimized thepresent at the expenseof their downstreamcompetitiveness.The company rewarded managers for achieving short-term objectives in amanner thatwas bad for the company. Itwould have been better to take intoaccount thewhite box.Thewhite boxgoesbeyond thenumbers andgets intohowtheorganizationproducedthenumbers.Itpenalizesmanagerswhosacrificethefuturefortheshorttermandrewardsthosewhoinvestinthefutureevenifthatinvestmentcannotbeeasilymeasured.

CLOSINGTHOUGHT

Itiseasytoseethattherearemanywaysforleaderstobemisinterpreted.Togetthingsright,youmustrecognizethatanythingyoumeasureautomaticallycreatesasetofemployeebehaviors.Onceyoudeterminetheresultyouwant,youneedto test the description of the result against the employee behaviors that thedescriptionwilllikelycreate.Otherwise,theside-effectbehaviorsmaybeworsethanthesituationyouweretryingtofix.

MANAGEMENTDEBT

ThankstoWardCunningham,thecomputerprogrammerwhodesignedthefirstwiki, themetaphor “technical debt” is now a well-understood concept.Whileyou may be able to borrow time by writing quick and dirty code, you willeventuallyhavetopayitback—withinterest.Oftenthistrade-offmakessense,butyouwillrunintoserioustroubleifyoufailtokeepthetrade-offinthefrontofyourmind.Therealsoexistsalessunderstoodparallelconcept,whichIwillcallmanagementdebt.Like technical debt, management debt is incurred when you make an

expedient, short-term management decision with an expensive, long-termconsequence. Like technical debt, the trade-off sometimes makes sense, butoften does not. More important, if you incur the management debt withoutaccountingforit,thenyouwilleventuallygomanagementbankrupt.Like technicaldebt,managementdebtcomes in toomanydifferent forms to

elaborateentirely,butafewsalientexampleswillhelpexplaintheconcept.Herearethreeofthemorepopulartypesamongstartups:

1.Puttingtwointhebox2.Overcompensatingakeyemployee,becauseshegetsanotherjoboffer3.Noperformancemanagementoremployeefeedbackprocess

PUTTINGTWOINTHEBOX

Whatdoyoudowhenyouhavetwooutstandingemployeeswhologicallybothfitinthesameplaceontheorganizationalchart?Perhapsyouhaveaworld-classarchitectwho is running engineering, but she does not have the experience toscaletheorganizationtothenextlevel.Youalsohaveanoutstandingoperationalpersonwhoisnotgreattechnically.Youwanttokeepbothinthecompany,butyouonlyhaveoneposition.Soyouget thebright ideatoput“twointhebox”andtakeonalittlemanagementdebt.Theshort-termbenefitsareclear:youkeepbothemployees,youdon’thavetodevelopeitherbecausetheywilltheoretically

helpeachotherdevelop,andyouinstantlyclosetheskillsetgap.Unfortunately,youwillpayforthosebenefitswithinterestandataveryhighrate.Forstarters,bydoingthisyouwillmakeeveryengineer’sjobmoredifficult.

Ifanengineerneedsadecisionmade,whichbossshouldshegoto?Ifthatbossdecides,willtheotherbossbeabletooverrideit?Ifit’sacomplexdecisionthatrequiresameeting,doesshehavetoschedulebothheadsofengineeringforthemeeting?Whosetsthedirectionfortheorganization?Willthedirectionactuallygetsetifdoingsorequiresaseriesofmeetings?In addition, you have removed all accountability. If schedules slip, who is

accountable? If engineering throughput becomes uncompetitive, who isresponsible?Iftheoperationalheadisresponsibleforthescheduleslipandthetechnical head is responsible for throughput, what happens if the operationalhead thrashes the engineers to make the schedule and kills throughput? Howwouldyouknowthatshedidthat?Thereallyexpensivepartaboutbothofthesethingsisthattheytendtogetworseovertime.Intheveryshorttermyoumightmitigatetheireffectswithextrameetingsorbyattemptingtocarveupthejobinaclearway.However,asthingsgetbusy,thoseonce-clearlineswillfadeandtheorganization will degenerate. Eventually, you’ll either make a lump-sumpayment by making the hard decision and putting one in the box or yourengineeringorganizationwillsuckforever.

OVERCOMPENSATINGAKEYEMPLOYEEBECAUSESHEGETSANOTHERJOBOFFER

An excellent engineer decides to leave the company because she gets a betteroffer.Forvariousreasons,youwereundercompensatingher,buttheofferfromthe other company pays more than any engineer in your company and theengineerinquestionisnotyourbestengineer.Still,sheisworkingonacriticalprojectandyoucannotaffordtoloseher.Soyoumatchtheoffer.Yousavetheproject,butyoupileonthedebt.Here’s how the payment will come due. You probably think that your

counteroffer was confidential because you’d sworn her to secrecy. Let meexplainwhyitwasnot.Shehasfriendsinthecompany.Whenshegottheofferfromtheothercompany,sheconsultedwithherfriends.Oneofherbestfriendsadvisedhertotaketheoffer.Whenshedecidedtostay,shehadtoexplaintohimwhyshedisregardedhisadviceorelselosepersonalcredibility.Soshetoldhimandsworehimtosecrecy.Heagreedtohonorthesecretbutwasincensedthat

shehad to threaten toquit inorder toget aproper raise.Furthermore, hewasfurious that you overcompensated her. So he told the story to a few of hisfriends,butkepthernameconfidentialtopreservethesecret.Andnoweveryoneinengineeringknowsthatthebestwaytogetaraiseistogenerateanofferfromanothercompanyandthenthreatentoquit.It’sgoingtotakeawhiletopayoffthatdebt.

NOPERFORMANCEMANAGEMENTOREMPLOYEEFEEDBACKPROCESS

Yourcompanynowemploystwenty-fivepeopleandyouknowthatyoushouldformalize theperformancemanagementprocess,butyoudon’twant topaytheprice. You worry that doing so will make it feel like a “big company.”Moreover, you do not want your employees to be offended by the feedback,because you can’t afford to lose anyone right now.And people are happy, sowhyrocktheboat?Whynottakeonalittlemanagementdebt?The first noticeable payments will be due when somebody performs below

expectations:CEO:“Hewasgoodwhenwehiredhim;whathappened?”Manager:“He’snotdoingthethingsthatweneedhimtodo.”CEO:“Didweclearlytellhimthat?”Manager:“Maybenotclearly...”However, the larger payment will be a silent tax. Companies execute well

wheneverybodyisonthesamepageandeverybodyisconstantlyimproving.Inavacuumoffeedback,thereisalmostnochancethatyourcompanywillperformoptimally across either dimension. Directions with no corrections will seemfuzzy and obtuse. People rarely improve weakness they are unaware of. Theultimate price you will pay for not giving feedback: systematically crappycompanyperformance.

INTHEEND

Every really good, really experienced CEO I know shares one importantcharacteristic:They tend toopt for thehardanswer toorganizational issues. Iffacedwithgivingeveryonethesamebonustomakethingseasyorwithsharplyrewardingperformanceandrufflingmanyfeathers,they’llrufflethefeathers.Ifgiventhechoiceofcuttingapopularprojecttoday,becauseit’snotinthelong-term plans or you’re keeping it around for morale purposes and to appear

consistent, they’ll cut it today. Why? Because they’ve paid the price ofmanagementdebt,andtheywouldrathernotdothatagain.

MANAGEMENTQUALITYASSURANCE

Everyoneinthetechnologyindustryseemstoagreethatpeopleareparamount,yetnobodyseemstobeonthesamepagewithwhatthepeopleorganization—humanresources—shouldlooklike.TheproblemisthatwhenitcomestoHR,mostCEOsdon’treallyknowwhat

theywant. In theory, theywantawell-managedcompanywithagreatculture.InstinctivelytheyknowthatanHRorganizationprobablycan’tdeliverthat.Asaresult, CEOs usually punt on the issue and implement something that’ssuboptimal,ifnotworthless.Ironically one of the first things you learn when you run an engineering

organization is thatagoodqualityassuranceorganizationcannotbuildahigh-quality product, but it can tell you when the development team builds a lowqualityproduct.Similarly, ahighqualityhuman resourcesorganizationcannotmakeyouawell-managedcompanywithagreatculture,butitcantellyouwhenyouandyourmanagersarenotgettingthejobdone.

THEEMPLOYEELIFECYCLE

Thebestwaytoapproachmanagementqualityassuranceisthroughthelensoftheemployeelifecycle.Fromhiretoretire,howgoodisyourcompany?Isyourmanagementteamworld-classinallphases?Howdoyouknow?A great HR organization will support, measure, and help improve your

managementteam.Someofthequestionstheywillhelpyouanswer:RecruitingandHiring

Doyousharplyunderstandtheskillsandtalentsrequiredtosucceedineveryopenposition?Areyourinterviewerswellprepared?Doyourmanagersandemployeesdoaneffectivejobofsellingyourcompanytoprospectiveemployees?Dointerviewersarriveontime?

Domanagersandrecruitersfollowupwithcandidatesinatimelyfashion?Doyoucompeteeffectivelyfortalentagainstthebestcompanies?

CompensationDoyourbenefitsmakesenseforyourcompanydemographics?Howdoyoursalaryandstockoptionpackagescomparewiththecompaniesthatyoucompetewithfortalent?Howwelldoyourperformancerankingscorrespondtoyourcompensationpractices?

TrainingandIntegrationWhenyouhireanemployee,howlongdoesittakethemtobecomeproductivefromtheperspectiveoftheemployee,herpeers,andhermanager?Shortlyafterjoining,howwelldoesanemployeeunderstandwhat’sexpectedofher?

PerformanceManagementDoyourmanagersgiveconsistent,clearfeedbacktotheiremployees?Whatisthequalityofyourcompany’swrittenperformancereviews?Didallofyouremployeesreceivetheirreviewsontime?Doyoueffectivelymanageoutpoorperformers?

MotivationAreyouremployeesexcitedtocometowork?Doyouremployeesbelieveinthemissionofthecompany?Dotheyenjoycomingtoworkeveryday?Doyouhaveanyemployeeswhoareactivelydisengaged?Doyouremployeesclearlyunderstandwhat’sexpectedofthem?Doemployeesstayalongtimeordotheyquitfasterthannormal?Whydoemployeesquit?

REQUIREMENTSTOBEGREATATRUNNINGHR

Whatkindofpersonshouldyoulookfor tocomprehensivelyandcontinuouslyunderstand the quality of your management team? Here are some keyrequirements:

World-classprocessdesignskillsMuchliketheheadofqualityassurance,theheadofHRmustbeamasterfulprocessdesigner.Onekeytoaccuratelymeasuringcriticalmanagementprocessesisexcellentprocessdesignandcontrol.AtruediplomatNobodylikesatattletaleandthereisnowayforanHRorganizationtobeeffectiveifthemanagementteamdoesn’timplicitlytrustit.ManagersmustbelievethatHRistheretohelpthemimproveratherthanpolicethem.GreatHRleadersgenuinelywanttohelpthemanagersandcouldn’tcarelessaboutgettingcreditforidentifyingproblems.TheywillworkdirectlywiththemanagerstogetqualityupandonlyescalatetotheCEOwhennecessary.IfanHRleaderhoardsknowledge,makespowerplays,orplayspolitics,hewillbeuseless.IndustryknowledgeCompensation,benefits,bestrecruitingpractices,etc.areallfast-movingtargets.TheheadofHRmustbedeeplynetworkedintheindustryandstayabreastofallthelatestdevelopments.IntellectualhefttobetheCEO’strustedadviserNoneoftheotherskillsmatteriftheCEOdoesnotfullybacktheheadofHRinholdingthemanagerstoahighqualitystandard.Inorderforthistohappen,theCEOmusttrusttheHRleader’sthinkingandjudgment.

UnderstandingthingsunspokenWhenmanagementqualitystartstobreakdowninacompany,nobodysaysanythingaboutit,butsuper-perceptivepeoplecantellthatthecompanyisslipping.Youneedoneofthose.

—CHAPTER6—

CONCERNINGTHEGOINGCONCERN

“Thisain’tfornofuckniggasIfyouarealniggathenfuckwithme.”

—TRINIDADJAMES,“ALLGOLDEVERYTHING”

Oneday inastaffmeeting in theLoudcloud/Opswaredays, someonebroughtupan issue thathadbeenbotheringhimfor some time.“Thisplace isentirelytooprofane.It’smakingmanyoftheemployeesuncomfortable.”Otherschimedin: “It makes the environment unprofessional. We need to put a stop to it.”Althoughthecomplaintswereabstract,theywereclearlydirectedatmesinceIwasthebiggestabuserofprofanityinthecompanyandperhapsintheindustry.Inthosedays,Idirectedtheteamwithsuchurgencythatitwasrareformetosaymorethanafewsentenceswithoutanexpletiveinjectedsomewhere.Partofitwasintentional.Ionlyhadsomuchtimewitheachemployeeandit

wascriticallyimportantthatIbecrystalclearinthosemoments.Nothingmakesthingsclearlikeafewchoicecursewords.“Thatisnotthepriority”isradicallyweaker than “That is not the fucking priority.” When the CEO drops the F-bomb, it gets repeated. And that’s good if you want your message to spreadthroughout the company. (On the other hand, it’s extremely bad if you don’twantyouremployeestalkinglikeabunchofgangstarappers.)Butpartofitwasalsounintentional.Atthispoint,Icouldbarelycontrolmyself.Thiswasnotaneasy company to run, and I’d developed CEO Tourette’s syndrome—theprofanitywasinvoluntary.Sincethecomplaintsseemedbroadanddeep,Ihadtotakethemseriously.I

thoughthardaboutitthatnightandconsideredthefollowing.

Inthetechnologybusiness,someemployeeswouldbecomfortablewithprofanitywhileotherswouldnot.

Ifweoutlawedprofanity,thensomeemployeeswhouseditwouldnotcometoworkforusorquitoncetheygottherebecausewewouldseemold-fashionedandprudish.Ifwekeptprofanity,somepeoplemightleave.Myjudgmentwasbiased,becauseIwasthemainoffender.

Aftermuchconsideration,Irealizedthatthebesttechnologycompaniesoftheday,IntelandMicrosoft,wereknowntobehighlyprofaneplaces,sowe’dbeoffculturewith themand the restof themodern industry ifwestoppedprofanity.Obviously, that didn’t mean that we had to encourage it, but prohibiting itseemed both unrealistic and counterproductive. Attracting the very bestengineersmeant recruiting fromhighlyprofane environments.The choicewasbetweenoptimizingfortoptalentorcleanculture.Easydecision.Idecidedtokeepthecursing,butIalsoneededtomakeastatement.People

hadcomplainedandhadrunthisissueallthewayuptheorganization,andtheydeserved an explanation.Explaining thingswould be tricky, because profanitydidnotwork in all contexts.We certainly couldnot tolerate profanity used tointimidate or sexually harass employees, so I needed to make the distinctionclear.Approvingprofanityonlyincertaincontextswasatrickymessagetocraft.ThatnightIwatchedadisturbingmoviefromthelate1970scalledShortEyes,

which told the graphic story of a child molester who went to prison andconfronted the one clear prison ethic: Child molesters must die. One of thecharacters in themoviewas a youngman referred to by the other inmates as“Cupcakes.”Hardtobelieve,butwatchingthatmovie,Ifoundmysolution.Thenextday,Igavethefollowingspeechattheall-companymeeting:“It has come tomy attention that many people are uncomfortable with the

amountofprofanitythatweuse.Beingthenumber-oneabuser,thesecomplaintshave causedme to reflect on my own behavior as well as the company as awhole.AsIseeit,wehavetwochoices:(a)wecanbanprofanityor(b)wecanaccept profanity. Anything in between is very unlikely to work. ‘Minimalprofanity’ cannot be enforced. I’ve said before that we cannot win unless weattract the very best people in the world. In the technology industry, almosteverybody comes from a culture that allows profanity. Therefore, banningprofanitywill likely limit our talent poolmore than accepting profanity.As aresult,wewill allowprofanity.However, this doesnotmean that you canuseprofanity to intimidate, sexually harass people, or do other bad things. In this

way, profanity is no different from other language. For example, consider theword‘cupcakes.’It’sfineformetosaytoShannon,‘Thosecupcakesyoubakedlookdelicious.’But it is not okay forme to say toAnthony, ‘Hey,Cupcakes,youlookmightyfineinthemjeans.’”AndthatwasallIsaidaboutthat.After that day, I never heard another complaint about profanity and I don’t

thinkwelostanybodybecauseofthepolicy.Sometimesanorganizationdoesn’tneedasolution;itjustneedsclarity.OnceImadeitclearthatcursingwasokay—solongasitwasn’tusedtointimidateorharass—nobodyhadaproblemwithitanymore.AtleastasfarasIknew.Bottomline,theresultsofthepolicyweregood: a comfortable work environment, low attrition, and no complaining.SometimestherightpolicyistheonethattheCEOcanfollow.Asacompanygrows,itwillchange.Nomatterhowwellyousetyourculture,keep your spirit, or slow-roll your growth, your company won’t be the samewhenit’sonethousandpeopleasitwaswhenitwastenpeople.Butthatdoesn’tmean that it can’t be a good companywhen it reaches 1,000, 10,000, or even100,000 employees. It will just be different. Making it good at scale meansadmittingthatitmustbedifferentandembracingthechangesthatyou’llneedtomake to keep things from falling apart. This chapter explains some of thechangesthatyouwillneedtomake.

HOWTOMINIMIZEPOLITICSINYOURCOMPANY

In allmyyears inbusiness, I haveyet tohear someone say, “I lovecorporatepolitics.”Ontheotherhand,Imeetplentyofpeoplewhocomplainbitterlyaboutcorporate politics—sometimes even in the companies they run. So, if nobodylovespolitics,whyallthepolitics?Political behavior almost always starts with the CEO. Now you may be

thinking,“Ihatepolitics,I’mnotpolitical,butmyorganizationisverypolitical.I clearly didn’t cause this.” Sadly, you needn’t be political to create extremepoliticalbehaviorinyourorganization.Infact,it’softentheleastpoliticalCEOswhorunthemostferociouslypoliticalorganizations.ApoliticalCEOsfrequently—andaccidentally—encourageintensepoliticalbehavior.WhatdoImeanbypolitics?Imeanpeopleadvancingtheircareersoragendas

bymeansotherthanmeritandcontribution.Theremaybeothertypesofpolitics,butpoliticsofthisformseemtobetheonesthatreallybotherpeople.

HOWITHAPPENS

A CEO creates politics by encouraging and sometimes incentivizing politicalbehavior—often unintentionally. As a very simple example, let’s considerexecutivecompensation.WhenyouareCEO,senioremployeeswillcometoyoufrom time to timeandask foran increase incompensation.Theymaysuggestthatyouarepayingthemfarlessthantheircurrentmarketvalue.Theymayevenhaveacompetitiveofferinhand.Facedwiththisconfrontation,iftherequestisreasonable, you might investigate the situation. You might even give theemployeearaise.Thismaysound innocent,butyouhave justcreatedastrongincentiveforpoliticalbehavior.Specifically, you will be rewarding behavior that has nothing to do with

advancingyourbusiness.Theemployeewillearnaraisebyaskingforoneratherthanasaresultofyourrewardingthemforoutstandingperformance.Whyisthisbad?Letmecounttheways:

1.Theotherambitiousmembersofyourstaffwillimmediatelygetthepointandagitateforraisesaswell.Wordalwaysgetsout.Notethatneitherthiscampaignnortheprioroneneedbecorrelatedwithactualperformance.Youwillnowspendtimedealingwiththepoliticalissuesratherthanactualperformanceissues.Importantly,ifyouhaveacompetentboard,youwillnotbeabletogivethemallout-of-cycleraises,soyourcompanyexecutiveraiseswilloccuronafirst-come,first-servedbasis.2.Thelessaggressive(butperhapsmorecompetent)membersofyourteamwillbedeniedoff-cycleraisessimplybybeingapolitical.3.Theobjectlessonforyourstaffandthecompanywillbethatthesqueakywheelgetsthegrease,andthatthemostpoliticallyastuteemployeesgettheraises.Getreadyforawholelotofsqueakywheels.

Nowlet’smoveontoamorecomplicatedexample.YourCFOcomestoyouand says that hewants to continue developing as amanager.He says that hewouldliketoeventuallybecomeaCOOandwouldliketoknowwhatskillshemust demonstrate in order to earn that position in your company. Being apositiveleader,youwouldliketoencouragehimtopursuehisdream.Youtellhimthatyouthinkthathe’dbeafineCOOsomedayandthatheshouldworktodevelopafewmoreskills.Inaddition,youtellhimthathe’llneedtobeastrongenoughleader,suchthatotherexecutivesinthecompanywillwanttoworkforhim.Aweek later,oneofyourother executivescomes toyou inapanic.ShesaysthattheCFOjustaskedherifshe’dworkforhim.ShesaysthathesaidyouaregroominghimtobetheCOOandthat’shisfinalstep.Didthatjusthappen?Welcometothebigtime.

HOWTOMINIMIZEPOLITICS

Minimizing politics often feels totally unnatural. It’s counter to excellentmanagement practices such as being open-minded and encouraging employeedevelopment.The difference between managing executives and managing more junior

employees can be thought of as the difference between being in a fight withsomeonewithnotrainingandbeinginaringwithaprofessionalboxer.Ifyouare in a fight with a regular person, then you can do natural things and theywon’t get you into much trouble. For example, if you want to take a stepbackward, you can pick your front foot up first. If you do this against a

professional boxer, you will get your block knocked off. Professional boxerstrainforyearstotakeadvantageofsmallerrorsintechnique.Liftingyourfrontfoot first to take a stepbackwardwill takeyou slightly off balance for a splitsecondandthat’sallyouropponentwillneed.Similarly, if you manage a junior employee and they ask you about their

careerdevelopment,youcansaywhatcomesnaturallyandgenerallygetawaywithit.Aswesawabove,thingschangewhenyoudealwithhighlyambitious,seasonedprofessionals.Inordertokeepfromgettingknockedoutbycorporatepolitics,youneedtorefineyourtechnique.

THETECHNIQUE

AsIdevelopedasaCEO,Ifoundtwokeytechniquestobeusefulinminimizingpolitics.

1.Hirepeoplewiththerightkindofambition.ThecasesthatIdescribedabovemightinvolvepeoplewhoareambitiousbutnotnecessarilyinherentlypolitical.Allcasesarenotlikethis.ThesurestwaytoturnyourcompanyintothepoliticalequivalentoftheU.S.Senateistohirepeoplewiththewrongkindofambition.AsdefinedbyAndyGrove,therightkindofambitionisambitionforthecompany’ssuccesswiththeexecutive’sownsuccessonlycomingasaby-productofthecompany’svictory.Thewrongkindofambitionisambitionfortheexecutive’spersonalsuccessregardlessofthecompany’soutcome.2.Buildstrictprocessesforpotentiallypoliticalissuesanddonotdeviate.Certainactivitiesattractpoliticalbehavior.Theseactivitiesinclude:PerformanceevaluationandcompensationOrganizationaldesignandterritoryPromotions

Let’sexamineeachcaseandhowyoumightbuildandexecuteaprocessthatinsulatesthecompanyfrombadbehaviorandpoliticallymotivatedoutcomes.PerformanceevaluationandcompensationOftencompaniesdeferputting

performance management and compensation processes in place. This doesn’tmeanthattheydon’tevaluateemployeesorgivepayraises;itjustmeanstheydoso in an ad hocmanner that’s highly vulnerable to politicalmachinations.By

conductingwell-structured,regularperformanceandcompensationreviews,youwillensurethatpayandstockincreasesareasfairaspossible.Thisisespeciallyimportantforexecutivecompensation,sincedoingsowillalsoservetominimizepolitics.Intheexampleabove,theCEOshouldhavehadanairtightperformanceand compensation policy and simply told the executive that his compensationwould be evaluatedwith everyone else’s. Ideally, the executive compensationprocess should involve the board of directors. This will help ensure goodgovernanceandmakeexceptionsevenmoredifficult.OrganizationaldesignandterritoryIfyoumanageambitiouspeople,from

time to time they will want to expand their scope of responsibility. In theexample above, theCFOwanted to become theCOO. In other situations, thehead of marketing might want to run sales and marketing or the head ofengineering may want to run engineering and product management. Whensomeoneraisesanissuelikethiswithyou,youmustbeverycarefulaboutwhatyou say, because everything that you say can be turned into political cannonfodder.Generally,it’sbesttosaynothingatall.Atmost,youmightask“why?”butifyoudosobesurenottoreacttothereasons.Ifyouindicatewhatyouarethinking,thatinformationwillleak,rumorswillspread,andyouplanttheseedsfor all kinds of unproductive discussions. You should evaluate yourorganizationaldesignonaregularbasisandgathertheinformationthatyouneedtodecidewithout tippingpeopleoff towhatyouplan todo.Onceyoudecide,you should immediately execute the reorg: Don’t leave time for leaks andlobbying.PromotionsEverytimeyourcompanygivessomeoneapromotion,everyone

else at that person’s organizational level evaluates the promotion and judgeswhethermeritorpoliticalfavorsyieldedit.Ifthelatter,thentheotheremployeesgenerallyreactinoneofthreeways:

1.Theysulkandfeelundervalued.2.Theyoutwardlydisagree,campaignagainsttheperson,andunderminethemintheirnewposition.3.Theyattempttocopythepoliticalbehaviorthatgeneratedtheunwarrantedpromotion.

Clearly,youdon’twantanyof thesebehaviors inyourcompany.Therefore,you must have a formal, visible, defensible promotion process that governsevery employeepromotion.Often this processmust bedifferent for peopleon

your own staff. (The general process may involve various managers who arefamiliar with the employee’s work; the executive process should include theboardofdirectors.)Thepurposeoftheprocessistwofold.First,itwillgivetheorganization confidence that the company at least attempted to base thepromotiononmerit.Second,theprocesswillproducetheinformationnecessaryforyourteamtoexplainthepromotiondecisionsyoumade.Be careful with “he said, she said” Once your organization grows to a

significant size,membersofyour teamwill from time to timecomplain abouteach other. Sometimes this criticismwill be extremely aggressive. Be carefulabouthowyoulistenandthemessagethatitsends.Simplybyhearingthemoutwithoutdefendingtheemployeeinquestion,youwillsendthemessagethatyouagree.Ifpeopleinthecompanythinkthatyouagreethatoneofyourexecutivesis less than stellar, that information will spread quickly and withoutqualification.Asaresult,peoplewillstoplisteningtotheexecutiveinquestionandtheexecutivewillsoonbecomeineffective.Therearetwodistincttypesofcomplaintsthatyouwillreceive:

1.Complaintsaboutanexecutive’sbehavior2.Complaintsaboutanexecutive’scompetencyorperformance

Generally, the best way to handle the first type of complaint is to get thecomplainingexecutiveandthetargetedexecutiveintheroomtogetherandhavethemexplain themselves.Usually, simplyhaving thismeetingwill resolve theconflictandcorrectthebehaviorandimprovetherelationship(ifitwasactuallybroken).Donotattempttoaddressbehavioralissueswithoutbothexecutivesintheroom.Doingsowillinvitemanipulationandpolitics.Complaintsofthesecondtypearebothmorerareandmorecomplex.Ifoneof

yourexecutivessummonsthecouragetocomplainaboutthecompetencyofoneof their peers, then there is a good chance that either the complainer or thetargeted executivehas amajorproblem. If you receive this typeof complaint,youwillgenerallyhaveoneoftworeactions:theywillbetellingyousomethingthatyoualreadyknow,orthey’llbetellingyoushockingnews.Iftheyaretellingyousomethingthatyoualreadyknow,thenthebignewsis

thatyouhaveletthesituationgotoofar.Whateveryourreasonsforattemptingto rehabilitate thewayward executive, you have taken too long and nowyourorganizationhasturnedonthem.Youmustresolvethesituationquickly.Almostalways, this means firing the executive. While I’ve seen executives improve

their performance and skill sets, I’ve never seen one lose the support of theorganizationandthenregainit.Ontheotherhand,ifthecomplaintisnewnews,thenyoumustimmediately

stoptheconversationandmakecleartothecomplainingexecutivethatyouinnowayagreewiththeirassessment.Youdonotwanttocrippletheotherexecutivebefore you reevaluate his performance. You do not want the complaint tobecomeaself-fulfillingprophecy.Onceyou’veshutdowntheconversation,youmust quickly reassess the employee in question. If you find he is doing anexcellentjob,youmustfigureoutthecomplainingexecutive’smotivationsandresolvethem.Donotletanaccusationofthismagnitudefester.Ifyoufindthatthe employee is doing a poor job, there will be time to go back and get thecomplainingemployee’sinput,butyoushouldbeonatracktoremovethepoorperformeratthatpoint.As CEO, you must consider the systemic incentives that result from your

wordsandactions.Whileitmayfeelgoodinthemomenttobeopen,responsive,andactionoriented,becarefulnottoencourageallthewrongthings.

THERIGHTKINDOFAMBITION

Whenhiringamanagementteam,moststartupsfocusalmostexclusivelyonIQ,butabunchofhigh-IQpeoplewith thewrongkindofambitionwon’twork. Ihavealreadystressedthatyoushouldstrivetohirepeoplewiththerightkindofambition.AsI’vetalkedabouttheseideasinthelastfewyears,Ihavereceivedamixedresponse.Somethinkit’sgoodadvice,whileothersquestionit.Atamacro level,acompanywillbemostsuccessful if theseniormanagers

optimize for the company’s success (think of this as a global optimization) asopposedtotheirownpersonalsuccess(localoptimization).NomatterhowwelltheCEOdesignsthepersonalincentiveprograms,theywillneverbeperfect.Inaddition, career incentives likepromotions and territoryownership fall outsidethescopeofbonusplansandotherapriorimanagementtools.Inanequity-basedcompensation structure, optimizing for the company’s success should yieldbetter results for individuals as well. As my Opsware head of sales MarkCranneyusedtosay,“Twopercentofzeroiszero.”It is particularly important that managers have the right kind of ambition,

because anything elsewill be exceptionally demotivating for their employees.Asanemployee,whywouldIwanttoworklonghourstoadvancethecareerofmymanager?Ifthemanagercaresmoreabouthiscareerthanthecompany,thenthat’s what I’d be doing. Nothing motivates a great employee more than amissionthat’ssoimportantthatitsupersedeseveryone’spersonalambition.Asaresult, managers with the right kind of ambition tend to be radically morevaluable than those with the wrong kind. For a complete explanation of thedangersofmanagerswiththewrongkindofambition,IstronglyrecommendDr.Seuss’smanagementmasterpieceYertletheTurtle.

SCREENINGFORTHERIGHTKINDOFAMBITION

Aswithanycomplexcharactertrait,thereisnowaytoperfectlyscreenfortherightkindofambitioninaninterview,buthopefullysomeofthesethoughtswillproveuseful.Atamacrolevel,everybodyviewstheworldthroughherownpersonalprism.

When interviewing candidates, it’s helpful towatch for small distinctions thatindicate whether they view the world through the “me” prism or the “team”prism.Peoplewho view theworld through the “me” prismmight describe a prior

company’sfailureinaninterviewasfollows:“Mylastjobwasmye-commerceplay.Ifeltthatitwasimportanttoroundoutmyrésumé.”Notetheuseofmytopersonalize the company in a way that it’s unlikely that anyone else at thecompanywouldagreewith.Infact, theotheremployeesinthecompanymightevenbeoffendedbythisusage.Peoplewiththerightkindofambitionwouldnotlikely use the word play to describe their effort to work as a team to buildsomethingsubstantial.Finally,peoplewhousethe“me”prismfinditnaturalandobvioustospeakintermsof“buildingoutmyrésumé”whilepeoplewhousethe“team” prism find such phrases to be somewhat uncomfortable and awkward,because they clearly indicate an individual goal that is separate from the teamgoal.Ontheotherhand,peoplewhoviewtheworldpurelythroughtheteamprism

will very seldomuse thewords I orme evenwhenansweringquestions abouttheiraccomplishments.Eveninaninterview,theywilldeflectcredittoothersontheir previous team. They will tend to be far more interested in how yourcompanywillwin than in how theywill be compensated orwhat their careerpathwillbe.Whenaskedaboutapreviouslyfailedcompany,theywillgenerallyfeel such great responsibility that they will describe in detail their ownmisjudgmentsandbaddecisions.Whenwe hired the head ofworldwide sales forOpsware, using this screen

provedtobequitevaluable.Sincethiswasasalesposition,Ishouldmention(inreference to thecommenterabove) thatambition for thecompanyaboveone’sowngoalsisparticularlyimportantfortheheadofsales.Thereasonsaremany:

Thelocalincentivesinsalesareparticularlystronganddifficulttobalancewithouttherightkindofleadership.Thesalesorganizationisthefaceofthecompanytotheoutsideworld.Ifthatgroupoptimizesforitself,yourcompanywillhaveamajorproblem.Inhigh-techcompanies,fraudgenerallystartsinsalesduetomanagersattemptingtoperfecttheultimatelocaloptimization.

Throughoutourinterviewprocess,wemetwithalotofcandidateswhotooksole credit for landing extremely large deals, achieving impressive goals, and

generating company success. Invariably, the candidateswho claimed themostcreditfordealswouldhavethemostdifficulttimedescribingthedetailsofhowthe deal was actually won and orchestrated. During reference checks, othersinvolvedinthedealswouldtellaverydifferentstory.WhenIspoketoMarkCranney,ontheotherhand,itwasdifficulttogethim

todiscusshispersonalaccomplishments.Infact,someoftheotherinterviewersfelt that Mark was standoffish and even obnoxious in the way he bristled atcertainquestions.One interviewercomplained,“Ben, Iknow thathe increasedthesizeoftheNikedealfromonemilliontofivemillion,becauseourcontactatNike told me that, but Mark wouldn’t go into any detail on it.” When IinterviewedMark,hereallyonlywantedtodiscusshowhisoldcompanywon.Hewent intogreatdetailabouthowhis teamdiagnosedweaknessesversusthecompetitionandhowheworkedwithanotherexecutivetoadvancetheproduct.HethentalkedabouthowheworkedwiththeCEOtorevisethewaythesalesforcewastrainedandorganized.When the conversation turned to Opsware, Mark had already interviewed

sales repsatournumber-onecompetitor’scompanyandknewwhatdeals theywerein.Herelentlesslyquestionedmeonhowweweregoingtowinthedealsthattheywereinandhowweplannedtogetintothedealsthatweweren’tin.Hewantedtoknowthestrengthsandweaknessesofeveryoneelseontheteam.Hewanted to know the game plan for winning. The topics of his potentialcompensationandcareeradvancementdidn’tcomeupuntiltheveryendoftheprocess. And then he only wanted assurances that compensation wasperformance-andnotpoliticallybased.ItwasclearthatMarkwasallabouttheteamanditssuccess.During Mark’s tenure, sales increased more than tenfold, and our market

capitalizationincreasedtwentyfold.Moretothepoint,voluntaryattritioninthesales organization was extremely low, customers were managed fairly andhonestly, and our legal and finance teams often commented that first andforemost,Markprotectedthecompany.

FINALTHOUGHT

While itmaywork to have individual employeeswho optimize for their owncareers,countingonseniormanagerstodoalltherightthingsforallthewrongreasonsisadangerousidea.

TITLESANDPROMOTIONS

OftenwhenImeetwithstartups, theemployeeshavenojobtitles.Thismakessense,becauseeverybodyisjustworkingtobuildthecompany.Rolesneedn’tbeclearly defined and, in fact, can’t be, because everyone does a little bit ofeverything. In an environment like this there are no politics and nobody isjockeyingforpositionorauthority.It’srathernice.Sowhydoallorganizationseventuallycreatejobtitlesandwhatistheproperwaytomanagethem?(ThankstoMarkZuckerbergforcontributingtomythinkingonthissubject.)

WHYDOTITLESMATTER?

Twoimportantfactorsdriveallcompaniestoeventuallycreatejobtitles:

1.Employeeswantthem.Whileyoumayplantoworkatyourcompanyforever,atleastsomeofyouremployeesneedtoplanforlifeafteryourcompany.Whenyourheadofsalesinterviewsforhernextjob,shewon’twanttosaythatdespitethefactthatsheranaglobalsalesforcewithhundredsofemployees,hertitlewas“Dude.”2.Eventually,peopleneedtoknowwhoiswho.Ascompaniesgrow,everybodywon’tknoweverybodyelse.Importantly,employeeswon’tknowwhateachpersondoesandwhomtheyshouldworkwithtogettheirjobsdone.Jobtitlesprovideanexcellentshorthandfordescribingrolesinthecompany.Inaddition,customersandbusinesspartnerscanalsomakeuseofthisshorthandtofigureouthowtobestworkwithyourcompany.

Beyond these core reasons, employeeswill use titles to calibrate theirvalueandcompensationagainsttheircolleagues.IfanemployeewithatitleofJuniorEngineerbelievesthatsheisafarbetterprogrammerthanhercounterpartwiththetitleSeniorArchitect,thiswillindicatetoherthatshemaybeunderpaidandundervalued.Becausetitleswillbeusedtocalculaterelativevalue,theymustbemanagedcarefully.

THEDANGERS:THEPETERPRINCIPLEANDTHELAWOFCRAPPYPEOPLE

Thebasicsseemobvious,sowhydoesalmosteverycompanyeventuallymakeseriousmistakes regarding titles? If youhave everworked in a company, youhaveprobablythoughttoyourselfaboutsomeoverlypromotedexecutive:“Howdidhegettobeavicepresident?Iwouldn’tlethimmanagealemonadestand.”One challenge is the Peter Principle. Coined by Dr. Laurence J. Peter and

RaymondHullintheir1969bookofthatname,thePeterPrincipleholdsthatinahierarchy,membersarepromotedsolongastheyworkcompetently.Soonerorlatertheyarepromotedtoapositionatwhichtheyarenolongercompetent(their“level of incompetence”), and there they remain being unable to earn furtherpromotions.AsAndyGrovepointsoutinhismanagementclassicHighOutputManagement, the Peter Principle is unavoidable, because there is no way toknowaprioriatwhatlevelinthehierarchyamanagerwillbeincompetent.AnotherchallengeisaphenomenonthatIcalltheLawofCrappyPeople.The

Law of Crappy People states:For any title level in a large organization, thetalentonthatlevelwilleventuallyconvergetothecrappiestpersonwiththetitle.Therationalebehindthelawisthattheotheremployeesinthecompanywith

lowertitleswillnaturallybenchmarkthemselvesagainstthecrappiestpersonatthenextlevel.Forexample,ifJasperistheworstvicepresidentinthecompany,thenallof thedirectorswillbenchmark themselvesagainstJasperanddemandpromotionsassoonastheyreachhislowlevelofcompetency.AswiththePeterPrinciple,thebestthatyoucandoistomitigatetheLawof

CrappyPeopleand thatmitigationwillbecritically important to thequalityofyourcompany.

PROMOTIONPROCESS

ThebestwaytomitigateboththePeterPrincipleandtheLawofCrappyPeopleiswithaproperlyconstructedandhighlydisciplinedpromotionprocess.Ideally,thepromotionprocessshouldyieldaresultsimilartotheverybestkaratedojos.In top dojos, in order to achieve the next level (for example, being promotedfromabrownbelttoablackbelt),youmustdefeatanopponentincombatatthatlevel. This guarantees that a new black belt is never aworse fighter than theworstcurrentblackbelt.Frustratingly,thereisnoexactanaloguetoafistfightinbusiness,sohowcan

wepreservequalitywithoutactualcombat?

To begin, start with an extremely crisp definition not only of theresponsibilitiesateachlevelbutalsooftheskillrequiredtoperformtheduties.Whendescribingtheskills,avoidthegenericcharacterizationssuchas“mustbecompetentatmanagingaP&L”or“musthaveexcellentmanagementskills.”Infact, the best leveling tools get extremely specific and even name names:“shouldbeasuperstarrecruiter—asgoodasJennyRogers.”Next,defineaformalprocessforallpromotions.Onekeyrequirementofthe

process should be that promotions will be leveled across groups. If you let amanagerorasinglechainofcommanddeterminepromotionsunilaterally, thenit’s possible that, for example, HR will have five vice presidents andEngineering only one. One way to level across groups is to hold a regularpromotions council that reviews every significant promotion in the company.Whenamanagerwishestopromoteanemployee,shewillsubmitthatemployeefor reviewwith an explanationofwhy shebelievesher employee satisfies theskill criteria required for the level. The committee should then compare theemployee with both the level’s skill description and the skills of the otheremployees at that level to determine whether to approve the promotion. Inadditiontoensuringfairnessandlevelquality,thisprocesswillservetoeducateyour entire management team on the skills and accomplishments of theemployeesbeingsubmittedforpromotion.

ANDREESSENVS.ZUCKERBERG:HOWBIGSHOULDTHETITLESBE?

Should your company make Vice President the top title or should you haveChiefMarketingOfficers,ChiefRevenueOfficers,Chief PeopleOfficers, andChief Snack Officers? There are two schools of thought regarding this, onerepresentedbyMarcAndreessenandtheotherbyMarkZuckerberg.Andreessenargues thatpeopleask formany things fromacompany: salary,

bonus, stock options, span of control, and titles. Of those, title is by far thecheapest, so it makes sense to give the highest titles possible. The hierarchyshould have Presidents, Chiefs, and Senior Executive Vice Presidents. If itmakes people feel better, let them feel better. Titles cost nothing. Better yet,whencompetingfornewemployeeswithothercompanies,usingAndreessen’smethodyoucanalwaysoutbidthecompetitioninatleastonedimension.AtFacebook,bycontrast,MarkZuckerbergpurposelydeploys titles thatare

significantly lower than the industry standard. SeniorVice Presidents at other

companiesmusttaketitlehaircutsdowntoDirectorsorManagersatFacebook.Why does he do this? First, he guarantees that every new employee getsreleveledastheyenterhiscompany.Inthisway,heavoidsaccidentallygivingnew employees higher titles and positions than better-performing existingemployees.Thisboostsmoraleand increases fairness.Second, it forcesall themanagersofFacebooktounderstandandinternalizeFacebook’slevelingsystem,which serves the company extremely well in their own promotion andcompensationprocesses.He also wants titles to bemeaningful and reflect who has influence in the

organization. As a company grows quickly, it’s important to provideorganizationalclaritywhereverpossibleandthatgetsmoredifficultiftherearefiftyVPsandtenChiefs.Next, he finds that businesspeople often carry inflated titles versus their

engineering counterparts. While he recognizes that big titles help them outexternallywithgettingmeetings,hestillwantstohaveanorganizationwheretheproductpeopleandengineersformtheculturalcore,sohestrivestokeepthisincheckaswell.Does Facebook ever miss out on a new hire due to its low titles? Yes,

definitely.But onemight argue that theymiss out on precisely the employeestheydon’twant.Infact,boththehiringandonboardingprocessesatFacebookhavebeencarefullydesignedtoencouragetherightkindofemployeestoselectthemselvesinandthewrongonestoselectthemselvesout.Sowhichmethodisbetter,Andreessen’sorZuckerberg’s?Theansweristhat

itdepends.Facebookhassomanyadvantagesinrecruitingemployeesthatbeingdisciplinedaboutabsolutetitlelevelsdoesnotsignificantlyimpairitsabilitytoattract theverybest talent.Yourcompanymightnothavetheseadvantages,soloftytitlesmaybeagoodtactic.Ineitherscenario,youshouldstillrunahighlydisciplinedinternallevelingandpromotionprocess.

FINALTHOUGHT

Youmight think that somuch time spent on promotions and titles places toomuchimportanceandfocusonsillyformalisms.Theoppositeistrue.Withoutawellthoughtout,disciplinedprocessfortitlesandpromotions,youremployeeswill become obsessed with the resulting inequities. If you structure thingsproperly,nobodyotherthanyouwillspendmuchtimethinkingabouttitlesotherthanEmployeeoftheMonth.

WHENSMARTPEOPLEAREBADEMPLOYEES

Inbusiness, intelligence is alwaysa critical element in anyemployee,becausewhat we do is difficult and complex and the competitors are filled withextremelysmartpeople.However,intelligenceisnottheonlyimportantquality.Beingeffectiveinacompanyalsomeansworkinghard,beingreliable,andbeinganexcellentmemberoftheteam.WhenIwasaCEO,thiswasoneofthemostdifficultlessonsformetolearn.

I felt that itwasmy job tocreateanenvironmentwherebrilliantpeopleofallbackgrounds,personalitytypes,andworkstyleswouldthrive.AndIwasright.Thatwasmyjob.Companieswherepeoplewithdiversebackgroundsandworkstyles can succeed have significant advantages in recruiting and retaining toptalentoverthosethatdon’t.Still,youcantakeittoofar.AndIdid.Here are three examples of the smartest people in the company being the

worstemployees.

EXAMPLE1:THEHERETIC

Anysizablecompanyproducessomenumberof strategies,projects,processes,promotions, and other activities that don’t make sense. No large organizationachievesperfection.Asaresult,acompanyneedslotsofsmart,super-engagedemployeeswhocanidentifyitsparticularweaknessesandhelpitimprovethem.However,sometimesareallysmartemployeedevelopsanagendaotherthan

improving thecompany.Rather than identifyingweaknesses so thathecan fixthem,helooksforfaultstobuildhiscase.Specifically,hebuildshiscasethatthecompanyishopelessandrunbyabunchofmorons.Thesmartertheemployee,themoredestructive this typeofbehaviorcanbe.Simplyput, it takesa reallysmart person to bemaximally destructive, because otherwise nobody elsewilllistentohim.Why would a smart person try to destroy the company that he works for?

Thereareactuallymanyreasons.Hereareafew:

1.Sheisdisempowered.Shefeelsthatshecannotaccessthepeopleinchargeand,asaresult,complainingisheronlyvehicletogetthetruthout.2.Sheisfundamentallyarebel.Shewillnotbehappyunlesssheisrebelling;thiscanbeadeeppersonalitytrait.SometimesthesepeopleactuallymakebetterCEOsthanemployees.3.Sheisimmatureandnaive.Shecannotcomprehendthatthepeoplerunningthecompanydonotknoweveryminutedetailoftheoperationandthereforetheyarecomplicitineverythingthat’sbroken.

Often, it’s very difficult to turn these kinds of cases around. Once anemployee takes a public stance, the social pressure for him tobe consistent isenormous. If he tells fifty of his closest friends that the CEO is the stupidestpersonontheplanet,thenreversingthatpositionwillcosthimagreatamountofcredibilitythenexttimehecomplains.Mostpeoplearenotwillingtotakethehittotheircredibility.

EXAMPLE2:THEFLAKE

Somebrilliantpeople canbe totallyunreliable.AtOpsware,weoncehiredanunequivocalgenius—Arthur (nothis realname)wasanengineer in anareaoftheproductwherea typicalnewhirewould take threemonths tobecomefullyproductive.Arthurgotfullyuptospeedintwodays.Onhisthirdday,wegavehim a project that was scheduled to take one month. Arthur completed theproject in three days with nearly flawless quality. More specifically, hecompleted the project in seventy-two hours. Seventy-two nonstop hours: Nostops,nosleep,nonothingbutcoding.Inhisfirstquarteronthejob,hewasthebestemployeethatwehadandweimmediatelypromotedhim.ThenArthurchanged.Hewouldmissdaysofworkwithoutcallingin.Then

he would miss weeks of work. When he finally showed up, he apologizedprofusely, but the behavior didn’t stop. His work product also degraded. Hebecame sloppy and unfocused. I could not understand how such a stellaremployee could go so haywire. Hismanager wanted to fire him, because theteamcouldno longer countonArthur for anything. I resisted. Iknew that thegeniuswasstillinhimandIwantedustofindit.Weneverdid.ItturnedoutthatArthurwasbipolarandhad twosignificantdrugproblems: (1)Hedidnot liketakinghisbipolarmedicationand(2)hewasaddictedtococaine.Ultimately,wehad to fireArthur, but even now, it painsme to think aboutwhatmight have

been.Oneneednotbebipolartobeaflake,butflakybehavioroftenhasaseriously

problematicrootcause.Causesrangefromself-destructivestreakstodrughabitstomoonlightingforotheremployers.Acompanyisateameffortand,nomatterhowhighanemployee’spotential,youcannotgetvaluefromhimunlesshedoeshisworkinamannerinwhichhecanbereliedupon.

EXAMPLE3:THEJERK

Thisparticularsmart-bad-employeetypecanoccuranywhereintheorganizationbut is most destructive at the executive level.Most executives can be pricks,dicks,a-holes,oravarietyofotherprofanenounsat times.Beingdramaticallyimpolitecanbeusedtoimproveclarityoremphasizeanimportantlesson.That’snotthebehaviorthatIamtalkingabout.When used consistently, asinine behavior can be crippling. As a company

grows, its biggest challenge always becomes communication.Keeping a hugenumberofpeopleonthesamepageexecutingthesamegoalsisnevereasy.Ifamemberofyourstaffisaragingjerk,itmaybeimpossible.Somepeoplearesobelligerentintheircommunicationstylethatpeoplejuststoptalkingwhentheyare in the room. If every time anyone brings up an issue with the marketingorganization, the VP ofmarketing jumps down their throats, then guess whattopicwillnevercomeup?Thisbehaviorcanbecomesobadthatnobodybringsupanytopicwhenthejerkisintheroom.Asaresult,communicationacrosstheexecutive staff breaks down and the entire company slowly degenerates.Notethat this only happens if the jerk in question is unquestionably brilliant.Otherwise,nobodywillcarewhensheattacksthem.Thebiteonlyhasimpactifit comes from a big dog. If one of your big dogs destroys communication onyourstaff,youneedtosendhertothepound.

WHENDOYOUHOLDTHEBUS?

The great football coach John Madden was once asked whether he wouldtolerate a player like Terrell Owens on his team.Owenswas both one of themost talented players in the game and one of the biggest jerks. Maddenanswered,“Ifyouholdthebusforeveryoneontheteam,thenyou’llbesolateyou’ll miss the game, so you can’t do that. The bus must leave on time.However, sometimesyou’llhaveaplayer that’ssogood thatyouhold thebusforhim,butonlyhim.”

PhilJackson,thecoachwhohaswonthemostNBAchampionships,wasonceasked about his famously flaky superstar Dennis Rodman, “Since DennisRodman is allowed to miss practice, does this mean other star players likeMichaelJordanandScottiePippencanmisspractice,too?”Jacksonreplied,“Ofcoursenot.ThereisonlyroomforoneDennisRodmanonthisteam.Infact,youreally can only have a very few Dennis Rodmans in society as a whole;otherwise,wewoulddegenerateintoanarchy.”You may find yourself with an employee who fits one of the above

descriptions but nonetheless makes a massive positive contribution to thecompany. You may decide that you will personally mitigate the employee’snegative attributes and keep her from polluting the overall company culture.That’sfine,butremember:Youcanonlyholdthebusforher.

OLDPEOPLE

Yourstartup isgoingwellandasyourbusinessexpandsyouhear thedreadedwords from someone on your board: “You need to hire some senior people.Somereal‘beenthere,donethat’executivestohelpyougetthecompanytothenext level.”Really? Isnow the time? If so,wheredo Ibegin?Andonce Igetthem,whatdoIdowith them?Andhowwill Iknowif theyaredoingagoodjob?The first question you might ask is “Why do I need senior people at all?

Won’ttheyjustruintheculturewiththeirfancyclothes,politicalambitions,andneed to go home to see their families?” To some extent, the answer to all ofthosemaybe“yes,”which iswhy thisquestionmustbe takenquite seriously.However,bringingintherightkindofexperienceattherighttimecanmeanthedifferencebetweenbankruptcyandglory.Let’sgobacktothefirstpartofthequestion.Whyhireaseniorperson?The

shortansweristime.Asatechnologystartup,fromthedayyoustartuntilyourlastbreath,youwillbeinafuriousraceagainsttime.Notechnologystartuphasa long shelf life.Even thebest ideasbecome terrible ideas after a certain age.Howwould Facebook go if Zuckerberg started it lastweek?AtNetscape,wewentpublicwhenwewerefifteenmonthsold.Hadwestartedsixmonthslater,wewouldhavebeenlatetoamarketwiththirty-sevenotherbrowsercompanies.Even if nobody beats you to the punch, nomatter how beautiful your dreammostemployeeswilllosefaithafterthefirstfiveorsixyearsofnotachievingit.Hiringsomeonewhohasalreadydonewhatyouare trying todocan radicallyspeedupyourtimetosuccess.ButCEO,beware:Hiringseniorpeopleintoastartupiskindoflikeanathlete

taking performance-enhancing drugs. If all goes well, you will achieveincrediblenewheights. Ifallgoeswrong,youwill startdegenerating from theinsideout.Inordertomakeallgowell,ifyouareconsideringhiringaseniorpersondo

not chase an abstract rationale like “adult supervision” or “becoming a realcompany.” Aweak definition of what you are looking for will lead to a badoutcome.Theproperreasontohireaseniorpersonistoacquireknowledgeand

experienceinaspecificarea.For example, as a technical founder, you probably do not have terrific

knowledge of how to build a worldwide sales channel, how to create aninvincible brand, or how to identify and negotiate ecosystem-altering businessdevelopment deals. Acquiring a world-class senior person can dramaticallyaccelerateyourcompany’sabilitytosucceedintheseareas.Onegood test fordeterminingwhether togowithoutsideexperienceversus

internal promotion is to figure out whether you value inside knowledge oroutsideknowledgemorefortheposition.Forexample,forengineeringmanagersthecomprehensiveknowledgeofthecodebaseandengineeringteamisusuallymore importantanddifficult toacquire thanknowledgeofhowtorunscalableengineeringorganizations.Asaresult,youmightverywellvaluetheknowledgeofyourownorganizationmorethanthatoftheoutsideworld.In hiring someone to sell your product to large enterprises, the opposite is

true. Knowing how your target customers think and operate, knowing theirculturaltendencies,understandinghowtorecruitandmeasuretherightpeopleintherightregionsof theworldtomaximizeyoursales—thesethingsturnout tobe farmore valuable than knowing your own company’s product and culture.Thisiswhywhentheheadofengineeringgetspromotedfromwithin,sheoftensucceeds.Whentheheadofsalesgetspromotedfromwithin,shealmostalwaysfails.Askingyourself,“DoIvalueinternalorexternalknowledgemoreforthisposition?”willhelpyoudeterminewhethertogoforexperienceoryouth.

ONCETHEYARRIVE

Bringingseniorpeopleonboardcanbefraughtwithperil,asIhaveoutlinedinthe sections “Why It’s Hard to Bring Big Company Execs into LittleCompanies” and “HiringExecutives: IfYou’veNeverDone the Job,HowDoYouHireSomebodyGood?”(seepages119and124).Equally difficult is managing them effectively once they come on board.

Seniorpeopleposeseveralimportantchallenges:

Theycomewiththeirownculture.Theywillbringthehabits,thecommunicationstyle,andvaluesfromthecompanytheygrewupin.It’sveryunlikelythesewillmatchyourenvironmentexactly.Theywillknowhowtoworkthesystem.Becauseseniorpeoplecomefromlargerenvironments,theyusuallydeveloptheskillstonavigateandbe

effectiveinthoseenvironments.Theseskillsmayseempoliticalandunusualinyourenvironment.Youdon’tknowthejobaswellastheydo.Infact,youarehiringthempreciselybecauseyoudon’tknowhowtodothejob.Sohowdoyouholdthemaccountablefordoingagoodjob?

Inordertopreventtheinternaldegenerationmentionedearlier,it’simportanttobeawareoftheaboveissuesandthenemployappropriatecountermeasurestomakesuretheydon’tmetastasize.First,youshoulddemandculturalcompliance.It’sfinethatpeoplecomefrom

othercompanycultures.It’struethatsomeofthosecultureswillhavepropertiesthataresuperiortoyourown.Butthisisyourcompany,yourculture,andyourwayofdoingbusiness.Donotbeintimidatedbyexperienceonthisissue;stickto your guns and stick to your culture. If youwant to expand your culture toincorporate someof thenew thinking, that’s fine,butdosoexplicitly—donotdrift.Next,watchforpoliticallymotivatedtacticsanddonottoleratethem.Perhapsmostimportant,setahighandclearstandardforperformance.Ifyou

want to have aworld-class company, youmustmake sure that the people onyourstaff—betheyyoungorold—areworld-class.It isnotnearlyenoughthatsomeone on your staff can do the job better than you can, because you areincompetentatthejob—that’swhyyouhiredtheminthefirstplace.Becarefulnottosetalowbarbecauseyouhavenotdonetheworktoknow

what good is. For example, I’ve seen many a young CEO excited about hercompany’scompetencyinmarketingandPRbecauseshegotabunchofpositivestoriesonherlaunch.That’snotahighPRstandard.Anybodycangetreporterstowritenicethingsaboutasweet,cuddlybabyofacompany.Onlyworld-classPRpeoplecandealwithgangly,pimple-ridden,teenagecompanies.World-classPR people can turn around negative stories. World-class PR people can turnchickenshitintochickensalad.Turningchickenshitintochickensaladrequireslong-term, trusted relationships, deep know-how, and the confidence to makeuseofbothappropriately.PRkidsdon’thaveanyofthethree.Oneexcellentwaytodevelopahighstandardistointerviewpeoplewhoyou

seedoingagreatjobintheirfield.Findoutwhattheirstandardisandaddittoyourown.Onceyoudetermineahighyetachievableperformancebar,holdyourexecutivetothathighstandardevenifyouhavenoideahowtheymightachieveit. It’s not your job to figure out how to create an incredible brand, tilt theplaying field by cutting a transformational deal, or achieve a sales goal that

nobody thought possible—that’swhat you are paying them to do. That’swhyyouhiredthem.Finally,you’llneedyournewexecutivetobemorethanjustagoalachiever.

Shewillneedtobewellroundedandpartoftheteam.BillCampbelldevelopedanexcellentmethodologyformeasuringexecutivesinabalancedwaythatwillhelpyouachievethis.Hebreaksperformancedownintofourdistinctareas:

1.ResultsagainstobjectivesOnceyou’vesetahighstandard,itwillbestraightforwardtomeasureyourexecutiveagainstthatstandard.2.ManagementEvenifanexecutivedoesasuperbjobachievinghergoals,thatdoesn’tmeansheisbuildingastrongandloyalteam.It’simportanttounderstandhowwellsheismanaging,evenifsheishittinghergoals.3.InnovationIt’squitepossibleforanexecutivetohithergoalforthequarterbyignoringthefuture.Forexample,agreatwayforanengineeringmanagertohithergoalsforfeaturesanddatesisbybuildingahorriblearchitecture,whichwon’tevensupportthenextrelease.Thisiswhyyoumustlookbeyondtheblack-boxresultsandintothesausagefactorytoseehowthingsgetmade.4.WorkingwithpeersThismaynotbeintuitiveatfirst,butexecutivesmustbeeffectiveatcommunicating,supporting,andgettingwhattheyneedfromtheotherpeopleonyourstaff.Evaluatethemalongthisdimension.

AW,MAN,YOUSOLDYOURSOUL

Hiringthefirstseniorpeopleintoyourcompanymayfeellikesellingyoursoul,and if you are not careful, you may well end up selling the soul of yourcompany.But if youwant tomake something fromnothing, youhave to takerisksandyouhavetowinyourraceagainsttime.Thismeansacquiringtheverybest talent, knowledge, and experience even if it requires dealing with someseriousagediversity.

ONE-ON-ONE

After I firstwrote about one-on-ones, people floodedmewith feedback aboutone-on-ones.About half the responders chastisedme, saying that one-on-oneswereuselessandthatIshouldn’tputsomuchemphasisonthem.Theotherhalfwantedtoknowhowtorunmoreeffectiveone-on-ones.Itseemstomethatbothgroupsarelikelytalkingabouttwosidesofthesamecoin.PerhapstheCEO’smostimportantoperationalresponsibilityisdesigningand

implementingthecommunicationarchitectureforhercompany.Thearchitecturemight include the organizational design, meetings, processes, email, yammer,and even one-on-onemeetings withmanagers and employees. Absent a well-designed communication architecture, information and ideaswill stagnate, andyourcompanywilldegenerateintoabadplacetowork.Whileitisquitepossibleto design a great communication architecturewithout one-on-onemeetings, inmost cases one-on-ones provide an excellent mechanism for information andideastoflowuptheorganizationandshouldbepartofyourdesign.Generally, peoplewho thinkone-on-onemeetings are a bad idea have been

victimsofpoorlydesignedones.Thekey toagoodone-on-onemeeting is theunderstanding that it is the employee’s meeting rather than the manager’smeeting.Thisisthefree-formmeetingforallthepressingissues,brilliantideas,andchronicfrustrationsthatdonotfitneatlyintostatusreports,email,andotherlesspersonalandintimatemechanisms.Ifyouareanemployee,howdoyougetfeedbackfromyourmanageronan

excitingbutonly20percentformedideathatyou’renotsureisrelevant,withoutsounding likea fool?Howdoyoupointout thatacolleagueyoudonotknowhowtoworkwithisblockingyourprogresswithoutthrowingherunderthebus?Howdoyougethelpwhenyouloveyourjobbutyourpersonallifeismeltingdown?Throughastatusreport?Onemail?Yammer?Asana?Really?Fortheseandotherimportantareasofdiscussions,one-on-onescanbeessential.If you like structured agendas, then the employee should set the agenda.A

goodpracticeistohavetheemployeesendyoutheagendainadvance.Thiswillgiveherachancetocancelthemeetingifnothingispressing.Italsomakesclearthatitishermeetingandwilltakeasmuchoraslittletimeassheneeds.During

the meeting, since it’s the employee’s meeting, the manager should do 10percent of the talking and 90 percent of the listening. Note that this is theoppositeofmostone-on-ones.While it’s not the manager’s job to set the agenda or do the talking, the

manager should try to draw the key issues out of the employee. The moreintroverted the employee, the more important this becomes. If you manageengineers,drawingoutissueswillbeanimportantskilltomaster.SomequestionsthatI’vefoundtobeveryeffectiveinone-on-ones:

Ifwecouldimproveinanyway,howwouldwedoit?What’sthenumber-oneproblemwithourorganization?Why?What’snotfunaboutworkinghere?Whoisreallykickingassinthecompany?Whomdoyouadmire?Ifyouwereme,whatchangeswouldyoumake?Whatdon’tyoulikeabouttheproduct?What’sthebiggestopportunitythatwe’remissingouton?Whatarewenotdoingthatweshouldbedoing?Areyouhappyworkinghere?

In the end, the most important thing is that the best ideas, the biggestproblems, and the most intense employee life issues make their way to thepeoplewhocandealwith them.One-on-onesarea time-testedway todo that,butifyouhaveabetterone,goaheadwithyourbadself.

PROGRAMMINGYOURCULTURE

Ask ten foundersabout companycultureandwhat itmeansandyou’llget tendifferent answers. It’s about office design, it’s about screening out thewrongkinds of employees, it’s about values, it’s about fun, it’s about alignment, it’saboutfindinglike-mindedemployees,it’saboutbeingcultlike.Sowhat is culture?Does culturematter? If so, howmuch time should you

spendonit?Let’s start with the second question first. The primary thing that any

technologystartupmustdo isbuildaproduct that’sat least ten timesbetteratdoing something than the current prevailing way of doing that thing. Two orthree timesbetterwill not begood enough to get people to switch to the newthingfastenoughorinlargeenoughvolumetomatter.Thesecondthingthatanytechnologystartupmustdoistotakethemarket.Ifit’spossibletodosomethingtentimesbetter,it’salsopossiblethatyouwon’tbetheonlycompanytofigurethatout.Therefore,youmusttakethemarketbeforesomebodyelsedoes.Veryfewproducts are ten timesbetter than the competition’s, sounseating thenewincumbentismuchmoredifficultthanunseatingtheoldone.Ifyoufail todobothofthosethings,yourculturewon’tmatteronebit.The

worldisfullofbankruptcompanieswithworld-classcultures.Culturedoesnotmakeacompany.So,whybotherwithcultureatall?Threereasons:

1.Itmatterstotheextentthatitcanhelpyouachievetheabovegoals.2.Asyourcompanygrows,culturecanhelpyoupreserveyourkeyvalues,makeyourcompanyabetterplacetowork,andhelpitperformbetterinthefuture.3.Perhapsmostimportant,afteryouandyourpeoplegothroughtheinhumanamountofworkthatitwilltaketobuildasuccessfulcompany,itwillbeanepictragedyifyourcompanycultureissuchthatevenyoudon’twanttoworkthere.

CREATINGACOMPANYCULTURE

WhenIrefertocompanyculture,Iamnotreferringtootherimportantactivitieslikecompanyvaluesandemployeesatisfaction.Specifically,Iamwritingaboutdesigningawayofworkingthatwill:

DistinguishyoufromcompetitorsEnsurethatcriticaloperatingvaluespersistsuchasdelightingcustomersormakingbeautifulproductsHelpyouidentifyemployeeswhofitwithyourmission

Culturemeans lots of other things in other contexts, but the above will beplentytodiscusshere.Whenyou start implementing your culture, keep inmind thatmost ofwhat

willberetrospectivelyreferredtoasyourcompany’sculturewillnothavebeendesignedintothesystem,butratherwillhaveevolvedovertimebasedonyourbehaviorandthebehaviorofyourearlyemployees.Asaresult,youwillwanttofocus on a small number of cultural design points that will influence a largenumberofbehaviorsoveralongperiodoftime.InhisbestsellingbookBuilttoLast,JimCollinswrotethatoneofthethings

thatlong-lastingcompanieshestudiedhaveincommonisa“cultlikeculture.”Ifoundthisdescriptiontobeconfusingbecauseitseemstoimplythataslongasyourcultureisweirdenoughandyouarerabidenoughaboutit,youwillsucceedontheculturalfront.That’s related to the truth,butnotactually true. In reality,Collinswas right

thataproperlydesignedcultureoftenendsuplookingcultlikeinretrospect,butthat’snottheinitialdesignprinciple.Youneedn’tthinkhardabouthowyoucanmakeyourcompanyseembizarre tooutsiders.However,youdoneed to thinkabouthowyoucanbeprovocativeenoughtochangewhatpeopledoeveryday.Ideally, a cultural design point will be trivial to implement but have far-

reaching behavioral consequences. Key to this kind of mechanism is shockvalue.Ifyouputsomethingintoyourculturethatissodisturbingthatitalwayscreatesaconversation,itwillchangebehavior.AswelearnedinTheGodfather,askaHollywoodmogultogivesomeoneajobandhemightnotrespond.Putahorse’shead inhis bed andunemploymentwill dropbyone.Shock is a greatmechanismforbehavioralchange.Herearethreeexamples:

Desksmade out of doors Very early on, Jeff Bezos, founder andCEO ofAmazon.com, envisioned a company thatmademoney by delivering value toratherthanextractingvaluefromitscustomers.Inordertodothat,hewantedtobeboththepriceleaderandcustomerserviceleaderforthelongrun.Youcan’tdothatifyouwastealotofmoney.Jeffcouldhavespentyearsauditingeveryexpense and raining hell on anybody who overspent, but he decided to buildfrugality into his culture.He did itwith an incredibly simplemechanism:AlldesksatAmazon.comforall timewouldbebuiltbybuyingcheapdoors fromHome Depot and nailing legs to them. These door desks are not greatergonomically, nor do they fit with Amazon.com’s $150 billion–plus marketcapitalization,butwhenashockednewemployeeaskswhyshemustworkonamakeshiftdeskconstructedoutofrandomHomeDepotparts,theanswercomesbackwithwitheringconsistency:“Welookforeveryopportunitytosavemoneyso thatwecandeliver thebestproducts for the lowestcost.” Ifyoudon’t likesittingatadoor,thenyouwon’tlastlongatAmazon.TendollarsperminuteWhenwestartedAndreessenHorowitz,MarcandI

wantedthefirmtotreatentrepreneurswithgreatrespect.Werememberedhowpsychologicallybrutal theprocessofbuildinga companywas.Wewanted thefirmtorespectthefactthatinthebacon-and-eggbreakfastofastartup,wewerewith thechickenand theentrepreneurwas thepig:Wewere involved,but shewascommitted.Wethought thatonewaytocommunicaterespectwouldbe toalwaysbeontimetomeetingswithentrepreneurs.Ratherthanmakethemwaitin our lobby for thirtyminuteswhilewe attended tomore important businesslikesomanyventurecapitaliststhatwevisited,wewantedourpeopletobeontime, prepared, and focused. Unfortunately, anyone who has ever workedanywhereknowsthatthisiseasiersaidthandone.Inordertoshockthecompanyintotherightbehavior,weinstitutedaruthlesslyenforcedten-dollar-per-minutefineforbeinglatetoameetingwithanentrepreneur.So,forexample:Youareonareally importantcallandwillbe tenminutes late?Noproblem, justbringone hundred dollars to themeeting and pay your fine.When new employeescomeonboard, they find this shocking,whichgivesus a great opportunity toexplainindetailwhywerespectentrepreneurs.Ifyoudon’tthinkentrepreneursare more important than venture capitalists, we can’t use you at AndreessenHorowitz.MovefastandbreakthingsMarkZuckerbergbelievesininnovationandhe

believestherecanbenogreatinnovationwithoutgreatrisk.So,intheearlydaysofFacebook,hedeployedashockingmotto:Movefastandbreakthings.Didthe

CEOreallywantus tobreakthings?Imean,he’s tellingus tobreakthings!Amotto that shocking forces everyone to stop and think.When they think, theyrealize that if you move fast and innovate, you will break things. If you askyourself, “Should I attempt this breakthrough? Itwill be awesome, but itmaycauseproblemsintheshortterm,”youhaveyouranswer.Ifyou’dratherberightthaninnovative,youwon’tfitinatFacebook.Priortofiguringouttheexactformofyourcompany’sshocktherapy,besure

that yourmechanism agreeswith your values. For example, JackDorseywillnevermakehisowndesksoutofdoorsatSquarebecauseatSquare,beautifuldesigntrumpsfrugality.WhenyouwalkintoSquare,youcanfeelhowseriouslytheytakedesign.

WHYDOGSATWORKANDYOGAAREN’TCULTURE

Startups today do all kinds of things to distinguish themselves. Many great,many original,many quirky, butmost of themwill not define the company’sculture.Yes, yogamaymakeyour companyabetter place towork for peoplewholikeyoga.Itmayalsobeagreatteam-buildingexerciseforpeoplewholikeyoga.Nonetheless, it’snotculture.Itwillnotestablishacorevaluethatdrivesthebusinessandhelpspromoteitinperpetuity.Itisnotspecificwithrespecttowhatyourbusinessaimstoachieve.Yogaisaperk.Somebodykeepingapitbullinhercubemaybeshocking.Thelessonlearned

—thatanimalloversarewelcomeorthatemployeescanlivehowevertheywant—mayprovidesomesocietalvalue,butitdoesnotconnecttoyourbusinessinadistinguishingway.Everysmartcompanyvaluesitsemployees.Perksaregood,buttheyarenotculture.

THEPOINTOFITALL

Inthelatersection“HowtoEvaluateCEOs”(seepage235),IdescribetheCEOjob as knowing what to do and getting the company to do what you want.Designingapropercompanyculturewillhelpyougetyourcompanytodowhatyouwantincertainimportantareasforaverylongtime.

TAKINGTHEMYSTERYOUTOFSCALINGACOMPANY

Ifyouwanttobuildanimportantcompany,thenatsomepointyouhavetoscale.People in startup land often talk about themagic of how fewpeople built theoriginalGoogle or the original Facebook, but today’sGoogle employs twentythousand people and today’s Facebook employs more than fifteen hundredpeople.So,ifyouwanttodosomethingthatmatters,thenyouaregoingtohavetolearntheblackartofscalingahumanorganization.Oftenboardmembersgiveentrepreneurstwobitsofadviceregardingscale:

1.Getamentor.2.Findsome“beenthere,donethat”executiveswhoalreadyknowhowtoscale.

Theseanswers,whilefineasfarastheygo,havesomeimportantlimitations.First, ifyoudon’tknowanythingaboutscalinganorganization, thenitwillbevery difficult for you to evaluate people for that job. Imagine trying to find akillerengineerifyou’dneverwrittenasingleprogram.Second,manyinvestor-boardmembersdon’tknowanythingaboutscalingacompany,either,andcanbesuckers for peoplewho have the experience but not the skills. If you’ve everworked in a largeorganization, youknow that there areplentyofpeoplewithexperiencerunningthembutnoneoftherequisiteskillstorunthemwell.Thisadviceisstillgood,buttherightwaytopickboththebestmentorsand

bestemployeesisbyfirstlearningthebasics;thenyoucanapplythemyriadofscalingtechniquesinthemanagementliteraturedependingonthecontext.

THEBASICIDEA:GIVEGROUNDGRUDGINGLY

When anorganization grows in size, things thatwere previously easy becomedifficult.Specifically, the following things that causeno troublewhenyouaresmallbecomebigchallengesasyougrow:

CommunicationCommonknowledgeDecisionmaking

In order to get a clear understanding of the problem, let’s start with theboundarycondition.Imagineacompanyofoneemployee.Thatemployeewritesandtestsallthecode,doesallthemarketingandsales,andmanagesherself.Shehascompleteknowledgeofeverythinginthecompany,makesallthedecisions,needn’t communicate with anyone, and is totally alignedwith herself. As thecompanygrows,thingswillonlygetworseineachdimension.Ontheotherhand,ifthecompanydoesn’texpand,thenitwillneverbemuch

ofacompany,sothechallengeistogrowbutdegradeasslowlyaspossible.Thereisagreatanalogueto thisconcept inAmericanfootball.Anoffensive

lineman’sjobistoprotectthequarterbackfromonrushingdefensivelinemen.Iftheoffensive linemanattempts todo thisbyholdinghisground, thedefensivelineman will easily run around him and crush the quarterback. As a result,offensive linemen are taught to lose the battle slowly or to give groundgrudgingly. They are taught to back up and allow the defensive lineman toadvance,butjustalittleatatime.When you scale an organization, you will also need to give ground

grudgingly. Specialization, organizational structure, and process all complicatethingsandimplementingthemwillfeellikeyouaremovingawayfromcommonknowledge and quality communication. It is very much like the offensivelinemantakingastepbackward.Youwillloseground,butyouwillpreventyourcompanyfromdescendingintochaos.

HOWTODOIT

Atthepointwhenaddingpeopleintothecompanyfeelslikemoreworkthantheworkthatyoucanoffloadtothenewemployees,thedefensivelinemanhasrunaroundyouandyouprobablyneedtostartgivinggroundgrudgingly.

SPECIALIZATION

The first scale technique to implement is specialization. In startups,everybodystartsoutasajack-of-all-trades.Forexample,engineerswritecode,managethebuild system, test theproduct, and, increasingly, deploy it andoperate it.This

workswellinthebeginningbecauseeverybodyknowseverythingandtheneedtocommunicateisminimized;therearenocomplicatedhandoffs,becausethereisnobodytohandanythingto.Asthecompanygrows,itbecomesincreasinglydifficult to add new engineers, because the learning curve starts to get super-steep.Gettinganewengineerup tospeedstarts tobecomemoredifficult thandoingtheworkyourself.Atthispoint,youneedtospecialize.Bydedicatingpeople and teams to such tasks as thebuild environment, the

test environment, and operations, you will create some complexity—handoffsacross groups, potentially conflicting agendas, and specialized rather thancommonknowledge.Inordertomitigatetheseissues,youwillneedtoconsiderotherscaletechniqueslikeorganizationaldesignandprocess.

ORGANIZATIONALDESIGN

Thefirstruleoforganizationaldesignisthatallorganizationaldesignsarebad.With any design, you will optimize communication among some parts of theorganization at the expense of other parts. For example, if you put productmanagementintheengineeringorganization,youwilloptimizecommunicationbetweenproductmanagementandengineeringattheexpenseofcommunicationbetweenproductmanagementandmarketing.Asaresult,assoonasyourollouttheneworganization,peoplewillfindfaultwithitandtheywillberight.Nonetheless, at some point themonolithic design of one huge organization

runsoutofgasandyouwillneedtosplitthingsintosmallersubgroups.Atthemost basic level, you’ll want to consider giving the groups that you’vespecializedtheirownmanagersastheygrow.YoumaywantaQAmanager,forexample. After that, things becomemore complex. Do client engineering andserverengineeringhave theirowngroupsordoyouorganizebyusecasesandincludealltechnicalcomponents?Whenyougetreallybig,you’llneedtodecidewhether to organize the entire company around functions (for example, sales,marketing, product management, engineering) or around missions—self-containedbusinessunitsthatcontainmultiplefunctions.Yourgoalistochoosetheleastofallevils.Thinkoftheorganizationaldesign

as the communications architecture for your company. If you want people tocommunicate, the best way to accomplish that is to make them report to thesamemanager. By contrast, the further away people are in the organizationalchart, the less they will communicate. The organizational design is also thearchitecture for how the company communicates with the outside world. For

example,youmightwant toorganizeyoursales forcebyproduct tomaximizecommunication with the relevant product groups and maximize the productcompetencyofthesalesforce.Ifyoudothat,thenyouwilldosoattheexpenseofsimplicityforcustomerswhobuymultipleproductsandwillnowhavetodealwithmultiplesalespeople.Withthisinmind,herearethebasicstepstoorganizationaldesign:

1.Figureoutwhatneedstobecommunicated.Startbylistingthemostimportantknowledgeandwhoneedstohaveit.Forexample,knowledgeoftheproductarchitecturemustbeunderstoodbyengineering,QA,productmanagement,marketing,andsales.2.Figureoutwhatneedstobedecided.Considerthetypesofdecisionsthatmustgetmadeonafrequentbasis:featureselection,architecturaldecisions,howtoresolvesupportissues.Howcanyoudesigntheorganizationtoputthemaximumnumberofdecisionsunderthedomainofadesignatedmanager?3.Prioritizethemostimportantcommunicationanddecisionpaths.Isitmoreimportantforproductmanagerstounderstandtheproductarchitectureorthemarket?Isitmoreimportantforengineerstounderstandthecustomerorthearchitecture?Keepinmindthattheseprioritieswillbebasedontoday’ssituation.Ifthesituationchanges,thenyoucanreorganize.4.Decidewho’sgoingtoruneachgroup.Noticethatthisisthefourthstep,notthefirst.Youwanttooptimizetheorganizationforthepeople—forthepeopledoingthework—notforthemanagers.Mostlargemistakesinorganizationaldesigncomefromputtingtheindividualambitionsofthepeopleatthetopoftheorganizationaheadofthecommunicationpathsforthepeopleatthebottomoftheorganization.Makingthisstepfourwillupsetyourmanagers,buttheywillgetoverit.5.Identifythepathsthatyoudidnotoptimize.Asimportantaspickingthecommunicationpathsthatyouwilloptimizeisidentifyingtheonesthatyouwillnot.Justbecauseyoudeprioritizedthemdoesn’tmeantheyareunimportant.Ifyouignorethementirely,theywillsurelycomebacktobiteyou.6.Buildaplanformitigatingtheissuesidentifiedinstepfive.Onceyou’veidentifiedthelikelyissues,youwillknowtheprocessesyouwillneedto

buildtopatchtheimpendingcross-organizationalchallenges.

These six steps should get you pretty far. When we examine advancedorganizationaldesign,we’llalsoneedtoconsidertrade-offssuchasspeedversuscost, how to roll out organizational changes, and how often you shouldreorganize.

PROCESS

The purpose of process is communication. If there are five people in yourcompany,youdon’tneedprocess,becauseyoucanjusttalktoeachother.Youcan hand off taskswith a perfect understanding ofwhat’s expected, you passimportant informationfromoneperson toanother,andyoucanmaintainhigh-quality transactionswithnobureaucraticoverhead.Withfour thousandpeople,communicationbecomesmoredifficult.Adhoc,point-to-point communicationnolongerworks.Youneedsomethingmorerobust—acommunicationbusor,tousetheconventionaltermforhumancommunicationbuses,aprocess.A process is a formal, well-structured communication vehicle. It can be a

heavily engineered Six Sigma process or it can be a well-structured regularmeeting.Thesizeoftheprocessshouldbescaledupordowntomeettheneedsofthecommunicationchallengethatitfacilitates.When communication in an organization spans across organizational

boundaries,processeswillhelpensurethatthecommunicationhappensandthatithappenswithquality.Ifyouarelookingforthefirstprocesstoimplementinyour company, consider the interview process. It usually runs acrossorganizationalboundaries(thehiringgroup,humanresources—orwherevertherecruiter lives, and supporting groups), involves people from outside thecompany (the candidate), and is critically important to the success of thecompany.Whoshoulddesignaprocess?Thepeoplewhoarealreadydoingtheworkin

an adhocmanner.Theyknowwhat needs tobe communicated and towhom.Naturallytheywillbetherightgrouptoformalizetheexistingprocessandmakeitscalable.Whenshouldyoustartimplementingprocesses?Whilethatvariesdepending

onyoursituation,keep inmind that it’smucheasier toaddnewpeople tooldprocesses than new processes to old people. Formalizewhat you are doing tomakeiteasytoonboardnewpeople.Much has beenwritten about process design, so I won’t repeat that here. I

have found the “TheBasics ofProduction,” the first chapter ofAndyGrove’sHighOutputManagement, tobeparticularlyhelpful.Fornewcompanies,hereareafewthingstokeepinmind:

Focusontheoutputfirst.Whatshouldtheprocessproduce?Inthecaseoftheinterviewprocess,anoutstandingemployee.Ifthat’sthegoal,what’stheprocesstogetthere?Figureouthowyou’llknowifyouaregettingwhatyouwantateachstep.Areyougettingenoughcandidates?Areyougettingtherightcandidates?Willyourinterviewprocessfindtherightpersonforthejob?Onceyouselecttheperson,willtheyacceptthejob?Oncetheyacceptthejob,willtheybecomeproductive?Oncetheybecomeproductive,willtheystaywithyourcompany?Howwillyoumeasureeachstep?Engineeraccountabilityintothesystem.Whichorganizationandwhichindividualisresponsibleforeachstep?Whatcanyoudotoincreasethevisibilityoftheirperformance?

FINALTHOUGHT

Theprocessofscalingacompanyisnotunliketheprocessofscalingaproduct.Different sizes of company impose different requirements on the company’sarchitecture. If you address those requirements too early, your company willseem heavy and sluggish. If you address those requirements too late, yourcompanymaymeltdownunderthepressure.Bemindfulofyourcompany’struegrowthrateasyouaddarchitecturalcomponents.It’sgoodtoanticipategrowth,butit’sbadtooveranticipategrowth.

THESCALEANTICIPATIONFALLACY

Theotherday Iwas talking to a coupleof friendsofmine,oneaVCand theother a CEO. During the meeting, we discussed one of the executives at theCEO’s company. The executive in question performs exceptionally but lacksexperiencemanagingat larger scale.My friend theVC innocently advised theCEO to carefully consider whether the executive would scale to meet thecompany’s needs in the future. I responded swiftly, aggressively, and loudly,saying,“That’sahorrible ideaandmakesnosenseatall.”Bothofmyfriendswerestartledatmyoutburst.NormallyIamdisciplinedenoughtorefrainfromlettingmyfeelingspassstraightthroughmymouthwithoutstoppingatmybrainforreview.Whytheoutburst?Hereismyanswer.As CEO, you must constantly evaluate all the members of your team.

However,evaluatingpeopleagainstthefutureneedsofthecompanybasedonatheoreticalviewofhowtheywillperformiscounterproductive,forthefollowingreasons:

Managingatscaleisalearnedskillratherthananaturalability.Nobodycomesoutofthewombknowinghowtomanageathousandpeople.Everybodylearnsatsomepoint.It’snearlyimpossibletomakethejudgmentinadvance.Howdoyoutellinadvanceifanexecutivecanscale?WasitobviousthatBillGateswouldlearnhowtoscalewhenhewasaHarvarddropout?Howdoyougoaboutmakingthatdecision?Theactofjudgingpeopleinadvancewillretardtheirdevelopment.Ifyoumakeajudgmentthatsomeoneisincapableofdoingsomethingsuchasrunningalargerorganization,willitmakesensetoteachthemthoseskillsorevenpointouttheanticipateddeficiencies?Probablynot.You’vealreadydecidedtheycan’tdoit.Hiringscalableexecstooearlyisabadmistake.Thereisnosuchthingasagreatexecutive.Thereisonlyagreatexecutiveforaspecificcompanyataspecificpointintime.MarkZuckerbergisaphenomenalCEOfor

Facebook.HewouldnotbeagoodCEOforOracle.Similarly,LarryEllisondoesaterrificjobatOraclebuthewouldnotbetherightpersontomanageFacebook.Ifyoujudgeyourteaminadvanceandhaveahighsenseofurgency,youwillbringinexecutiveswhocanmanageathighscaleinadvanceofneedingthem.Unfortunately,youwillprobablyignoretheirabilitytodothejobforthenexttwelvemonths,whichistheonlyrelevantmeasure.Asaresult,youwillswapoutgoodexecutivesforworseones.Youstillhavetomakethejudgmentattheactualpointintimewhenyouhitthehigherlevelofscale.Evenifyouavoidthetrapofhiringascalableexecutivetooearlyorretardingthenewexecutive’sdevelopment,youstillhaven’tactuallyboughtyourselfanythingbymakingtheprejudgment.RegardlessofwhatyoudecidedatpointintimeA,youstillhavetoevaluatethesituationwithfarbetterdataatpointintimeB.It’snowaytoliveyourlifeorrunanorganization.Deciding(withwoefullyincompletedata)thatsomeonewhoworkstheirbuttoff,doesaterrificjob,andloyallycontributestoyourmissionwon’tbewithyouthreeyearsfromnowtakesyoutoadarkplace.It’saplaceofinformationhiding,dishonesty,andstiltedcommunication.It’saplacewhereprejudicesubstitutesforjudgment.It’saplacewherejudgmentreplacesteaching.It’saplacewhereteamworkbecomesinternalwarfare.Don’tgothere.So,ifyoudon’tprejudgepeople’sabilitytoscale,howdoyoumakethejudgment?Youshouldevaluateyourteamatleastonceaquarteronalldimensions.Twokeyscanhelpyouavoidthescaleanticipationtrap:Don’tseparatescalefromtherestoftheevaluation.Therelevantquestionisn’twhetheranexecutivecanscale;it’swhethertheexecutivecandothejobatthecurrentscale.Youshouldevaluateholisticallyandthiswillpreventyoufromseparatingoutscale,whichoftenleadstoanunwisepredictionoffutureperformance.Makethejudgmentonarelativeratherthananabsolutescale.Askingyourselfwhetheranexecutiveisgreatcanbeextremelydifficulttoanswer.Abetterquestion:Forthiscompanyatthisexactpointintime,doesthereexistanexecutivewhoIcanhirewhowillbebetter?Ifmybiggestcompetitorhiresthatperson,howwillthatimpactourabilitytowin?Predictingwhetheranexecutivecanscalecorruptsyourabilitytomanage,isunfair,anddoesn’twork.

—CHAPTER7—

HOWTOLEADEVENWHENYOUDON’TKNOWWHEREYOUAREGOING

“ThisforeveryghettointhehoodNastheDon,SuperCattheDonDada,understood.”

—NAS,“THEDON”

After selling theLoudcloud business toEDS,we immediately plunged into anewcrisis.Ourinvestorscouldnotunderstandhowsellingallofourrevenueandallofourcustomerscouldpossiblyleaveuswithanythingworthinvestingin.Asa result, institutional investors sold all of theirOpsware shares, and our stockpricefellto$0.35pershare.Thisturnedouttobeanoteworthyprice,becauseitcomputedtoamarketcapitalizationequaltohalfofthecashwehadinthebank.This signaled that investors believed that theOpsware business had no value,and they further expected us to burn up half our cash before coming to oursenses and returning the cash to investors.Tomakemattersmoremiserable, Ireceived a notice fromNASDAQ informingme that if I didnot get our stockpriceabove$1withinthenextninetydays,theywoulddelistusandwewouldtradewiththepennystocks.Ibroughtthischeerynewstotheboardwiththreeoptions:

1.Reversesplit.Wecouldreversesplitthestock10:1andhavetentimesfewersharesandatentimeshigherstockprice.2.Givein.Wecouldbecomeapennystock.3.Hittheroad.Icouldgoontheroadandtrytogetenoughpeopletobuysothatthestockpricewouldtriple.

The board was extremely sympathetic and open to every option. Andy

Rachleff pointed out that the negative connotation with reverse splits amonginvestorshadlessenedduetothesheernumberofthem.Marchypothesizedthatbeingdelistedinapost-newspapereconomymightnotmattersomuch.Still,Ididnotwanttoreversesplitthestock.Morethananything,itfeltlikea

capitulationandadramaticsignofweakness.Reversesplittingwouldsaytothemarket that I believedwe really wereworth half the cash that we had in thebank.Ialsodidn’twantthecompanytogetdelisted.IknewthatMarcwouldberightoneday,butIalsoknewthatmanyinstitutionalinvestorswereprohibitedfrombuyingpennystocksinthecurrentdayandage.Idecidedtohittheroad.Thefirstbigquestionwas“Hittheroadtogoseewhom?”Atthattime,most

institutionalinvestorswouldn’tinvestinstockswithpricesunder$10,letaloneunder$1.SoMarcandIcalledournetworkingguru,thefamousangelinvestorRonConway,foradvice.Wetoldhimthestory,explainingthatthe$20milliona year EDS contract alonemade the Opsware business worth something, andadding thatwe had a great team and huge amount of intellectual property, sotherewasno reason for us to trade at half of cash.Ron listened carefully andthensaid,“IthinkyoushouldgoseeHerbAllen.”I had heard about Herb’s investment bank, Allen & Company, but didn’t

reallyknowmuchabouthim.Allen&Companywasfamousforrunningthebestbusiness conference in the world. The conference is invitation-only andconsistently attracts guestswhomyouwill never find at anyother conference.PeoplelikeBillGates,WarrenBuffett,andRupertMurdochareregulars.Allen&Companymayattractmoretop-tiergueststhanallotherbusinessconferencescombined—it’sthatgood.Marcand Iarrivedat theAllen&Companyoffice inManhattan, located in

theCoca-ColabuildingwhereHerb’sfather,Herbert,hadservedontheboardofdirectorsformanyyears.IfasinglewordcoulddescribetheAllen&Companyoffice itwould beclassy.Beautifully decorated, yet not flashy, the officewasbothelegantandcomfortable.Likehisoffice,Herbhimselfwasbothunassumingandclassy.Heopenedthe

meeting by complimenting Ron, saying that any referral from Ron waspersonally important to him.Marc and I then carefully tookHerb through thestoryofLoudcloud—howwehad sold the servicesbusiness toEDS, retainingthesoftwareandthekeypeople,andhadsecureda$20millionayearsoftwarelicense.Beyondthat,wehadatotallycleanbalancesheetandweresurelyworthmore than $0.35 per share.Herb listened attentively to the entire presentationandthensaid,“I’dliketobehelpful.I’llseewhatIcando.”Ihadnoideaifhe

meant, asmany in SiliconValleywouldmean, “Screw off, I’m not buying apennystock”orifhemeantwhathe’dsaid.Isoonfoundout.Over the next couple ofmonths,Allen&Company boughtOpsware stock,

Herb Allen personally bought Opsware stock, and several Allen & Companyclientsbecamemajor investors.Thisactivitypropelled thestockfrom$0.35to$3 per share in a matter of months. We’d avoided delisting, rebuilt theshareholder base, and given employees hope.Everythingwas largely due to asinglemeetingwithHerbAllen.Years later I asked Herb why he believed in our company at a time when

nobodyelsedid.Ipointedout that,at thetime,Allen&Companywasn’tveryinvolvedintechnology,letalonedatacenterautomation.Herbreplied,“Ididn’tunderstandanythingaboutyourbusinessandIunderstoodverylittleaboutyourindustry. What I saw was two guys come visit me when every other publiccompany CEO and chairman was hiding under their desk. Not only did youcomeseeme,butyouweremoredeterminedandconvincedyouwouldsucceedthanguysrunninggiantbusinesses.Investingincourageanddeterminationwasaneasydecisionforme.”That’s howHerbAllen does business.And that’swhy, if given the chance,

you’dbeafoolnottodobusinesswithHerb.PerhapsthemostimportantthingthatIlearnedasanentrepreneurwastofocusonwhat Ineeded toget rightandstopworryingaboutall the things that Ididwrongormightdowrong.Thissectionencapsulates thevariouspartsof thoselessonsandprovidesguidanceonhowtogettheimportantthingsright.

THEMOSTDIFFICULTCEOSKILL

By far themostdifficult skill I learnedasCEOwas the ability tomanagemyown psychology. Organizational design, process design, metrics, hiring, andfiringwereallrelativelystraightforwardskillstomastercomparedwithkeepingmymindincheck.IthoughtIwastoughgoingintoit,butIwasn’ttough.Iwassoft.Over the years, I’ve spoken to hundreds of CEOs, all with the same

experience. Nonetheless, very few people talk about it and I have never readanythingonthetopic.It’slikethefightclubofmanagement:ThefirstruleoftheCEOpsychologicalmeltdownisdon’ttalkaboutthepsychologicalmeltdown.Attheriskofviolatingthesacredrule,Iwillattempttodescribethecondition

and prescribe some techniques that helped me. In the end, this is the mostpersonalandimportantbattlethatanyCEOwillface.

IFI’MDOINGAGOODJOB,WHYDOIFEELSOBAD?

Generally, someone doesn’t become a CEO unless she has a high sense ofpurposeandcaresdeeplyabouttheworkshedoes.Inaddition,aCEOmustbeaccomplished enough or smart enough that peoplewill want to work for her.NobodysetsouttobeabadCEO,runadysfunctionalorganization,orcreateamassivebureaucracythatgrindshercompanytoascreechinghalt.YetnoCEOever has a smooth path to a great company. Along the way, many things gowrongandallofthemcouldhaveandshouldhavebeenavoided.ThefirstproblemisthateverybodylearnstobeaCEObybeingaCEO.No

trainingasamanager,generalmanager,orinanyotherjobactuallypreparesyoutorunacompany.Theonlythingthatpreparesyoutorunacompanyisrunningacompany.Thismeans thatyouwill faceabroadsetof things thatyoudon’tknowhowtodothatrequireskillsyoudon’thave.Nevertheless,everybodywillexpect you to know how to do them, because, well, you are the CEO. IrememberwhenIfirstbecameCEO,aninvestoraskedmetosendhimthe“captable.”Ihadavagueideaofwhathemeant,butIdidn’tactuallyknowwhattheformatwassupposedtolooklikeorwhatshouldbeincludedorexcluded.Itwas

asillylittlethingandIhadmuchbiggerthingstoworryabout,buteverythingishardwhenyoudon’tactuallyknowwhatyouaredoing.Iwastedquiteabitoftimesweatingoverthatstupidspreadsheet.Even if you knowwhat you are doing, things gowrong. Things go wrong

because building amultifaceted human organization to compete andwin in adynamic, highly competitivemarket turnsout tobe reallyhard. IfCEOsweregradedonacurve, themeanon the testwouldbe22outof100.Thiskindofmean can be psychologically challenging for a straight-A student. It isparticularlychallengingbecausenobodytellsyouthatthemeanis22.Ifyoumanageateamoftenpeople,it’squitepossibletodosowithveryfew

mistakes or bad behaviors. If you manage an organization of one thousandpeople,itisquiteimpossible.Atacertainsize,yourcompanywilldothingsthatare sobad that younever imagined that you’dbe associatedwith that kindofincompetence.Seeingpeople fritterawaymoney,wasteeachother’s time,anddosloppyworkcanmakeyoufeelbad.Ifyouare theCEO, itmaywellmakeyousick.Andtorubsaltintothewoundandmakemattersworse,it’syourfault.

NOBODYTOBLAME

“Youcan’tblameJazzmusiciansorDavidSternwithhisNBAfashionissues.”

—NAS,“HIPHOPISDEAD”

Whenpeopleinmycompanywouldcomplainaboutonethingoranotherbeingbroken, suchas theexpense reportingprocess, Iwould joke that itwasallmyfault.Thejokewasfunny,becauseitwasn’treallyajoke.Everyprobleminthecompany was indeed my fault. As the founding CEO, every hire and everydecision that the company evermade happened undermy direction. Unlike ahired gunwho comes in and blames all of the problems on the prior regime,therewasliterallynobodyformetoblame.Ifsomeonewaspromotedforallthewrongreasons,thatwasmyfault.Ifwe

missedthequarterlyearningstarget,thatwasmyfault.Ifagreatengineerquit,thatwasmyfault.Ifthesalesteammadeunreasonabledemandsontheproductorganization,thatwasmyfault.Iftheproducthadtoomanybugs,thatwasmyfault.Itkindofsuckedtobeme.Beingresponsibleforeverythingandgettinga22ontheteststartstoweighon

yourconsciousness.

TOOMUCHBROKENSTUFF

Giventhisstress,CEOsoftenmakeoneofthefollowingtwomistakes:

1.Theytakethingstoopersonally.2.Theydonottakethingspersonallyenough.

In the first scenario, the CEO takes every issue incredibly seriously andpersonally and urgently moves to fix it. Given the volume of the issues, thismotionusuallyresultsinoneoftwoscenarios.IftheCEOisoutwardlyfocused,sheendsupterrorizingtheteamtothepointwherenobodywantstoworkatthecompanyanymore.IftheCEOisinwardlyfocused,sheendsupfeelingsosickfromalltheproblemsthatshecanbarelymakeittoworkinthemorning.Inthesecondscenario,inordertodampenthepainoftherollingdisasterthat

is thecompany, theCEO takesaPollyannaishattitude: It’snot sobad. In thisview,noneof theproblems isactually thatbadand theyneedn’tbedealtwithurgently.By rationalizing away the issues, theCEO feels better about herself.The problem is that she doesn’t actually fix any of the problems and theemployees eventually become quite frustrated that the chief executive keepsignoringthemostbasicproblemsandconflicts.Ultimately,thecompanyturnstocrap.Ideally, theCEOwill be urgent yet not insane. Shewillmove aggressively

and decisively without feeling emotionally culpable. If she can separate theimportance of the issues from how she feels about them, she will avoiddemonizingheremployeesorherself.

IT’SALONELYJOB

In your darkestmoments asCEO, discussing fundamental questions about theviability of your company with your employees can have obvious negativeconsequences.Ontheotherhand,talkingtoyourboardandoutsideadviserscanbefruitless.Theknowledgegapbetweenyouandthemissovastthatyoucannotactually bring them fully up to speed in amanner that’s useful inmaking thedecision.Youareallalone.AtLoudcloud,whenthedot-combubbleburstandsubsequentlysentmostof

our customers into bankruptcy, it crippled our business and devastated our

balancesheet.Orrather,thatwasoneinterpretation.Anotherinterpretation,andnecessarily the official story for the company,was thatwe still had plenty ofmoney in the bank andwere signing up traditional enterprise customers at animpressiverate.Whichinterpretationwasclosertothetruth?Intheabsenceofsomeone to talk to, that’s a question that I askedmyself about three thousandtimes.(Asanaside,askingoneselfanythingthreethousandtimesturnsouttobeabadidea.)Inthiscase,Ihadtwospecificdifficultquestions:

1.Whatiftheofficialinterpretationwaswrong?WhatifIwasmisleadingeveryonefrominvestorstoemployees?Inthatcase,Ishouldberemovedfrommypositionimmediately.2.Whatiftheofficialinterpretationwasright?WhatifIwasgrindingmybrainintosawdustfornoreasonatall?WhatifIwastakingthecompanyofftrackbyquestioningmyowndirection?Inthatcase,Ishouldberemovedfrommypositionimmediately.

As is usually the case, therewas noway to knowwhich interpretationwasright until much later. It turned out that neither was actually right. The newcustomers didn’t save us, but we figured out another way to survive andultimately succeed.The key to getting to the right outcomewas to keep fromgettingmarriedtoeitherthepositiveorthedarknarrative.MyfriendJasonRosenthaltookoverasCEOofNingin2010.Assoonashe

became CEO, he faced a cash crisis and had to choose among three difficultchoices: (1) radically reduce thesizeof thecompany, (2) sell thecompany,or(3)raisemoneyinahighlydilutiveway.Thinkaboutthosechoices:

1.Layoffalargesetoftalentedemployeeswhomheworkedveryhardtorecruitand,asaresult,likelyseverelydamagethemoraleoftheremainingpeople.2.Selloutalloftheemployeeswhomhehadbeenworkingsidebysidewithforthepastseveralyears(Jasonwaspromotedintotheposition)bysellingthecompanywithoutgivingthemachancetoperformorfulfilltheirmission.3.Drasticallyreducetheownershippositionoftheemployeesandmaketheirhardworkeconomicallymeaningless.

Choiceslikethesecausemigraineheadaches.Tiptoaspiringentrepreneurs:Ifyoudon’tlikechoosingbetweenhorribleandcataclysmic,don’tbecomeCEO.Jason sought advice from some of the best minds in the industry, but

ultimatelyhewascompletelyaloneinthefinaldecision.Nobodyhadtheanswerand whatever the answer, Jason was the one who had to live with theconsequences.Sofarhisdecision to reducestaffby lettinggoofprimarily themost recenthireshaspaidoff.RevenueatNing is soaringand teammorale ishigh.Ifithadgoneworse(orultimatelygoesbad),itwouldbeallJason’sfaultanditwouldbeuptoJasontofindanewanswer.WheneverIseeJason,Iliketosay,“Welcometotheshow.”JasoneventuallysoldNingtoGlamandwentontobecomeCEOofLytro.Attimeslikethis,it’simportanttounderstandthatnearlyeverycompanygoes

through life-threatening moments. My partner at Andreessen Horowitz, ScottWeiss,relayedthatit’ssocommonthatthereisanacronymforit,WFIO,whichstands for “We’re Fucked, It’s Over” (it’s pronounced “whiff-ee-yo”). As hedescribes it, every company goes through at least two and up to five of theseepisodes (although I’m pretty sure that I went through at least a dozen atOpsware).Inallcases,WFIOsfeelmuchworsethantheyare—especiallyfortheCEO.

TECHNIQUESTOCALMYOURNERVES

The problemwith psychology is that everybody’s is different.With that as acaveat,over theyears Idevelopeda few techniques fordealingwithmyself. Ihopeyoufindthemuseful,too.Makesomefriends.Althoughit’snearlyimpossibletogethigh-qualityadvice

onthetoughdecisionsthatyoumake,itisextremelyusefulfromapsychologicalperspective to talk to people who have been through similarly challengingdecisions.Get itoutofyourheadandontopaper.WhenIhad toexplain tomyboard

that,sincewewereapubliccompany,Ithoughtthatitwouldbebestifwesoldallofourcustomersandallofourrevenueandchangedbusiness,itwasmessingwith my mind. In order to finalize that decision, I wrote down a detailedexplanation of my logic. The process of writing that document separated mefrommyownpsychologyandenabledmetomakethedecisionswiftly.Focusontheroad,notthewall.Whensomeonelearnstodrivearacecar,one

ofthefirstlessonstaughtisthatwhenyouaregoingaroundacurveat200mph,

donotfocusonthewall; focusontheroad.Ifyoufocusonthewall,youwilldriverightintoit.Ifyoufocusontheroad,youwillfollowtheroad.Runningacompanyislikethat.Therearealwaysathousandthingsthatcangowrongandsinktheship.Ifyoufocustoomuchonthem,youwilldriveyourselfnutsandlikelycrashyourcompany.Focusonwhereyouaregoingrather thanonwhatyouhopetoavoid.

DON’TPUNKOUTANDDON’TQUIT

AsCEO,therewillbemanytimeswhenyoufeellikequitting.IhaveseenCEOstrytocopewiththestressbydrinkingheavily,checkingout,andevenquitting.Ineachcase,theCEOhasamarvelousrationalizationaboutwhyitwasokayforhimtopunkoutorquit,butnoneofthemwilleverbegreatCEOs.Great CEOs face the pain. They deal with the sleepless nights, the cold

sweats,andwhatmyfriend thegreatAlfredChuang(legendarycofounderandCEOofBEASystems)calls“thetorture.”WheneverImeetasuccessfulCEO,Iaskthemhowtheydidit.MediocreCEOspointtotheirbrilliantstrategicmovesor their intuitive business sense or a variety of other self-congratulatoryexplanations.ThegreatCEOstendtoberemarkablyconsistentintheiranswers.Theyallsay,“Ididn’tquit.”

THEFINELINEBETWEENFEARANDCOURAGE

“Itellmykids,whatisthedifferencebetweenaheroandacoward?Whatis the difference between being yellow and being brave? No difference.Only what you do. They both feel the same. They both fear dying andgettinghurt.Themanwhoisyellowrefusestofaceuptowhathe’sgottoface.Thehero ismoredisciplinedandhe fights those feelingsoffandhedoes what he has to do. But they both feel the same, the hero and thecoward. Peoplewhowatch you judge you onwhat you do, not how youfeel.”

—CUSD’AMATO,LEGENDARYBOXINGTRAINER

WhenmypartnersandImeetwithentrepreneurs,thetwokeycharacteristicsthatwelookforarebrillianceandcourage. InmyexperienceasCEO,I foundthatthemostimportantdecisionstestedmycouragefarmorethanmyintelligence.The right decision is often obvious, but the pressure to make the wrong

decisioncanbeoverwhelming.Itstartswithsmallthings.Whenfounderscomein topitchourfirm—oneas theCEOandtheotheras

president—the conversation often goes like this: “Who is running thecompany?”“Weare,”theybothsay.“Whomakesthefinaldecision?”“Wedo.”“Howlongdoyouexpecttorunthatway?”“Forever.”“Soyou’vedecidedtomakeitmoredifficultforeveryemployeetogetwork

donesothatyoudon’thavetodecidewhoisincharge,isthatright?”Thatusuallyresultsinsilence.Intellectually, it should be clear that it is easier for employees to go to one

decisionmaker than two. It’snot reallyverycomplicatedatall.Unfortunately,theclearandpresentsocialpressureoftenoverwhelmsthelong-termbenefitsof

organizingthecompanyproperly.Becausethefoundersdonothavethecouragetodecidewhoisincharge,everyemployeesufferstheinconvenienceofdoubleapproval.More important, decisions only get scarier as a company grows.When we

decidedtotakeLoudcloudpublicwithonly$2millioninrevenue,itwasnotahard choice intellectually—the alternative was to go bankrupt. It wasnonetheless terrifying to do something that most employees, everyone in thepress,andmanyinvestorsthoughtwasnuts.

WHENMAKINGTHERIGHTCHOICEREQUIRESINTELLIGENCEANDCOURAGE

Sometimes the decision itself is rather complicated,whichmakes the couragechallengeevenmoredifficult.CEOspossessadifferentsetofdata,knowledge,and perspective than anybody else in the company. Frequently, some of theemployeesandboardmembersaremoreexperiencedandmore intelligent thantheCEO.Theonly reason theCEOcanmakeabetterdecision isher superiorknowledge.Tomakemattersworse,when aCEO faces a particularly difficult decision,

she may have only a slight preference for one choice over another—say 54percentkillaproduct line,46percentkeepit.If thereallysmartpeopleontheboardandonher staff take theother side,hercouragewillbe severely tested.Howcanshekilltheproductwhensheisnotevensureifsheismakingtherightdecisionandeveryoneisagainsther?Ifshe’swrong,shewillhavebeenwrongin the face of advice fromher top advisers. If she is right,will anybody evenknow?Recently,alargecompanyofferedtobuyoneofourportfoliocompanies.The

deal was lucrative and compelling given the portfolio company’s progress todate and revenue level. The founder/CEO (I’ll call him Hamlet—not his realname) thought that selling did not make sense due to the giant marketopportunitythathewaspursuing,buthestillwantedtomakesurethathemadethe best possible choice for investors and employees.Hamletwanted to rejecttheoffer,butonlymarginally.Tocomplicatematters,mostofthemanagementteamandtheboardthought theopposite. Itdidnothelp that theboardandthemanagementteamwerefarmoreexperiencedthanHamlet.Asaresult,Hamletspentmany sleepless nightsworrying aboutwhetherhewas right.He realizedthatitwasimpossibletoknow.Thisdidnothelphimsleep.Intheend,Hamlet

made the best and most courageous decision he could and did not sell thecompany.Ibelievethatwillprovetobethedefiningmomentofhiscareer.Interestingly, as soon as Hamlet made the decision, the entire board and

executive team immediatelyembraced thechoice.Why? If theywanted to sellthecompanyenough toadvise theCEO togiveuphisdream,howcould theyreverse themselves so quickly? It turns out that themost important data pointdriving their earlier preference for selling the company was Hamlet’s initialambivalence—the team supported the decision they thought the CEOwanted.Hamletdidnotrealizethisandinterpretedtheirdesiretoselltobetheresultofathoroughanalysis.Luckilyforeverybodyinvolved,hehadthecouragetomaketherightdecision.The general problem can be seen in the social credit matrix below. The

expectedsocialrewardsformakingthecrowd-influenceddecisionappearbetterthanthoseformakingthedecisionyouthinkisright:

Youareright Youarewrong

Youdecideagainstthecrowd.

Fewrememberthatyoumadethedecision,butthecompanysucceeds.

Everybodyremembersthedecisionandyouaredowngraded,ostracized,orfired.

Youdecidewiththecrowd.

Everyonewhoadvisedyouremembersthedecisionandthecompanysucceeds.

Youreceivetheminimumblamepossibleforgettingitwrong,butthecompanysuffers.

Onthesurface,itappearsthatifthedecisionisaclosecall,it’smuchsaferto

gowiththecrowd.Inreality,ifyoufallintothistrap,thecrowdwillinfluenceyourthinkingandmakea70-30decisionseemlikea51-49decision.Thisiswhycourageiscritical.

COURAGE,LIKECHARACTER,CANBEDEVELOPED

InallthedifficultdecisionsthatImadethroughthecourseofrunningLoudcloudandOpsware,Ineveroncefeltbrave.Infact,Ioftenfeltscaredtodeath.Ineverlost those feelings, but after much practice I learned to ignore them. Thatlearningprocessmightalsobecalledthecouragedevelopmentprocess.In life, everybody faces choices between doing what’s popular, easy, and

wrongversusdoingwhat’slonely,difficult,andright.Thesedecisionsintensify

when you run a company, because the consequences get magnified athousandfold. As in life, the excuses for CEOsmaking the wrong choice arealwaysplentiful.LifeExcuse CEOExcuse

Othersmartpeoplemadethesamemistake.

Itwasaclosecall.

Allmyfriendswantedtodoit.

TheteamwasagainstmeandIcouldn’tgoagainsttheteam.

Allthecoolkidsaredoingit.

Itwasindustrybestpractice;Ididn’trealizeitwasillegal.

Itwasn’tperfect,soIdecidednottocompete.

Weneverachievedtotalproduct-marketfit,sowenevertriedtosellourproduct.

Every time you make the hard, correct decision you become a bit more

courageousandeverytimeyoumaketheeasy,wrongdecisionyoubecomeabitmore cowardly. If you are CEO, these choices will lead to a courageous orcowardlycompany.

LASTTHOUGHT

Over thepast tenyears, technological advanceshavedramatically lowered thefinancialbarforstartinganewcompany,butthecouragebarforbuildingagreatcompanyremainsashighasithaseverbeen.

ONESANDTWOS

Jim Collins, in his bestselling book Good to Great, demonstrates throughmassive research and comprehensive analysis that when it comes to CEOsuccession,internalcandidatesdramaticallyoutperformexternalcandidates.Thecore reason is knowledge. Knowledge of technology, prior decisions, culture,personnel, and more tends to be far more difficult to acquire than the skillsrequired to manage a larger organization. Collins does not, however, explainwhy internal candidates sometimes fail aswell. Iwill attempt to do sohere. Iwill focus thediscussionon two core skills for running anorganization:First,knowingwhattodo.Second,gettingthecompanytodowhatyouknow.WhilebeingagreatCEOrequiresbothskills,mostCEOstendtobemorecomfortablewithoneortheother.IcallmanagerswhoarehappiersettingthedirectionofthecompanyOnesand thosewhomoreenjoymaking thecompanyperformat thehighestlevelTwos.

WHATONESLIKEANDDON’TLIKE

Oneslikespendingmostoftheirtimegatheringinformationfromabroadvarietyof sources, from employees to customers to competitors. Ones love makingdecisions.Although theyprefer tohave comprehensive informationwhen theymakeadecision, they comfortablymakedecisionswithvery little informationwhennecessary.Oneshavegreatstrategicmindsandenjoynothingmorethanagoodgameofeight-dimensionalchessagainsttheirbestcompetitors.Ones sometimes get bored with many of the important execution details

required to run a company, such as process design, goal setting, structuredaccountability,training,andperformancemanagement.MostfoundingCEOstendtobeOnes.WhenfoundingCEOsfail,asignificant

reasonis that theynever investedthe timetobecompetentenoughin theTwotasks todirect thoseactivitieseffectively.TheresultingcompaniesbecometoochaotictoreachtheirfullpotentialandtheCEOendsupbeingreplaced.

WHATTWOSLIKEANDDON’TLIKE

Twos,ontheotherhand,thoroughlyenjoytheprocessofmakingthecompanyrunwell.They insist upon super-clear goals and stronglyprefer not to changegoalsordirectionunlessabsolutelynecessary.Twosliketoparticipateinstrategicdiscussionsbutoftenhavedifficultywith

thestrategicthinkingprocessitself.WhereaOnemightbeperfectlycomfortablespendingonedayaweekreading,studying,andthinking,doingsowouldmakeaTwoverynervous,becauseitwouldnotfeellikeworktothem.ATwowouldgetantsyatthethoughtofalltheprocessesthatmightbeimproved,peoplewhomightbeheldaccountabletoachievingthestandard,orsalescallsthatcouldbemadewhilehewaswastingtimejustthinkingaboutstrategy.BigdecisionsworryTwosmuchmore thantheyworryOnes.Circumstances

oftenforcebothOnesandTwostomakecriticaldecisionswithinsufficientdata,butOnesgenerallyfeelfineaboutdoingthatanddonotgetoverlyanxiousaboutthe consequences. Twos, by contrast, can become highly agitated about suchthings and sometimes overcomplicate the decision-making process in order toprovideafalsefeelingofthoroughnessaboutthechoice.CEOs who are Twos, despite their love of action, can sometimes bring

decisionmakinginacompanytoahalt.

YOUNEEDBOTHCHARACTERISTICSTOBEAGOODCEO

While people tend to beOnesorTwos,with discipline andhardworknaturalTwoscanbecompetentatOnetasksandOnescanbecompetentatTwotasks.Ifa CEO ignores the dimension of management she doesn’t like, she generallyfails.OnesendupinchaosandTwosfailtopivotwhennecessary.

FUNCTIONALONES

OftenTwoexecutivesactasOnesfortheirfunctions,butTwosasmembersofthe executive team. For example, the head of salesmight easilymake all thedecisionsthatarelocaltothesalesorganizationbutprefertotakedirectionwithrespect to the overall company plans. This is the best kind of multilayerleadershippossible,becausedirectionsareclearanddecisionsaremaderapidlywithprecision.

HOWORGANIZATIONSTENDTOBECONSTRUCTED

Theprimarypurposeof theorganizationalhierarchy inacompany isdecision-

makingefficiency.It followsthatmostCEOstendtobeOnes.If thepersonatthetopofthedecision-makinghierarchydoesn’tlikemakingextremelycomplexdecisions,thecompany’sprocesseswillbeslowandunwieldy.Ifyou’reaOne,itcanbecounterproductivetohaveanotherOneonyourstaff,

because shewillwant to set her own direction rather than follow yours. Thiskindofstrategiccontentioncanconfusetheorganizationandsendemployeesinopposingdirections.Asaresult,manygreatOneCEOsemployprimarilyTwosandFunctionalOnesontheirstaff.

WHATHAPPENSATSUCCESSION?

Thisbringsustothequestionofsuccession.SincemostorganizationsarerunbyOnesandhaveateamofTwos(sometimesFunctionalOnes)reportingtothem,replacingtheCEOcanbeextremelytricky.Doyoupromotesomeonefromtheexecutive staff even though they are likely aTwo?Microsoft did this in2000whentheyreplacedBillGates,aprototypicalOne,withSteveBallmer,literallyhisnumbertwo.OrdoyoureachdeepintotheorganizationandpullaOnefromalevellowerwheretheyarelikelytoexist?GeneralElectricfamouslydidthiswithJackWelchin1981.ItwasanincrediblyboldmovebyGE—notonlydidtheypromoteanexecutivetwolevelsdownintheorganizationalchartpastallofhissuperiors,butindoingsotheynamedtheyoungestCEOinthehistoryofGE.It’s difficult for most boardmembers to even conceive of the possibility thatthereisaOnedeepintheorganizationwhoismorequalifiedtorunthecompanythananyoneontheexecutivestaff.Bothmethodscanbeproblematic.Thefirstapproach leaves thecompany in

thechargeofaTwo.Asthecompanyfacesforks in theroad,decisionmakingmayslowdownandthecompanymayloseitsedge.Inaddition,thenaturalOnes(inMicrosoft’scase,stellarexecutivessuchasPaulMaritzandBradSilverberg)willeventuallyleave.Inscenariotwo,bypromotingsomeonepasteveryoneontheexecutiveteam

andmaking themCEO(asGEdid),youwill likelycausemassive turnoveroftheexecutivestaff. Infact, inveryshortorder,almostnoneof theoriginalGEexecutives remained underWelch. In a diversified conglomerate likeGE, thiskind of rough transition is possible. For companies in the highly dynamictechnologybusiness,thesuper-high-turnoverscenarioismoredangerous.

THEBIGCONCLUSION

Thebigconclusionwillbeabigdisappointmentforthoselookingforananswer.The answer is there is no easy answer. CEO transition is hard. If you bringpeopleinfromtheoutside,youloweryourchancesforsuccess.Ifyoupromotefrom within, you must deal with the One-Two phenomenon. Ideally, you’llpromoteaOneandtherestoftheexecutiveteamwillbegladyoudid.Toobadthingsarerarelyideal.

FOLLOWTHELEADER

There is no prototype for the perfect CEO. Radically different styles—thinkSteve Jobs, Bill Campbell, and AndyGrove—can all lead to great outcomes.Perhaps the most important attribute required to be a successful CEO isleadership.SowhatisleadershipandhowdowethinkaboutitinthecontextoftheCEOjob?Aregreatleadersbornormade?Most people define leadership in the sameway that Supreme Court justice

PotterStewart famouslydefinedpornographywhenhesaid,“IknowitwhenIseeit.”Forourpurposes,wecangeneralizethistobethemeasureofthequalityofaleader:thequantity,quality,anddiversityofpeoplewhowanttofollowher.Sowhatmakespeoplewanttofollowaleader?Welookforthreekeytraits:

TheabilitytoarticulatethevisionTherightkindofambitionTheabilitytoachievethevision

Let’staketheseinorder.

THEABILITYTOARTICULATETHEVISION:THESTEVEJOBSATTRIBUTE

Can the leader articulate a vision that’s interesting, dynamic, and compelling?Moreimportant,cantheleaderdothiswhenthingsfallapart?Morespecifically,whenthecompanygetstoapointwhenitdoesnotmakefinancialsenseforanyemployeetocontinueworkingthere,willtheleaderbeabletoarticulateavisionthat’scompellingenoughtomakepeoplestay?IbelieveJobs’sgreatestachievementasavisionary leaderwas ingettingso

manysuper-talentedpeople tocontinue followinghimatNeXT, longafter thecompany lost itspatina,and ingetting theemployeesofApple tobuy intohisvision when the company was weeks away from bankruptcy. It’s difficult toimagine any other leader being so compelling that he could accomplish thesegoalsback-to-back,andthisiswhywecallthisonetheSteveJobsattribute.

THERIGHTKINDOFAMBITION:THEBILLCAMPBELLATTRIBUTE

One of the biggest misperceptions in our society is that a prerequisite forbecomingaCEO is tobe selfish, ruthless, andcallous. In fact, theopposite istrueandthereasonisobvious.ThefirstthingthatanysuccessfulCEOmustdoisgetreallygreatpeopletoworkforher.Smartpeopledonotwanttoworkforpeoplewhodonothavetheirinterestsinmindandinheart.Most of us have experienced this in our careers: a bright, ambitious,

hardworking executive whom nobody good wants to work for and who, as aresult,deliversperformancefarworsethanonemightimagine.Trulygreat leaderscreateanenvironmentwhere theemployees feel that the

CEOcaresmoreabouttheemployeesthanshecaresaboutherself.Inthiskindof environment, an amazing thing happens: A huge number of employeesbelieveit’stheircompanyandbehaveaccordingly.Asthecompanygrowslarge,theseemployeesbecomequalitycontrolfortheentireorganization.Theysettheworkstandardthatallfutureemployeesmustliveupto.Asin,“Hey,youneedtodoabetterjobonthatdatasheet—youarescrewingupmycompany.”Icall thischaracteristic theBillCampbellattributeafter themanwho is the

bestI’veeverseenatthis.Ifyoutalktopeoplewhoworkedinanyofthemanyorganizations that Bill has run, they refer to those organizations as “myorganization”or“mycompany.”Ahugepartofwhyhehasbeensoremarkablystrong in this dimension of leadership is that he’s completely authentic. Hewould happily sacrifice his own economics, fame, glory, and rewards for hisemployees.WhenyoutalktoBill,yougetthefeelingthathecaresdeeplyaboutyouandwhatyouhavetosay,becausehedoes.Andallofthatshowsupinhisactionsandfollow-through.

THEABILITYTOACHIEVETHEVISION:THEANDYGROVEATTRIBUTE

Thefinallegofourleadershipstooliscompetence,pureandsimple.IfIbuyintothevisionandbelievethattheleadercaresaboutme,doIthinkshecanactuallyachieve the vision?Will I follow her into the junglewith nomap forward orbackandtrustthatshewillgetmeoutofthere?IliketorefertothisastheAndyGroveattribute.AndyGrovewillalwaysbe

my model of CEO competence. He earned a Ph.D. in electrical engineering,wrotethebestmanagementbookI’veeverread(HighOutputManagement),and

tirelessly refined his craft. Not only did he write exceptional books onmanagement,buthetaughtmanagementclassesatIntelthroughouthistenure.Inhis classicbookOnly theParanoidSurvive,Grove tellshowhe led Intel

throughthedramatictransitionfromthememorybusinesstothemicroprocessorbusiness.Inmakingthatchangehewalkedawayfromnearlyallhisrevenue.Hehumbly credits others in the companywith coming to the strategic conclusionbefore he did, but the credit for swiftly and successfully leading the companythrough the transitiongoes toDr.Grove.Changingyourprimarybusinessasasixteen-year-oldlarge,publiccompanyraisesalotofquestions.Andydescribesanincidentwithoneofhisemployees:“Oneofthemattacked

meaggressively, asking, ‘Does itmean thatyoucanconceiveof Intelwithoutbeinginthememorybusiness?’Iswallowedhardandsaid,‘Yes,IguessIcan.’Allhellbrokeloose.”Despite shocking many of his best employees with this radical strategy,

ultimatelythecompanytrustedAndy.Theytrustedhimtorebuildtheircompanyaroundanentirelynewbusiness.Thattrustturnedouttobeverywellplaced.

SO,AREGREATLEADERSBORNORMADE?

Let’slookatthisoneattributeatatime:

ArticulatingthevisionThereisnoquestionthatsomepeoplearemuchbetterstorytellersthanothers.However,itisalsotruethatanybodycangreatlyimproveinthisareathroughfocusandhardwork.AllCEOsshouldworkonthevisioncomponentofleadership.AlignmentofinterestsIamnotsureiftheBillCampbellattributeisimpossibletolearn,butIamprettysurethatitisimpossibletoteach.Ofthethree,thisonemostfitsthebill“bornnotmade.”AbilitytoachievethevisionThisattributecanabsolutelybelearned;perhapsthisiswhyAndyGrove’stoleranceforincompetencewaslegendarilylow.Indeed,theenemyofcompetenceissometimesconfidence.ACEOshouldneverbesoconfidentthatshestopsimprovingherskills.

In theend, someattributesof leadershipcanbe improvedmore thanothers,buteveryCEOshouldworkonall three.Furthermore,eachattributeenhancesall three. If people trust you, they will listen to your vision even if it is less

articulate. If you are super-competent, theywill trust you and listen to you. Ifyoucanpaintabrilliantvision,peoplewillbepatientwithyouasyoulearntheCEOskillsandgiveyoumoreleewaywithrespecttotheirinterests.

PEACETIMECEO/WARTIMECEO

BillCampbellalwaysusedtosaytome,“Ben,you’rethebestCEOthatIworkwith.” This always seemed crazy to me, because he was working with SteveJobs, JeffBezos, andEric Schmidt at the timewhilemy companywas goingstraightintothewall.OnedayIcalledhimonitandsaid,“Bill,whywouldyousay that? Do results not count?” He said, “There are lots of good peacetimeCEOsandlotsofgoodwartimeCEOs,butalmostnoCEOsthatcanfunctioninbothpeacetimeandinwartime.You’reapeacetime/wartimeCEO.”Bymycalculation,IwasapeacetimeCEOforthreedaysandwartimeCEO

foreightyears.Istillhaveahardtimeshakingthewartimeflashbacks.I’mnotthe only one who has experienced this. Dennis Crowley, the founder ofFoursquare, told me that he thinks about this tension—between wartime andpeacetime—everyday.Thesamegoesforalotoftechcompanies.For instance, when Eric Schmidt stepped down as CEO of Google and

founder Larry Page took over,much of the news coverage focused on Page’sabilitytobethe“faceofGoogle”sincePageisfarmoreshyandintrovertedthanthe gregarious and articulateSchmidt.While an interesting issue, this analysismisses themain point. Schmidt wasmuchmore than Google’s front man; asGoogle’s peacetime chief executive, he led the greatest technology businessexpansioninthelasttenyears.LarryPage,incontrast,seemstohavedeterminedthatGoogleismovingintowarandheclearlyintendstobeawartimeCEO.ThishasbeenaprofoundchangeforGoogleandtheentirehigh-techindustry.

DEFINITIONSANDEXAMPLES

Peacetimeinbusinessmeansthosetimeswhenacompanyhasalargeadvantageover thecompetition in itscoremarket, and itsmarket isgrowing. In timesofpeace, the company can focus on expanding the market and reinforcing thecompany’sstrengths.Inwartime,acompany is fendingoffan imminentexistential threat.Sucha

threatcancomefromawiderangeofsources, includingcompetition,dramaticmacroeconomicchange,marketchange,supplychainchange,andsoforth.The

greatwartimeCEOAndyGrovemarvelouslydescribestheforcesthatcantakeacompanyfrompeacetimetowartimeinhisbookOnlytheParanoidSurvive.A classic peacetime mission is Google’s effort to make the Internet faster.

Google’spositioninthesearchmarketissodominantthattheydeterminedthatanything thatmakes the Internet faster accrues to theirbenefit since it enablesusers to do more searches. As the clear market leader, they focus more onexpanding themarket thandealingwith theirsearchcompetitors. Incontrast,aclassic wartime mission was Andy Grove’s drive to get out of the memorybusiness in the mid-1980s due to an irrepressible threat from the Japanesesemiconductorcompanies.Inthismission,thecompetitivethreat—whichcouldhave bankrupted the company—was so great that Intel had to exit its corebusiness,whichemployed80percentofitsstaff.Mygreatestmanagementdiscoverythroughthetransitionwasthatpeacetime

andwartime require radically differentmanagement styles. Interestingly,mostmanagementbooksdescribepeacetimeCEO techniques andvery fewdescribewartime.Forexample,abasicprincipleinmostmanagementbooksisthatyoushouldneverembarrassanemployeeinapublicsetting.Ontheotherhand,inaroomfilledwithpeople,AndyGroveoncesaidtoanemployeewhoenteredthemeeting late,“All Ihave in thisworld is time,andyouarewastingmy time.”Whysuchdifferentapproachestomanagement?Inpeacetime,leadersmustmaximizeandbroadenthecurrentopportunity.As

a result, peacetime leaders employ techniques to encourage broad-basedcreativity and contribution across a diverse set of possible objectives. Inwartime,bycontrast, thecompany typicallyhasa singlebullet in thechamberandmust,atallcosts,hitthetarget.Thecompany’ssurvivalinwartimedependsuponstrictadherenceandalignmenttothemission.When Steve Jobs returned to Apple, the company was weeks away from

bankruptcy—a classic wartime scenario. He needed everyone to move withprecisionandfollowhisexactplan;therewasnoroomforindividualcreativityoutsidethecoremission.Instarkcontrast,asGoogleachieveddominanceinthesearchmarket,Google’smanagementfosteredpeacetimeinnovationbyenablingand even requiring every employee to spend 20 percent of their time on theirownnewprojects.Peacetimeandwartimemanagement techniquescanbothbehighlyeffective

whenemployedintherightsituations,buttheyareverydifferent.ThepeacetimeCEOdoesnotresemblethewartimeCEO.

PEACETIMECEO/WARTIMECEO

Peacetime CEO knows that proper protocol leads to winning. Wartime CEOviolatesprotocolinordertowin.PeacetimeCEOfocusesonthebigpictureandempowersherpeopletomake

detaileddecisions.WartimeCEOcaresaboutaspeckofdustonagnat’sassifitinterfereswiththeprimedirective.Peacetime CEO builds scalable, high-volume recruiting machines.Wartime

CEOdoesthat,butalsobuildsHRorganizationsthatcanexecutelayoffs.PeacetimeCEOspendstimedefiningtheculture.WartimeCEOlets thewar

definetheculture.Peacetime CEO always has a contingency plan. Wartime CEO knows that

sometimesyougottarollahardsix.Peacetime CEO knows what to do with a big advantage.Wartime CEO is

paranoid.PeacetimeCEO strives not to use profanity.WartimeCEO sometimes uses

profanitypurposefully.PeacetimeCEO thinksof the competition asother ships in a bigocean that

may never engage.Wartime CEO thinks the competition is sneaking into herhouseandtryingtokidnapherchildren.PeacetimeCEO aims to expand themarket.WartimeCEO aims towin the

market.PeacetimeCEOstrivestotoleratedeviationsfromtheplanwhencoupledwith

effortandcreativity.WartimeCEOiscompletelyintolerant.Peacetime CEO does not raise her voice.Wartime CEO rarely speaks in a

normaltone.Peacetime CEO works to minimize conflict. Wartime CEO heightens the

contradictions.Peacetime CEO strives for broad-based buy-in. Wartime CEO neither

indulgesconsensusbuildingnortoleratesdisagreements.PeacetimeCEO sets big, hairy, audacious goals.WartimeCEO is too busy

fightingtheenemytoreadmanagementbookswrittenbyconsultantswhohavenevermanagedafruitstand.Peacetime CEO trains her employees to ensure satisfaction and career

development.WartimeCEO trainsheremployees so theydon’tget their assesshotoffinthebattle.PeacetimeCEOhasruleslike“We’regoingtoexitallbusinesseswherewe’re

notnumberoneortwo.”WartimeCEOoftenhasnobusinessesthatarenumberoneortwoandthereforedoesnothavetheluxuryoffollowingthatrule.

CANACEOBEBOTH?

CanaCEObuildtheskillsetstoleadinbothpeacetimeandwartime?Onecouldeasilyargue that I failedasapeacetimeCEObutsucceededasa

wartimeone.JohnChambershadagreatrunaspeacetimeCEOofCiscobuthasstruggled asCisco hasmoved intowarwith Juniper,HP, and a range of newcompetitors.Steve Jobs,whoemployedaclassicalwartimemanagement style,removedhimself asCEOofApple in the1980sduring their longestperiodofpeace before coming back toApple for a spectacular runmore than a decadelater,duringtheirmostintensewarperiod.I believe that the answer is yes, but it’s hard.Mastering both wartime and

peacetime skill sets means understanding the many rules of management andknowingwhentofollowthemandwhentoviolatethem.Be aware that management books tend to be written by management

consultantswho study successful companies during their times of peace.As aresult, the resulting books describe the methods of peacetime CEOs. In fact,otherthanthebookswrittenbyAndyGrove,Idon’tknowofanymanagementbooksthatteachyouhowtomanageinwartimelikeSteveJobsorAndyGrove.

BACKTOTHEBEGINNING

It turnedout thata littlewartimewas justwhat thedoctororderedforGoogle.Page’s precise and exacting leadership has led to brilliant execution inintegratingidentityacrossGoogle’sbroadproductline,fromtheriseofAndroidtobrilliantnewproductslikeGoogleGlass.Sometimesyouneedtogotowar.

MAKINGYOURSELFACEO

Theotherday,afriendofmineaskedmewhetherCEOswerebornormade.Isaid,“That’skindoflikeaskingifJollyRanchersaregrownormade.CEOisanunnatural job.”The surprised lookonhis facemademe realize thatperhaps itwasn’tasobviousasI’doriginallythought.Mostpeopleactuallyassumetheopposite—CEOsareborn,notmade.Ioften

listenasotherventurecapitalistsandboardmembersrapidlyevaluateafounderandconcludethatshe’snot“CEOmaterial.”Iamnotsurehowtheyfigurethesethingsoutsofast.ItgenerallytakesyearsforafoundertodeveloptheCEOskillsetanditisusuallyextremelydifficultformetotellwhethershewillmakeit.In athletics, some things, likebecominga sprinter, canbe learned relatively

quickly because they start with a natural motion and refine it. Others, likeboxing, take much longer to master because they require lots of unnaturalmotions and lots of specific technique. For example, as I mentioned earlier,when going backward in boxing it’s critically important to pick up your backfoot first, because if you get hit while walking backward the natural way—pickingupyourfrontfootfirst—itoftenleadstogettingknockedcold.Learningtomakethisunnaturalmotionfeelnaturaltakesagreatdealofpractice.IfyoudowhatfeelsmostnaturalasaCEO,youmayalsogetknockedcold.Being CEO requires lots of unnatural motion. From an evolutionary

standpoint,itisnaturaltodothingsthatmakepeoplelikeyou.Itenhancesyourchancesforsurvival.YettobeagoodCEO,inordertobelikedinthelongrun,you must do many things that will upset people in the short run. Unnaturalthings.EventhemostbasicCEObuildingblockswillfeelunnaturalatfirst.Ifyour

buddy tells you a funny story, it would feel quite weird to evaluate herperformance. It would be totally unnatural to say, “Gee, I thought that storyreallysucked.Ithadpotential,butyouwereunderwhelmingonthebuildupandthenyou totally flubbed thepunch line. I suggest thatyougoback, rework it,andpresentittomeagaintomorrow.”Doing sowould be quite bizarre, but evaluating people’s performances and

constantlygivingfeedbackispreciselywhataCEOmustdo.Ifshedoesn’t,the

morecomplexmotionssuchaswritingreviews,takingawayterritory,handlingpolitics, setting compensation, and firing people will be either impossible orhandledratherpoorly.Giving feedback turns out to be the unnatural atomic building block atop

whichtheunnaturalskillsetofmanagementgetsbuilt.Buthowdoesonemastertheunnatural?

THESHITSANDWICH

A popular and sometimes effective technique for feedback beginners issomething that experiencedmanagers call theShit Sandwich.The technique ismarvelouslydescribedintheclassicmanagementtextTheOneMinuteManager.The basic idea is that people open up to feedback far more if you start bycomplimenting them (slice of bread number one), then you give them thedifficult message (the shit), then wrap up by reminding them howmuch youvaluetheirstrengths(sliceofbreadnumbertwo).Theshitsandwichalsohasthepositive side effect of focusing the feedback on the behavior rather than theperson,becauseyouestablishupfrontthatyoureallyvaluetheperson.Thisisakeyconceptingivingfeedback.Theshitsandwichcanworkwellwithjunioremployeesbuthasthefollowing

challenges:

Ittendstobeoverlyformal.Becauseyouhavetopreplanandscriptthesandwichtomakeitcomeoutcorrectly,theprocesscanfeelformalandjudgmentaltotheemployee.Afteryoudoitacoupleoftimes,itwilllackauthenticity.Theemployeewillthink,“Ohboy,she’scomplimentingmeagain.Iknowwhat’scomingnext,theshit.”Moreseniorexecutiveswillrecognizetheshitsandwichimmediatelyanditwillhaveaninstantnegativeeffect.

Earlyinmycareer,IattemptedtodeliveracarefullycraftedshitsandwichtoasenioremployeeandshelookedatmelikeIwasalittlekidandsaid,“Sparemethecompliment,Ben,andjusttellmewhatIdidwrong.”Atthatpoint,IthoughtthatIwasdefinitelynotborntobeaCEO.

THEKEYS

Tobecomeelite at giving feedback,youmust elevateyourself beyondabasictechnique like the shit sandwich.Youmust develop a style thatmatches yourownpersonalityandvalues.Herearethekeystobeingeffective:

Beauthentic.It’sextremelyimportantthatyoubelieveinthefeedbackthatyougiveandnotsayanythingtomanipulatetherecipient’sfeelings.Youcan’tfakethefunk.Comefromtherightplace.It’simportantthatyougivepeoplefeedbackbecauseyouwantthemtosucceedandnotbecauseyouwantthemtofail.Ifyoureallywantsomeonetosucceed,thenmakeherfeelit.Makeherfeelyou.Ifshefeelsyouandyouareinhercorner,thenshewilllistentoyou.Don’tgetpersonal.Ifyoudecidetofiresomebody,fireher.Don’tpreparehertogetfired.Preparehertosucceed.Ifshedoesn’ttakethefeedback,that’sadifferentconversation.Don’tclownpeopleinfrontoftheirpeers.Whileit’sokaytogivecertainkindsoffeedbackinagroupsetting,youshouldstrivenevertoembarrasssomeoneinfrontoftheirpeers.Ifyoudoso,thenyourfeedbackwillhavelittleimpactotherthantocausetheemployeetobehorriblyashamedandtohateyourguts.Feedbackisnotone-size-fits-all.Everybodyisdifferent.Someemployeesareextremelysensitivetofeedbackwhileothershaveparticularlythickskinandoftenthickskulls.Stylistically,yourtoneshouldmatchtheemployee’spersonality,notyourmood.Bedirect,butnotmean.Don’tbeobtuse.Ifyouthinksomebody’spresentationsucks,don’tsay,“It’sreallygood,butcoulduseonemorepasstotightenuptheconclusion.”Whileitmayseemharsh,it’smuchbettertosay,“Icouldn’tfollowitandIdidn’tunderstandyourpointandherearethereasonswhy.”Watered-downfeedbackcanbeworsethannofeedbackatallbecauseit’sdeceptiveandconfusingtotherecipient.Butdon’tbeatthemuporattempttoshowyoursuperiority.Doingsowilldefeatyourpurposebecausewhendoneproperly,feedbackisadialogue,notamonologue.

FEEDBACKISADIALOGUE,NOTAMONOLOGUE

Youmaybe theCEOandyoumaybe telling somebodyabout something thatyou don’t like or disagree with, but that doesn’t mean you’re right. Your

employeeshouldknowmoreaboutherfunctionthanyou.Sheshouldhavemoredatathanyou.Youmaybewrong.Asaresult,yourgoalshouldbeforyourfeedbacktoopenupratherthanclose

downdiscussion.Encourage people to challenge your judgment and argue thepoint to conclusion.Culturally, youwanthigh standards thoroughlydiscussed.Youwanttoapplytremendouspressuretogetthehighest-qualitythinkingyetbeopenenoughtofindoutwhenyouarewrong.

HIGH-FREQUENCYFEEDBACK

Onceyou’vemastered thekeys,you shouldpracticewhatyou’vemasteredallthe time.AsCEO,you shouldhave anopinionon absolutely everything.Youshould have an opinion on every forecast, every product plan, everypresentation,andeveneverycomment.Letpeopleknowwhatyouthink.Ifyoulike someone’s comment, give her the feedback. If you disagree, give her thefeedback.Saywhatyouthink.Expressyourself.Thiswillhavetwocriticallyimportantpositiveeffects:

Feedbackwon’tbepersonalinyourcompany.IftheCEOconstantlygivesfeedback,theneveryonesheinteractswithwilljustgetusedtoit.Nobodywillthink,“Gee,whatdidshereallymeanbythatcomment?Doesshenotlikeme?”Everybodywillnaturallyfocusontheissues,notanimplicitrandomperformanceevaluation.Peoplewillbecomecomfortablediscussingbadnews.Ifpeoplegetcomfortabletalkingaboutwhateachotheraredoingwrong,thenitwillbeveryeasytotalkaboutwhatthecompanyisdoingwrong.High-qualitycompanyculturesgettheircuefromdatanetworkingroutingprotocols:Badnewstravelsfastandgoodnewstravelsslowly.Low-qualitycompanyculturestakeonthepersonalityoftheWickedWitchoftheWestinTheWiz:“Don’tnobodybringmenobadnews.”

MAKINGTHECEO

Being CEO also requires a broad set ofmore advanced skills, but the key toreaching the advanced level and feeling like you were born to be CEO ismasteringtheunnatural.IfyouareafounderCEOandyoufeelawkwardorincompetentwhendoing

someofthesethingsandbelievethereisnowaythatyou’llbeabletodoitwhen

your company is one hundred or one thousand people, welcome to the club.That’sexactlyhowIfelt.SodideveryCEOI’veevermet.Thisistheprocess.Thisishowyougetmade.

HOWTOEVALUATECEOs

Noposition inacompanyismore important than theCEOand,asaresult,nojobgetsmorescrutiny.Thejobissopoorlydefinedthatyoucanendupdoingallkindsofnuttythings(especiallyifyoulistentosomepeoplewhosaythingslike“theCEOshouldbethenumber-onesalesperson”).Sadly,littleofthisanalysisthat’sbeendonebenefitsCEOs,sincemostofthe

discussionshappenbehindtheirbacks.HereIwanttotakeastepintheoppositedirection.BydescribinghowIevaluateCEOs,IamatthesametimedescribingwhatIthinkthejoboftheCEOis.Herearethekeyquestionsweask:

1.DoestheCEOknowwhattodo?2.CantheCEOgetthecompanytodowhatsheknows?3.DidtheCEOachievethedesiredresultsagainstanappropriatesetofobjectives?

1.DOESTHECEOKNOWWHATTODO?

Oneshould interpret thisquestionasbroadlyaspossible.Does theCEOknowwhat to do in all matters all the time? This includes matters of personnel,financing,product strategy,goal sizing,andmarketing.Atamacro level,doesthe CEO set the right strategy for the company and know its implications ineverydetailofthecompany?Ievaluatetwodistinctfacetsofknowingwhattodo:

StrategyIngoodcompanies,thestoryandthestrategyarethesamething.Asaresult,theproperoutputofallthestrategicworkisthestory.DecisionmakingAtthedetailedlevel,theoutputofknowingwhattodoisthespeedandqualityoftheCEO’sdecisions.

ThestrategyandthestoryThe CEO must set the context within which every employee operates.Thecontextgivesmeaningtothespecificworkthatpeopledo,alignsinterests,enables decision making, and provides motivation. Well-structured goals andobjectives contribute to the context, but they do not provide the whole story.Moretothepoint,theyarenotthestory.Thestoryofthecompanygoesbeyondquarterlyorannualgoalsandgetstothehard-corequestionofwhy.WhyshouldIjointhiscompany?WhyshouldIbeexcitedtoworkhere?WhyshouldIbuyitsproduct?WhyshouldIinvestinthecompany?Why istheworldbetteroffasaresultofthiscompany’sexistence?When a company clearly articulates its story, the context for everyone—

employees,partners,customers,investors,andthepress—becomesclear.Whenacompanyfailstotellitsstory,youhearphraseslike

Thesereportersdon’tgetit.Whoisresponsibleforthestrategyinthiscompany?Wehavegreattechnology,butneedmarketinghelp.

TheCEOdoesn’thavetobethecreatorofthevision.Nordoesshehavetobethecreatorofthestory.Butshemustbethekeeperofthevisionandthestory.Assuch,theCEOensuresthatthecompanystoryisclearandcompelling.Thestoryisnotthemissionstatement;thestorydoesnothavetobesuccinct.

Itisthestory.Companiescantakeaslongastheyneedtotellit,buttheymusttell it and it must be compelling. A company without a story is usually acompanywithoutastrategy.Want to see a great company story?Read Jeff Bezos’s three-page letter he

wrotetoshareholdersin1997.IntellingAmazon’sstoryinthisextendedform—notasamissionstatement,notasatagline—JeffgotallthepeoplewhomatteredonthesamepageastowhatAmazonwasabout.

DecisionmakingSome employees make products, some make sales; the CEO makesdecisions.Therefore,aCEOcanmostaccuratelybemeasuredbythespeedandquality of those decisions. Great decisions come from CEOs who display anelitemixtureofintelligence,logic,andcourage.As already noted, courage is particularly important, because every decision

thataCEOmakesisbasedonincompleteinformation.Atthetimeofanygivendecision, theCEOwill generally have less than 10 percent of the informationtypically present in the post hoc Harvard Business School case study. As aresult, theCEOmusthave thecourage tobet thecompanyonadirectioneventhoughshedoesnotknowif thedirectionisright.Themostdifficultdecisions(andoftenthemostimportant)aredifficultpreciselybecausetheywillbedeeplyunpopularwiththeCEO’smostimportantconstituencies(employees,investors,andcustomers).ThebestdecisionthatImadeinmycareer—toselltheLoudcloudbusinessto

EDSandbecomeOpsware—wouldhavelostbyalandslidehadIputittoavotewithmyemployees,myinvestors,ormycustomers.AsCEO,thereisneverenoughtimetogatherallinformationneededtomake

adecision.Youmustmakehundredsofdecisionsbigandsmallinthecourseofa typical week. You cannot simply stop all other activities to gathercomprehensive data and do exhaustive analysis to make that single decision.Knowing this, youmust continuously and systematically gather knowledge inthecompany’sday-to-dayactivitiessothatyouwillhaveasmuchinformationaspossiblewhenthedecisionpointarrives.Inordertopreparetomakeanydecision,youmustsystematicallyacquirethe

knowledgeofeverything thatmight impactanydecision thatyoumightmake.Questionssuchas:

Whatarethecompetitorslikelytodo?What’spossibletechnicallyandinwhattimeframe?Whatarethetruecapabilitiesoftheorganizationandhowcanyoumaximizethem?Howmuchfinancialriskdoesthisimply?Whatwilltheissuesbe,givenyourcurrentproductarchitecture?

Willtheemployeesbeenergizedordespondentaboutthispromotion?

Great CEOs build exceptional strategies for gathering the requiredinformation continuously. They embed their quest for intelligence into all oftheir daily actions from staff meetings to customer meetings to one-on-ones. Winning strategies are built on comprehensive knowledge gathered inevery interaction theCEOhaswith an employee, a customer, a partner, or aninvestor.

2.CANTHECEOGETTHECOMPANYTODOWHATSHEKNOWS?

If theCEOpaints a compelling vision andmakes fast, high-quality decisions,canshethengetthecompanytoexecutehervision?Thefirstingredientinbeingabletodothisisleadership,asIoutlinedinthesection“FollowtheLeader.”In addition, executing well requires a broad set of operational skills. The

largertheorganization,themoreelaboratetherequisiteskillset.Inorder foracompany toexecuteabroadsetofdecisionsand initiatives, it

must:

Havethecapacitytodoso.Inotherwords,thecompanymustcontainthenecessarytalentintherightpositionstoexecutethestrategy.Beaplacewhereeveryemployeecangetthingsaccomplished.Theemployeesmustbemotivated,communicationmustbestrong,theamountofcommonknowledgemustbevast,andthecontextmustbeclear.

IstheCEObuildingaworld-classteam?TheCEOis responsible for theexecutive teamplus the fundamental interviewand hiring processes for all employees. She must make sure the companysourcesthebestcandidatesandthescreeningprocessesyieldthecandidateswiththerightcombinationof talentandskills.Ensuringthequalityof the teamisacorepartof running the company.GreatCEOsconstantly assesswhether theyarebuildingthebestteam.Theoutputofthiscapabilityisthequalityoftheteam.It’simportanttonote

that team quality is tightly tied to the specific needs of the company in thechallenges it faces at the point in time it faces them. As a result, it’s quitepossible that the executive team changes several times, but the team is high

qualitytheentirewayandthereisnoattritionproblem.

Isiteasyforemployeestocontributetothemission?Thesecondpartof theevaluationdetermineswhether theCEOcaneffectivelyrunthecompany.Totestthis,Iliketoaskthisquestion:“Howeasyisitforanygivenindividualcontributortogetherjobdone?”In well-run organizations, people can focus on their work (as opposed to

politicsandbureaucraticprocedures)andhaveconfidence that if theyget theirwork done, good things will happen both for the company and for thempersonally.Bycontrast,inapoorlyrunorganization,peoplespendmuchoftheirtimefightingorganizationalboundariesandbrokenprocesses.While it may be quite easy to describe, building a well-run organization

requires a high level of skill. The skills required range from organizationaldesigntoperformancemanagement.Theyinvolvetheincentivestructureandthecommunication architecture that drive and enable every individual employee.WhenaCEO“failstoscale,”it’susuallyalongthisdimension.Inpractice,veryfewCEOsgetanAonthisparticulartest.Netflix’sCEO,ReedHastings, put great effort into designing a system that

enablesemployeestobemaximallyeffective.Hispresentationofthisdesigniscalled ReferenceGuide onOur Freedom and Responsibility Culture. It walksthroughwhatNetflixvaluesintheiremployees,howtheyscreenforthosevaluesduring the interview process, how they reinforce those values, and how theyscalethissystemasthenumberofemployeesgrows.

3.DIDTHECEOACHIEVETHEDESIREDRESULTSAGAINSTANAPPROPRIATESETOFOBJECTIVES?

Whenmeasuring resultsagainstobjectives, startbymakingsure theobjectivesare correct. CEOs who excel at board management can “succeed” by settingobjectives artificially low. Great CEOs who fail to pay attention to boardmanagement can “fail” by setting objectives too high. Early in a company’sdevelopment, objectives can be particularly misleading since nobody reallyknows the true size of the opportunity. Therefore, the first task in accuratelymeasuringresultsissettingobjectivescorrectly.I also try to keep inmind that the size andnature of theopportunity varies

quiteabitacrosscompanies.Hopingthatahardwarecompanycanbeascapital-light as a consumer Internet company or trying to get Yelp to grow as fast

as Twitter doesn’t make sense and can be quite destructive. CEOs should beevaluatedagainst theircompany’sopportunity—notsomebodyelse’scompany.Let me share a funny story that illustrates a CEO really owning deliveringagainst results. This story is from Robin Li, CEO of Baidu, China’s biggestsearch engine company. In a lecture he gave at Stanford University in 2009,Robin recalls that on the day of Baidu’s IPO—usually one ofan entrepreneur’smost exhilarating days in his entire life—he sat at his deskterrified.Why?ListentohowRobinowneddeliveringresults:

In 2004, we raised our last round of VC money led by Draper FisherJurvetson...andGoogle,oneofourgreatcolleagues.Thenayearlater,in 2005, the company went public. The ideal price was $27 [the stock’sinitialofferprice]and it closedon the firstdayat$122. Itwasgreat formanyof theBaiduemployeesand forall of theBaidu investors. Itwasaverymiserable thing formebecausewhen Idecided to take the companypublic,Iwasonlypreparedtodeliverfinancialresultsthatmatchthepriceof$27ormaybea littlehigher,$30,$40.ButIwasreallyshockedtoseethat the price went to $122 on the first day. So that meant I needed todeliver real results thatmatched an expectationmuch,much higher thanwhatIhadpreparedtodo.Butinanycase,IthoughtIhadnochoice.SoIput my head down and focused on operations, focused on technology,focusedontheuser’sexperience,andIdelivered.

Once we’ve taken all of this into account, we see that the results againstobjectivesor“blackbox”resultsarealaggingindicator.Andastheysayinthemutualfundprospectuses,“pastperformanceisnoguaranteeoffutureresults.”Thewhite-boxCEOevaluationcriteria—“DoestheCEOknowwhattodo?”and“Can the CEO get the company to do it?”—will do a much better job ofpredictingthefuture.

CLOSINGTHOUGHT

CEOevaluationneednotbeabyzantine,unarticulatedart.Allpeople,includingCEOs,willperformbetteronatestiftheyknowthequestionsaheadoftime.

—CHAPTER8—

FIRSTRULEOFENTREPRENEURSHIP:THEREARENORULES

“Thatthethatthethatthatdon’tkillmeCanonlymakeamestrongerIneedyoutohurryupnow

’CauseIcan’twaitmuchlongerIknowIgottoberightnow

’CauseIcan’tgetmuchwrongerManIbeenwaitin’allnightnowThat’showlongIbeenonya.”

—KANYEWEST,“STRONGER”

Whenwewere in theprocessof sellingOpsware,HP’s initialbidcame inat$14 per share. BMC countered with $14.05. Then HP countered them with$14.25. John O’Farrell and I planned our strategy for closing the biddingprocess.Wefiguredthatifweexecutedcorrectly,thewinningbidwouldcomeinat$15pershareorhigher.Everybodywasextremelyexcited.Thendisasterstruck.Ormorespecifically,ourauditor,Ernst&Young,nearly

destroyedthedeal.DuringBMC’sdiligenceprocess, theydiscoveredthatwehadaccountedfor

three deals differently than they had. Specifically, each of our three dealscontained something that had come to be known in the industry as the “CAclause”inhonoroftheinfamoussoftwarecompanyComputerAssociates,orCAforshort.TheCAclausehadcomeaboutasaresultofsomeofCA’sbusinesspractices. Apparently CA had tricked their customers by selling themmaintenance contracts that gave them rights to free upgrades forever for

productsnamed“X.”CAwouldthenchangethenameofproduct“X”toproduct“Y”andchargetheircustomersforanupgradethecustomersthoughttheywereentitled to for free. Itwas very clever, and totally dirty. To fight back, savvycustomers started requiringall softwarevendors to include“theCAclause” intheir contracts. The clause stated that if you released a new version of yoursoftwarethatcontainsallofthefunctionalityofthepreviousversion,plussomenewfeaturesandanewname,thenthatproduct(despitethenewname)wouldbecoveredundertheexistingcontractwithnoadditionalfeesdue.Therewere two possibleways to interpret theCA clause.You could either

interpret it as itwas intended,asaworkaround forCA’sbadbehavior,oryoucouldinterpretitasapromiseoffuturefunctionality.Ifyouinterpreteditastheformer then you were required to recognize all the revenue up front. If youinterpreteditasthelatter,thenyouwouldrecognizetherevenueratablyoverthecourseofthecontract.Ineithercase,cashpaymentswouldbethesame.Weunderstoodthisambiguitywhenwesignedthethreedealsthatcontained

theclause,sowehadourE&Ypartner,DavePrice,auditthedealsandthentellus howwe should account for them.Dave correctlyunderstood the intent andrecommendedthatwetaketherevenueupfrontinallthreedeals.BMC’sE&Ypartner concluded the opposite and they accounted for similar deals ratably.Whenalertedtothedifference,theBMCpartnerescalatedtheissuetoErnst&Young’snationaloffice.E&Y’snationalauditpartnercalledmetoreport thathenowdisagreedwith

his partner’s audit and required that we restate revenue guidance in the nextforty-eight hours. I couldn’t believe what I had just heard. Restating revenuewouldtankthestockpriceanddestroythependingsale.TheaccountinghadnomaterialimpactoncashflowsandourtreatmentwasbasedonthejudgmentofErnst & Young in the first place. If they had accounted for it the other wayinitially, thestockpricewouldbethesameasit isnow.Itwastherestatementthatwouldkillus.Whatthehell?Icalmeddownandrepliedcarefully.Ben: “The accounting is meant to reflect the intent of both us and our

customersinthecontract,correct?”E&Y:“Correct.”Ben:“So,whydon’twecallall threecustomerson thephoneandask them

their intent? If it’swhatDavePrice reflected inourcontracts, thenwe’ll leavethemastheyare.Ifit’ssomethingelse,we’llrestaterevenue.”

E&Y:“No.That’snotgoodenough.You’llneedtogetallthreecustomerstoamendtheircontractstousenewlanguagethatwe’vecomeupwithatE&Ytoclearthingsup.”Ben: “But all three customers are large banks. They have risk departments.

Theywon’tjustamendtheircontractsquickly.Tomakethingsworse,weareinthemiddleofa$1.6billiontransaction.Youwillwreckourdeal.”E&Y:“Wedon’tcare.That’swhatyouhavetodo.”Ben:“Butwe’vebeenacustomerforeightyears,paidyoumillionsofdollars

in fees, and your partner made this judgment to begin with.Whywould youdestroyourtransactionifweandthecustomersverballyagreewiththecurrentinterpretation?”E&Y: “Either amend the contracts or restate your revenue.Youhave forty-

eighthours.”DavePricelookedlikehewasgoingtoburstintotears.E&Y’snationalofficedidnotcareaboutthespiritofthelaw,onlytheletter.

They refused todowhatwas right fromanaccountingorbusiness standpoint.TheyweredeterminedtodowhatwasoptimallyconvenientforErnst&Young.MyCFO,DaveConte,waspaleasaghost.Hundredsofpeoplehadworked

for eight years to get to this point and it looked like all that work would beflushed down the toilet by the accounting firm thatDave handpicked.He hadworked at Ernst&Young for fifteen years before joiningOpsware.Normallygregariousandoutgoing,hecouldbarelyspeak.IwasfuriousateverybodybutIknew thatnothing that Icouldsaywouldhelp theproblemormakeDave feelanyworsethanhealreadyfelt.IturnedtoJordanBreslow,mygeneralcounsel,and said, “Do we have to disclose this to the acquirers right away?”Horrifyingly,hesaid,“Yes.”We toldHPandBMCabout thediscrepancy and thatwe thoughtwe could

resolve the issue by amending the contracts in the next twenty-four hours.Neitherpartybelievedus.Idon’tknowifIbelievedit.Howcouldwepossiblyget three large banks to amend their contracts in twenty-four hours? Bothpotentialbuyersmadeplanstorespondtothenewsandupdatetheirpositionandpossiblytheirbids.Inthemeantime,Dave,MarkCranney,andIwenttoworktogetthecontracts

amended. We sat in the finance conference room drawing lines betweeneverybodyweknewintheworldandtriedtofigureouthowtogettotherightpeopleintimetosavethedeal.Icalledeveryboardmembertoseewhichbankhekepthismoneyinandwhetherhehadanyinfluenceorknewsomebodywith

influence.Markstayedonthephonewiththesalesrepsandthepeoplewhomweknewattheaccounts.JordanandDavecameupwithtendifferentwaystowordtheamendment.Westayedupallnightworkingonthedealandtheentiretime,Davelookedasthoughhewashavingamassiveheartattack.Miraculously,by11a.m.,wegotallthreecontractsamended,anditwasinlessthantwenty-fourhours.Wewouldnothavetorestaterevenue.Notsurprisingly,theentirethingspookedBMC,andtheywithdrewtheirbid.

Theydidnotbelievethattheissuehadbeenputtobed.HPwasn’tspooked,buttheyloweredtheirofferto$13.75pershareduetothe“taint”onthedeal.ThatnighttheboardmetinourofficestodiscussthenewofferfromHPand

informthemthatBMChaddroppedout.Theboardwasunanimousinitsadvicethatweshouldtakethe$13.75offer.Idisagreed.IsaidthatIwouldnottakeanickellessthanthe$14.25originallyoffered.BillCampbelllookedatmelikeIwasageneralwhohadspenttoomuchtimeinthefield.Atthatpoint,Ihadnotsleptanddidn’tknowifhewasrightorwrong.IjustknewthatIhadwaitedallnighttoberightandIcouldn’tgetmuchwronger.Igatheredmyselfandrestatedmyposition:“HPoffered$14.25,aprice that

equals sixteen times trailing twelve months revenue for one reason and onereason only. They offered it because we are the premium company, the goldstandard if youwill in an importantmarket. That is the entire premise of thisdeal.Thesecondthatweacceptadiscountedofferorinanyotherwaysuggestthatwearenotthegoldstandard,thisdealfallsapart.”JohnO’Farrellnoddedinagreement.Theboarduneasilyacceptedmyposition.IwentbacktoHPandtoldthemthedealwouldbe$14.25ornothing.After

about a two-hour pause, during which time the color never returned to DaveConte’sface,theyaccepted.Wehadadeal.Adealforabout$100millionlessthan if our so-called partner had not stabbed us in the back, but a dealnonetheless.IstillhateErnst&Young.I am telling this story todaybecause justwhenyou think thereare thingsyoucan count on in business, you quickly find that the sky is purple.When thishappens, itusuallydoesnogoodtokeeparguingthat theskyisblue.Youjusthave to get on anddealwith the fact that it’s going to look likeBarney for awhile.

SOLVINGTHEACCOUNTABILITYVS.CREATIVITYPARADOX

Asoftware engineer identifies aweakness inyour currentproduct architecturethatwillsignificantly impair itsability toscaledowntheroad.Shefiguresoutthat she’ll have to slip theproduct schedule threemonths to fix it.Everybodyagrees that three months is an acceptable slip to correct the problem. Thescheduleactuallyslipsninemonths,butshewasrightabouttheproblem.Doyourewardherforhercreativityandcourageorholdheraccountablefortheslip?Ifyoubecomeaprosecutingattorneyandholdhertotheletterofthelawon

hercommitments,youwillalmostcertainlydiscourageherandeverybodyelsefrom taking important risks in the future. If you take this stance consistently,don’t be surprised in the future if your people don’t have time to solve hardproblemsbecausetheywillbefartoobusycoveringtheirbutts.Ontheotherhand,ifyoudon’tholdheraccountableforhercommitment,then

thepeoplewhoactuallydotheworktomeettheircommitmentsmightfeellikeidiots.WhydidIstayupallnightmakingthedeadlineiftheCEOrewardsthepersonwhomissedherschedulebysixmonths?Ifyourhardestworking,mostproductiveemployeesfeellikechumpsandyouarelookingfortheculprit,lookin the mirror. You have failed to hold people accountable for their actions.WelcometotheAccountabilityvs.CreativityParadox.As we look to solve it, let’s start with themost basic assumption. Do you

assume that your employees are by and large creative, intelligent, andmotivated?Or do you assume that they are lazy, conniving, and counting theminutes toquitting time? Ifyoubelieve the latter, thenyoumight aswell justgiveuponcreativityandinnovationinyourorganization,becauseyouwillnotget it. It’s better to believe the former and assume that people have goodintentionsunlesstheyproveotherwise.Still,youmustholdpeopleaccountabletoavoidthechumpfactor.Howdoyouthinkaboutthat?Let’s look at accountability across the following dimensions: promises,

results,andeffort.

ACCOUNTABILITYFOREFFORT

Thisistheeasyone.Tobeaworld-classcompany,youneedworld-classeffort.Ifsomebodyisn’tgivingittoyou,theymustbechecked.

ACCOUNTABILITYFORPROMISES

Alotofgood-runningorganizationshavestatementslike“makeacommitment,keepacommitment.”It’struethatifyousignupforsomethingandyoudon’tdoit, you let everyone in the organization down. This type of letdown can becontagious.Holdingpeopleaccountablefortheirpromisesisacriticalfactoringetting things done. This changes as the degree of difficulty in fulfilling thepromiseincreases.Promisingtocompleteapieceofmarketingcollateralorsendan email is different from promising to meet an engineering schedule thatinvolvessolvingsomefundamentallyhardcomputerscienceproblem.Youmusthold people accountable for the former; the latter is more complicated andrelatestoresults.

ACCOUNTABILITYFORRESULTS

This iswhere thingsgetcomplicated. Ifsomeonefails todeliver theresultshepromised,as in theopeningstory,mustyouholdheraccountable?Shouldyouholdheraccountable?Theansweristhatitdepends.Itdependsupon:

SeniorityoftheemployeeYoushouldexpectexperiencedpeopletobeabletoforecasttheirresultsmoreaccuratelythanjuniorpeople.DegreeofdifficultySomethingsarejustplainhard.Makingyoursalesnumberwhenyourproductisinferiortothecompetitionandarecessionhitsmidquarterishard.Buildingaplatformthatautomaticallyandefficientlytakesserialprogramsandparallelizesthem,sothattheycanscaleout,ishard.It’shardtomakeagoodpredictionandhardtomeetthatprediction.Whendecidingtheconsequenceofmissingaresult,youmusttakeintoaccountthedegreeofdifficulty.AmountofstupidriskWhileyoudon’twanttopunishpeoplefortakinggoodrisks,notallrisksaregood.Whilethereisnorewardwithoutrisk,thereiscertainlyriskwithlittleornochanceofcorrespondingreward.DrinkingabottleofJackDaniel’sthengettingbehindthewheelofacaris

plentyrisky,butthere’snotmuchrewardifyousucceed.Ifsomeonemissedaresult,didshetakeobviouslystupidrisksthatshejustneglectedtoconsider,orweretheyexcellentrisksthatjustdidnotpanout?

REVISITINGTHEOPENINGPROBLEM

Solookingbackattheopeningproblem,herearesomethingstoconsider:

1.Howseniorisshe?Ifshe’syourchiefarchitect,you’llneedhertogetbetteratscopingherworkorshe’sgoingtotrashtheorganization.Ifsheismorejunior,thisshouldbemoreateachingmomentthanascoldingmoment.2.Howhardwasit?Ifitwasamiraclethatyouevermadethatpieceofcrapscale,thenyoushouldn’tyellather.Infact,youshouldthankher.Ifitwasarelativelytrivialprojectthatjusttooktoolong,thenyouneedtoaddressthat.3.Wastheoriginalrisktherightonetotake?Wouldtheproductreallyhaverunoutofscaleintheshort-to-mediumterm?Iftheanswerisyes,thenwhetherittookthreemonthsorninemonths,itwastherightrisktotakeandiffacedwiththesamesituationagain,youprobablyshouldnotchangeanyofyouractions.Youshouldn’tbewringingyourhandsaboutthat.

FINALPOINT

Inthetechnologybusiness,yourarelyknoweverythingupfront.Thedifferencebetween being mediocre and magical is often the difference between lettingpeople take creative risk and holding them too tightly accountable.Accountabilityisimportant,butit’snottheonlythingthat’simportant.

THEFREAKYFRIDAYMANAGEMENTTECHNIQUE

Manyyearsago,Iencounteredaparticularlytrickymanagementsituation.Twoexcellentteamsinthecompany,CustomerSupportandSalesEngineering,wentto war with each other. The sales engineers escalated a series of blisteringcomplaints arguing that the Customer Support team did not respond withurgency, refused to fix issues in theproduct, andgenerally inhibited sales andcustomersatisfaction.Meanwhile,theCustomerSupportgroupclaimedthatthesales engineers submitted bugs without qualification, did not listen to validsuggested fixes, andwere alarmistswho assigned every issue the top priority.Beyond theactual complaints, the teamsgenuinelydidnot likeeachother.Tomakemattersworse, thesegroupshad towork togetherconstantly inorder forthecompanytofunction.Bothteamsboastedsuperbpersonnelandoutstandingmanagers,sotherewasnobodytofireordemote.Icouldnotfigureoutwhattodo.Aroundthistime,Imiraculouslyhappenedtowatchthemotionpictureclassic

Freaky Friday, starring the underrated Barbara Harris and the incomparableJodieFoster.(Thereisalsoahigh-qualityremakestarringJamieLeeCurtisandthetroubledbuttalentedLindsayLohan.)Inthefilm,motheranddaughtergrowcompletelyfrustratedwitheachother’slackofunderstandingandwishthattheycouldswitchplacesand,throughthemagicoffilm,theydo.Through the course of themovie, by being inside each other’s bodies, both

charactersdevelopanunderstandingofthechallengesthattheotherfaces.Asaresult,thetwobecomegreatfriendswhentheyswitchback.Afterwatchingboththeoriginalandtheremake,IknewthatIhadfoundtheanswer:IwouldemployaFreakyFridaymanagementtechnique.TheverynextdayIinformedtheheadofSalesEngineeringandtheheadof

CustomerSupportthattheywouldbeswitchingjobs.Iexplainedthat,likeJodieFoster and Barbara Harris, they would keep their minds, but get new bodies.Permanently.Their initial reactionswerenotunlike the remakewhereLindsayLohanandJamieLeeCurtisbothscreaminhorror.

However, after just one week walking in the other’s moccasins, bothexecutives quickly diagnosed the core issues causing the conflict. They thenswiftlyactedtoimplementasimplesetofprocessesthatclearedupthecombatandgottheteamsworkingharmoniously.Fromthatdaytothedaywesoldthecompany, the Sales Engineering and Customer Support organizations workedbetter together than any other major groups in the company—all thanks toFreakyFriday,perhapsthemostinsightfulmanagementtrainingfilmevermade.

STAYINGGREAT

As CEO, you know that you cannot build a world-class company unless youmaintain aworld-class team.But how do you know if an executive isworld-class? Beyond that, if she was world-class when you hired her, will she stayworld-class?Ifshedoesn’t,willshebecomeworld-classagain?These are complex questions and are made more complex by the courting

process.EveryCEOsetsouttohiretheverybestpersonintheworldandthenrecruitsaggressivelytogethim.Ifhesaysyes,sheinevitablythinksshe’shitthejackpot.IfIhadatattooforeverytimeIheardaCEOclaimthatshe’djusthired“thebestVPintheindustry,”I’dbeLilWayne.Sowe beginwith a strong bias thatwhoeverwe hiredmust beworld-class

even before performing one day of work. Tomakematters worse, executiveswhostartoffworld-classoftendeteriorateovertime.Ifyouareasportsfan,youknowthatworld-classathletesdon’tstayworld-classforlong.OnedayyouareTerrellOwensandthenextdayyouareTerrellOwens.Whileexecutivesdon’tagenearlyasfastasathletesdo,companies,markets,andtechnologieschangeathousandtimesfasterthanthegameoffootball.Asaresult,theexecutivewhoisspectacular in this year’s hundred-person startup may be washed-up in nextyear’s versionwhen the company employs four hundred people and has $100millioninrevenue.

THESTANDARD

Thefirstthingtounderstandisthatjustbecausesomebodyinterviewedwellandreference-checkedgreat, thatdoesnotmeanshewillperformsuperbly inyourcompany.Therearetwokindsofculturesinthisworld:cultureswherewhatyoudomattersandcultureswhereall thatmatters iswhoyouare.Youcanbe theformeroryoucansuck.Youmust hold your people to a high standard, butwhat is that standard? I

discussed this in the section “Old People.” In addition, keep the following inmind:

Youdidnotknoweverythingwhenyouhiredher.Whileitfeelsawkward,itisperfectlyreasonabletochangeandraiseyourstandardsasyoulearnmoreaboutwhat’sneededandwhat’scompetitiveinyourindustry.Youmustgetleverage.Earlyon,it’snaturaltospendagreatdealoftimeintegratingandorientinganexecutive.However,ifyoufindyourselfasbusyasyouwerewiththatfunctionbeforeyouhiredorpromotedtheexecutive,thensheisbelowstandard.AsCEO,youcandoverylittleemployeedevelopment.OneofthemostdepressinglessonsofmycareerwhenIbecameCEOwasthatIcouldnotdevelopthepeoplewhoreportedtome.Thedemandsofthejobmadeitsuchthatthepeoplewhoreportedtomehadtobe99percentreadytoperform.UnlikewhenIranafunctionorwasageneralmanager,therewasnotimetodeveloprawtalent.Thatcanandmustbedoneelsewhereinthecompany,butnotattheexecutivelevel.Ifsomeoneneedslotsoftraining,sheisbelowstandard.

Itispossibletotakethestandardsettingtoofar.AsIdiscussedinthesection“The Scale Anticipation Fallacy,” it’s neither necessary nor a good idea toevaluateanexecutivebasedonwhather jobwillbe twoyears fromnow.Youcancross thatbridgewhenyoucome to it.Evaluateheronhowsheperformsrighthereandrightnow.

ONEXPECTATIONSANDLOYALTY

Ifyouhaveagreatandloyalexecutive,howdoyoucommunicateallthis?Howdoyoutellherthatdespitethemassiveeffortandgreatjobsheisdoingtoday,you might fire her next year if she doesn’t keep up with the changes in thebusiness?WhenIusedtoreviewexecutives,Iwouldtellthem,“Youaredoingagreat

job at your current job, but the plan says that we will have twice as manyemployeesnextyearaswehaverightnow.Therefore,youwillhaveanewandverydifferentjobandIwillhavetoreevaluateyouonthebasisofthatjob.Ifitmakesyoufeelbetter,thatrulegoesforeveryoneontheteam,includingme.”Inprovidingthiskindofdirection,it’simportanttopointouttotheexecutive

thatwhenthecompanydoublesinsize,shehasanewjob.Thismeansthatdoingthings thatmade her successful in her old jobwill not necessarily translate tosuccess in the new job. In fact, the number-oneway that executives fail is by

continuingtodotheiroldjobratherthanmovingontotheirnewjob.But, what about being loyal to the team that got you here? If your current

executive team helped you grow your company tenfold, how can you dismissthemwhentheyfallbehindinrunningthebehemoththeycreated?Theansweristhat your loyaltymust go to your employees—the people who report to yourexecutives.Yourengineers,marketingpeople,salespeople,andfinanceandHRpeoplewhoaredoingthework.Youowethemaworld-classmanagementteam.That’sthepriority.

SHOULDYOUSELLYOURCOMPANY?

OneofthemostdifficultdecisionsthataCEOevermakesiswhethertosellhercompany.Logically,determiningwhethersellingacompanywillbebetterinthelong term than continuing to run it stand-alone involves a huge number offactors,mostofwhicharespeculativeorunknown.Andifyouarethefounder,thelogicalpartistheeasypart.Thetaskwouldbefarsimpleriftherewerenoemotioninvolved.Butselling

yourcompanyisalwaysemotionalanddeeplypersonal.

TYPESOFACQUISITIONS

For the purpose of this discussion, it is useful to think about technologyacquisitionsinthreecategories:

1.Talentand/ortechnology,whenacompanyisacquiredpurelyforitstechnologyand/oritspeople.Thesekindsofdealstypicallyrangebetween$5millionand$50million.2.Product,whenacompanyisacquiredforitsproduct,butnotitsbusiness.Theacquirerplanstoselltheproductroughlyasitis,butwilldosoprimarilywithitsownsalesandmarketingcapability.Thesekindsofdealstypicallyrangebetween$25millionand$250million.3.Business,whenacompanyisacquiredforitsactualbusiness(revenueandearnings).Theacquirervaluestheentireoperation(product,sales,andmarketing),notjustthepeople,technology,orproducts.Thesedealsaretypicallyvalued(atleastinpart)bytheirfinancialmetricsandcanbeextremelylarge(suchasMicrosoft’s$30billion–plusofferforYahoo).

Mytakeonthesubjectismostapplicabletobusinessacquisitions,withsomerelevance to product acquisitions, andwill be fairly useless if you are sellingpeopleand/ortechnology.

THELOGICAL

When analyzingwhether you should sell your company, a good basic rule ofthumbisif(a)youareveryearlyoninaverylargemarketand(b)youhaveagoodchanceofbeingnumberoneinthatmarket,thenyoushouldremainstand-alone. The reason is that nobody will be able to afford to pay what you areworth,becausenobodycangiveyou thatmuchforwardcredit.Foraneasy-to-understand example, consider Google. When they were very early, theyreportedly receivedmultiple acquisition offers formore than$1billion.Thesewere considered very rich offers at the time and they amounted to a giganticmultiple.However,giventhesizeoftheultimatemarket,itdidnotmakesenseforGoogletosell.Infact,itdidn’tmakesenseforGoogletoselltoanysuitoratanypricethatthebuyercouldhavepaid.Why?BecausethemarketthatGooglewaspursuingwasactuallybiggerthanthemarketsthatallofthepotentialbuyersownedandGooglehadbuiltanearlyinvincibleproductleadthatenabledthemtobenumberone.Contrast this situationwithPointcast.Pointcastwasoneof the first Internet

applications to catch fire. They were the buzz of Silicon Valley and thetechnology industry in general. They received billion-dollar acquisition offersthat they passed on. Then, due to flaws in their product architecture, theircustomersstartedtoturnofftheirapplication.Overnight,theirmarketcollapsedandneverreturned.Theywereultimatelysoldforarelativelytinyamount.So, the judgment that you have to make is (a) is this market really much

bigger(morethananorderofmagnitude)thanhasbeenexploitedtodate?and(b)arewegoingtobenumberone?Iftheanswertoeither(a)or(b)isno,thenyoushouldconsider selling. If theanswers tobothareyes, thensellingwouldmeansellingyourselfandyouremployeesshort.Unfortunately,thesequestionsarenotassimpletoanswerasI’vemadethem

outtobe.Inordertogettheanswerright,youalsohavetoanswerthequestion“What is the market, really, and who are the competitors going to be?”WasGoogleinthesearchmarketortheportalmarket?Inretrospect,theywereinthesearchmarket, butmost people thought theywere in the portalmarket at thetime.Yahoowasatoughcompetitorintheportalmarket,butnotsomuchinthesearchmarket.IfGooglehadreallybeenintheportalmarket,thensellingmighthavebeenagoodidea.Pointcastthoughtthattheirmarketwasmuchlargerthanit turned out to be. Interestingly, Pointcast’s own product execution (or lackthereof)causedtheirmarkettoshrink.Let’s look at the case ofOpsware.Why did I sellOpsware?Another good

questioniswhydidn’tIsellOpswareuntilIdid?

AtOpsware,westarted in the serverautomationmarket.Whenwe receivedourfirst inquiriesandoffersfor theserverautomationcompany,wehadfewerthan fifty customers. I believed that there were at least ten thousand targetcustomers and that we had a decent shot at being number one. In addition,althoughIknewthemarketwouldberedefined,Ithoughtthatwecouldexpandtonetworksandstorage(datacenterautomation)fasterthanthecompetitionandwinthatmarketaswell.Therefore,assuming30percentmarketshare,somebodywouldhavehadtopaysixtytimeswhatwewereworthinforwardcredittobuyoutourpotential.Youwon’tbesurprisedtofindthatnobodywaswillingtopaythat.Oncewe grew to several hundred customers and expanded into data center

automation,wewerestillnumberoneandweremorevaluablestand-alonethanany of the prior acquisition offers. At that point both Opsware and our maincompetitor,BladeLogic,haddevelopedintofull-fledgedcompanies(worldwidesalesforces,built-outprofessionalservices,etc.).Thiswassignificant,becauseitmeant that a large company could buy one of us and potentially executesuccessfully (big enterprise companies can’t generally succeed with smallacquisitions,becausetoomuchoftheimportantintellectualpropertyisthesalesmethodology,andbigcompaniescan’tbuildthat).Atthispoint, itbecameclearthatBMCwasgoingtobuyeitherOpswareor

BladeLogic. As a result, the calculus, or whether Opsware was going to benumberoneinthemarket,neededtoberedefinedasfollows:

1.Wehadtobenumberoneinthesystemsandnetworkmanagementmarketratherthanthedatacenterautomationmarket,becauselikethewordprocessormarket,thedatacenterautomationmarketwasgoingtobesubsumedbyalargermarketthatcontainedit.2.Inordertobenumberone,wehadtobeatBMCandBladeLogictogether,whichwasasignificantlymoredifficultopponentthaneithercompanystand-alone.

Finally,themarketitselfwastransformingduetoanunderlyingtechnologicalshift: virtualization. Virtualization meant that the entire market needed to beretooled, so we were embarking on a new R&D race to build the bestmanagement for virtualized environments.Thismeant deferring earnings for averylongtime.Based on all these factors, it made sense for us to at least consider the

possibilityofacquisitionandrunashortprocesstounderstandtheinterestintheM&Amarket.Through that process, eleven companies made acquisition offers of some

form.ThistoldmethatwewereatalocalmaximaintermsofthemarketpriceforOpsware.Inotherwords,thesetofpotentialbuyerswasconvincedthatthemarketwasveryimportant,andtherewasnoextrapremiumthatwecouldhopetoachieve throughbetterawareness. In theend,basedona lotofanalysisandsoul-searching, Idetermined that thecurrent localmaximawashigher thanwecouldexpecttoachieveinthenextthreetofiveyearsandIsoldthecompanytoHewlett-Packardfor$1.65billion.Ithinkandhopethatwastherightdecision.

THEEMOTIONAL

The funny thing about the emotional part of the decision is that it’s soschizophrenic.Howcanyoueversellyourcompanyafteryou’vepersonallyrecruitedevery

employee and sold them on your spectacular vision of a thriving, stand-alonebusiness?Howcanyoueverselloutyourdream?Howcanyouwalkaway from total financial independence foryourself and

everymemberofyourcloseanddistantfamily?Aren’tyouinbusinesstomakemoney?Howmuchmoneydoesonepersonneed?HowcanyoureconcileDr.Stay-the-CourseandMr.Sell-the-Thing?Clearly

theyareirreconcilable,butthekeyistomutethemboth.Afewkeysonmutingtheemotions:

Getpaid(asalary).Mostventurecapitalistslikeentrepreneursthatare“allin,”meaningtheentrepreneurhaseverythinginvestedinthecompanyandwillhaveverylittletoshowforhereffortsifitdoesnotsucceed.Aspartofthis,theypreferthefoundingCEOtohaveaverylowsalary.Ingeneral,thisisagoodidea,becausethetemptationtowalkawaywhenthingsgopoorlyisintenseandtotalfinancialcommitmenthelpshimtokeephisothercommitments.However,oncethecompanystartstobecomeacompanyratherthananideaitmakessensetopaytheCEOatmarket.Morespecifically,oncethecompanyhasabusiness(asdefinedabove)andbecomesanattractiveacquisitiontarget,itmakessensetopaytheCEO,sothatthedecisiontokeeporsellthecompanyisn’tadirectresponsetotheCEO’spersonalfinancialsituation,asin“Idon’tthinkthatweshouldsellthecompany,butIliveinaneight-hundred-fifty-square-footapartment

withmyhusbandandtwokidsandit’sthatordivorce.”Beclearwiththecompany.OnequestionthateverystartupCEOgetsfromheremployeesis“Areyousellingthecompany?”Thisisanincrediblydifficultquestion.Ifshesaysnothing,theemployeewilllikelyinterpretthistomeanthecompanyisforsale.Ifshesays“attherightprice,”theemployeewillwonderwhatthatpriceisandmayevenask.Ifthecompanyeverreachesthatprice,theemployeewillassumethecompanywillbesold.Ifshedodgesthequestionwiththestandard“thecompanyisnotforsale,”theemployeemayfeelbetrayedifthecompanyiseversold.Moreimportant,theCEOmayfeellikesheisbetrayingtheemployeeandthatfeelingwillinfluenceherdecision-makingprocess.Onewaytoavoidthesetrapsistodescribetheanalysisinthepriorsection:Ifthecompanyachievesproduct-marketfitinaverylargemarketandhasanexcellentchancetobenumberone,thenthecompanywilllikelyremainindependent.Ifnot,itwilllikelybesold.Thisisonegoodmethodtodescribetheinterestsoftheinvestorsinawaythat’snotatoddswiththeinterestsoftheemployees,anditistrue.

FINALTHOUGHT

Whenfacedwiththedecisionofwhethertosellyourcompany,thereisnoeasyanswer.However,preparingyourselfintellectuallyandemotionallywillhelp.

—CHAPTER9—

THEENDOFTHEBEGINNING

“Wewalkthesamepath,butgotondifferentshoesLiveinthesamebuilding,butwegotdifferentviews.”

—DRAKE,“RIGHTABOVEIT”

AftersellingOpsware,IspentayearatHPrunningthebulkof theirsoftwarebusiness.And thenI tried to figureoutwhat todonext.ShouldIstartanothercompany? Should I be CEO of someone else’s company? Should I retire?ShouldIdosomethingcompletelydifferent?ThemoreIthoughtaboutmyfuture,themoreIthoughtaboutmypast.What

would have happened if I’d never met Bill Campbell? How would I havepossibly worked my way through all the challenges I’d faced? Why wasentrepreneurship suchablack art?Did everybodyhave the sameproblems I’dhad? If they did, why didn’t anybodywrite anything down?Why did so fewstartupadvisersandventurecapitalistshaveanyexperiencestartingcompanies?Asthesethoughtsrolledaroundinmyhead,IsentMarcAndreessenaninstant

message: “We ought to start a venture capital firm. Our motto for generalpartnerswouldbe‘someexperiencerequired’asinsomeexperienceinfoundingand running companies is required to advise people who are founding andrunning companies.” To my surprise, he replied, “I was thinking the samething.”

SOMEEXPERIENCENECESSARYFurthercontemplationtookmebacktooneofmyfirstseriousencounterswithventurecapital.Back in 1999, after raising our first round of funding for Loudcloud, my

cofoundersandIwent tovisitournewventurecapital firmandmeet their full

team.AsfoundingCEO,Irememberbeingexcitedtomeetourfinancialbackersand to talk about how we could partner to build a great company. Theconversationtookasharpdownhillturnwhenoneoftheseniorpartners,DavidBeirne,askedme,infrontofmycofounders,“WhenareyougoingtogetarealCEO?”Thecommentknockedthewindoutofme.Ourlargestinvestorhadbasically

calledme a fakeCEO in front ofmy team. I asked, “What do youmean?”—hoping hewould revise his statement and enableme to save face. Instead hepressed on: “Someone who has designed a large organization, someone whoknows great senior executives and brings prebuilt customer relationships,someonewhoknowswhattheyaredoing.”Icouldhardlybreathe.Itwasbadenoughthathehadunderminedmystanding

asCEO,but tomakemattersworse, I knew that at some levelhewas right. Ididn’thavethoseskills.Ihadneverdonethosethings.AndIdidnotknowthosepeople.IwasthefoundingCEO,notaprofessionalCEO.Icouldalmostheartheclock ticking in the background asmy time running the company quickly ranout.CouldIlearnthejobandbuildmynetworkfastenoughorwouldIlosethecompany?Thatquestiontorturedmeformonths.In the years that followed, I remained CEO, for better or worse. I worked

incrediblyhardtobecomethekindofCEOthateveryoneexpected.Thankstoalot of effort and help from friends andmentors, especiallyBill Campbell, thecompanysurvivedandbecamesuccessfulandvaluable.Notadaywentby,however,whenIdidn’t thinkabout that interactionwith

DavidBeirne. IalwayswonderedhowlongIhad togrowupandhowIcouldfindhelptobuildmyskillsandmakethenecessaryconnectionsalongtheway.MarcandIdiscussedthisparadoxoften.Wewonderedaloudwhyasfounders

wehadtoprovetoourinvestorsbeyondashadowofadoubtthatwecouldrunthecompany,ratherthanourinvestorsassumingthatwewouldrunthecompanywe’d created. This conversation ultimately became the inspiration forAndreessenHorowitz.Togetstarted,westudiedtheventurecapital industryandwecameacrossa

potential problem with our approach. Historically, all the returns in venturecapital had been concentrated in a tiny number of firms and consistently by asmallnumberofthesamefirms.Ofthemorethaneighthundredventurecapitalfirmsoftheday,onlyaboutsixhaddeliveredgreatreturnsfortheirinvestors.Aswe dug deeper, we uncovered an excellent reason for this: The bestentrepreneurswillonlyworkwith thebestventurecapital firms.Sinceventure

capital firmswerenotoriouslysecretiveabout theirmethodsandbeliefs—mostfirms did almost no PR and stated very little about what they did—the firmscompeted on their investing track records. Therefore, the firms with the besttrack records continued to have the best track records, thus making it nearlyimpossibleforanewfirm,withnotrackrecord,tocrackintothetoptier.We needed some way to break through to become the firm that great

entrepreneurswantedtoworkwith.Buthow?We needed to change the rules bywhich entrepreneurs evaluatedVCs.We

thoughttherewasanopeningtodothis,becausetimeshadchanged.WhenMarcandIfirstbecameentrepreneursbackinthemid-1990s,wedidnotknowmanyotherentrepreneurs.Wejustdidwhatwedid,withoutreallyseeingourselvesaspart of a larger “movement” or a community. We were entrepreneurs at thebeginning of the Internet and before Facebook, Twitter, and the other socialnetworkingplatformswerebuilt.Wedidnottalktootherentrepreneurs,becausetherewas no entrepreneurial community.Wewere completely heads downonthebusiness.All thathaschanged in the last tenyears.Entrepreneursarenowsocializing, friending eachother,meetingup, andhangingout.There is a realcommunity.Oncewerealized this,wefigured that ifwehadabetteroffering,word-of-mouthmarketingwouldworknowwhereithadn’tbefore.We needed to be better, butwe also needed to be different.Aswe thought

about what would make us both better and different, two core ideas greatlyinfluenced our thinking: First, technical founders are the best people to runtechnology companies. All of the long-lasting technology companies that weadmired—Hewlett-Packard, Intel, Amazon, Apple, Google, Facebook—hadbeen run by their founders. More specifically, the innovator was running thecompany. Second, itwas incredibly difficult for technical founders to learn tobecomeCEOswhile building their companies. Iwas a testament to that. But,most venture capital firmswere better designed to replace the founder than tohelpthefoundergrowandsucceed.Marc and I thought that if we created a firm specifically designed to help

technicalfoundersruntheirowncompanies,wecoulddevelopareputationandabrandthatmightvaultusintothetoptierofventurecapitalfirmsdespitehavingno track record.We identified twokey deficits that a founderCEOhadwhencomparedwithaprofessionalCEO:

1.TheCEOskillsetManagingexecutives,organizationaldesign,runningsalesorganizationsandthelikewereallimportantskillsthattechnical

founderslacked.2.TheCEOnetworkProfessionalCEOsknewlotsofexecutives,potentialcustomersandpartners,peopleinthepress,investors,andotherimportantbusinessconnections.Technicalfounders,ontheotherhand,knewsomegoodengineersandhowtoprogram.

Next,weasked,“HowmightaventurecapitalfirmhelpfounderCEOsclosethosegaps?”Addressing the skill set issueproved tobedifficultbecause, sadly, theonly

waytolearnhowtobeaCEOistobeaCEO.Sure,wemighttrytoteachsomeskills, but learning to be a CEO through classroom training would be likelearning tobeanNFLquarterback throughclassroom training.Even ifPeytonManning and Tom Brady were your instructors, in the absence of hands-onexperience,you’dgetkilledthemomentyoutookthefield.Wedecided thatwhilewewouldnotbeable togivea founderCEOall the

skillssheneeded,wewouldbeabletoprovidethekindofmentorshipthatwouldaccelerate the learningprocess.As a result,wedecided that all of our generalpartnerswouldneedtobeeffectivementorsforafounderstrivingtobeaCEO.(Of course, not all founders want to be CEO. For some companies, the rightthingistobringinaprofessionalCEO.Forthosecompanies,wewouldfocusonhelping the founders identify the right CEO, and then helping the CEOsuccessfully integrate into thecompanyandpartnerwith thefounders to retaintheiruniquestrengths.)ThisiswhysomanyofthegeneralpartnerswechooseareformerfoundersorCEOsorboth,andtheyareallhighlyfocusedonhelpingfounders become outstanding CEOs. The idea seemed so simple and obviousthatithadtowork.Next,wedecidedtosystematizeandprofessionalizethenetwork.Forthiswe

drewboth the inspiration and the formula frommy friend andOpsware boardmemberMichaelOvitz.Thirty-fouryearsearlier,MichaelhadfoundedCreativeArtists Agency (CAA), the powerhouse of Hollywood talent agencies. WhenMichaelstartedCAA,itwasnotanobviousidea.Thetalentagencybusinesshadexistedsincevaudevilleandhadchangedverylittleintheensuingseventy-fiveyears. Michael was a rising star at the William Morris Agency, the mostimportant agency in the industry at the time.Quitting that job to pursuewhatmusthave looked likeawindmill tiltmadeno sense to anybody.ButMichaelhadaclearvision: Ifhecouldbuilda firmsogood that itattractedall the toptalent in theworld, thenhewouldshift theeconomicsof theindustryfromthe

corporationstothetalent,wherehefeltthatitbelonged.The firms of the daywere essentially collections of loosely affiliated talent

agents.Agentsworkedunder the sameumbrella, but acted largely alone, eachagenttappingintoherownnetworkonbehalfofherownclients.Forexample,AgentAmightintroduceDustinHoffmantotheheadofWarnerBros.,butboththe relationship with Dustin Hoffman and the relationship withWarner Bros.were controlled entirely by the agent; otherWilliamMorris agents and clientsdidn’t automatically get access to either. This traditionalmodel sounded a lotlikethetraditionalventurecapitalbusiness,whereVCsworkedinthesamefirmbutmanagedtheirownnetworksandtheirownportfolios.Ovitz’s breakthrough idea was to build an integrated network that would

allowanyofthefirm’sagentstoconnecttheirclientstoafirm-widegridofnewopportunities.Asaresult,thefirmwouldbeahundredtimesmorepowerfulthanany one agent at any other agency. To implement the idea, Ovitz and hisfoundingpartnersagreedtodefertheirsalariesforseveralyearsandinvesttheircommissions into building what Ovitz referred to as “The Franchise.” TheFranchise consisted of specialists running networks and portfolios in eachrelevant area: book publishing, international, music, and more. His theoryworked,andwithinfifteenyears,CAArepresented90percentofthetoptalentinHollywoodandhad rewritten the rules—giving talentmore say indeals and abiggerpieceofthefinancialpie.We decided to copy CAA’s operating model nearly exactly—in fact,

Andreessen Horowitz employees have the same titles as the original CAAemployees:partner.Michaelthoughtitwasagreatidea,buthewastheonlyone.Everyoneelseofferedsomevariationofthefollowing:“ThisisSiliconValley,notHollywood.Youguysdon’tunderstandthebusiness.”Still,withMichael’sendorsementandenthusiasticsupport,wepushedforwardwiththeidea.Asweappliedittoventurecapital,wedecidedtobuildthefollowingnetworks:

LargecompaniesEverynewcompanyneedstoeithersellsomethingtoorpartnerwithalargercompany.ExecutivesIfyousucceed,atsomepointyouneedtohireexecutives.EngineersInthetechnologybusiness,youcanneverknowenoughgreatengineers.PressandanalystsWehaveasayingaroundthefirm:Showit,sellit;hideit,keepit.

InvestorsandacquirersBeingventurecapitalists,providingaccesstomoneywasobvious.

Oncewedesignedthefirm,weneededtohelpentrepreneursunderstandhowweweredifferent.Thisseemedtricky,becausenomajorventurecapitalfirmdidanymarketingofanykind.Wefiguredtheremustbeagoodexplanationforthis,but struggled to find one. Finally, Marc discovered that the original venturecapital firms in the late 1940s and early ’50sweremodeled after the originalinvestmentbankssuchasJ.P.MorganandRothschild.ThosebanksalsodidnotdoPRforaveryspecificreason:Thebanksfundedwars—andsometimesbothsidesofthesamewar—sopublicitywasnotagoodidea.Thisinsight,combinedwithourgeneral instinct tocounterprogramwhatever thebigguysweredoing,ledustolaunchAndreessenHorowitzwithgreatfanfare.Whendecidingonthename,thebiggestproblemwefacedwasthat,asafirm,wewerenobodies.Notrackrecord,noportfoliocompanies,nonothing.Butpeopleknewusandtheyespecially knewMarc. So I said, “Rather than trying to create a totally newbrand from scratch, why not just use your brand?” Marc thought that madesense, but nobody would ever be able to spell “Andreessen Horowitz” whentyping in the URL. Thinking back to old-time computer programming in thedaysbeforeprogramming languages supported internationalization,weused tohave to“internationalize”ourcode.Wecalled this internationalizationprocess“I18N”forshort(localizationwasL10N),whichmeantI followedbyeighteenlettersfollowedbyN.Wedecidedthatthefirm’snicknamewouldbe“a16z,”asinafollowedbysixteenlettersfollowedbyz.WehiredtheOutcastmarketingagency,headedupbyitsformidablefounder,

MargitWennmachers, to generatemedia interest.We needed people to knowwhatwewereaboutaswehaddecidedtodefytheconventionalventurecapitaltheoryofnoPR.ThedaughterofaGermanpigfarmer,Margitwasthefurthestthing froma swinewrangler imaginable.Smart and sophisticated, shewas theBabeRuthofPR.Sheworkedhercontacts,landingacoverstoryinFortunein2009 that featuredMarc posing as Uncle Sam. Andreessen Horowitz was anovernight sensation, and yetMarc and Iwere still the only two people in thefirm.AftereightyearsofrunningLoudcloudandOpsware,Ihadlearnedsomany

hard lessons that building the teamwas easy. I understood the importance ofhiring for strength rather than for lack of weakness, and I understood themeaningof “fit.”Thereare lotsof smartpeople in theworld,but smart isnot

good enough. I needed people who were great where I needed greatness. Ineeded people who really wanted to do the jobs they were hired for. And Ineeded people who believed in the mission—tomake Silicon Valley a betterplacetobuildacompany.The first person we hired was Scott Kupor, who had been the director of

financefromOpsware.Scottworkedformefornearlytheentireeightyears,andI’m not sure that he enjoyed any of it, but he performed phenomenallynonetheless.He ran customer support, planning, and technical field operationsduringthoseyears,butnoneofthejobswerewhathewantedtodo.Scottlovedthreetasks:runningthings,strategy,anddeals.Ifhecoulddothosethings,he’dalmostneversleep.ButatOpsware,he’donlygottentodotwoofthethree.Notgetting to work on transactions was torture for Scott. He was like a cagedanimal.AndI’dkepthiminthatcageforeightlongyears.So,whenwedesignedthefirm,thefirstthingthatcametomymindwas“IfinallyfoundtheperfectjobforKupor.”Scottbecamethefirm’schiefoperatingofficer.Wethenfilledouttherestoftheteam.WehiredMarkCranney,headofsales

atOpsware,torunthelarge-companynetwork;ShannonCallahan,formerheadofrecruitingandHR,toruntheengineeringnetwork;MargitWennmachers,theSultan of Swat, to run themarketing network; Jeff Stump, the best executiverecruiterwe knew, to run the executive network; and FrankChen,my formerheadofproductmanagement,torunacentralizedresearchgroup.Our theory about what a venture capital firm should offer turned out to

resonatewith thebest entrepreneurs in theworld. In fourvery shortyears,wehavegonefromnothingtobeingoneofthemostrespectedventurecapitalfirmsintheworld.

FINALLESSON“Iknowyouthinkmylifeisgoodcausemydiamondpiece

ButmylifebeengoodsinceIstartedfindingpeace.”—NAS,“LOCO-MOTIVE”

IoftenjokethatIamconsideredtobeamuchbetterCEOnowthanIwaswhenIwasactuallyCEO.Thesedayspeoplesometimesrefer tomeasamanagementguru, butwhen ImanagedOpswaremost people referred tome as,well, lessthanthat.AsFeliciaisfondofsaying,“TheycalledyoueverythingbutachildofGod.”Whathappened?DidIchangeordidperceptionchange?

ThereisnoquestionthatIlearnedagreatdealovertheyearsandIamprettyembarrassedabouthowIconductedmyself in theearlydays,butby theend Ibecame pretty good at running the company. There is plenty of evidencesupportingthisview.Icompletelychangedourbusinessmidstream—evenwhileitwasapubliclytradedcompany—andstillmanagedtogrowitsvaluefrom$29millionto$1.65billioninfiveyears.AlargepercentageoftheemployeesfromOpswareeitherworkformeatAndreessenHorowitzorinoneofourportfoliocompanies, so they must have liked something about working together. TheacquisitionbyHPwasthelargestoutcomeinthesector,sowewonourmarket.Still, during the years that I was good at running Opsware—from 2003 to

2007—you would be hard-pressed to find a single article or blog post ormessageboardcommentthatsaidanythingniceaboutme.Duringthattime,thepressdeclaredthecompanydeadandshareholderscalledformyresignation.No,Iwasnotconsideredtobeverygoodatall.Inretrospect,people’sperceptionschangedbecauseofthesaletoHPandthe

things that I’ve since written. Once I stopped being CEO, I was granted afreedom that I did not have before. As a venture capitalist, I have had thefreedom to say what I want and what I really think without worrying whateverybody else thinks.As aCEO, there is no such luxury.AsCEO, I had toworry about what everybody else thought. In particular, I could not showweaknessinpublic.Itwouldnothavebeenfairtotheemployees,theexecutives,orthepubliccompanyshareholders.Unrelentingconfidencewasnecessary.WhenwestartedAndreessenHorowitz, Icould letall thatgo.Sure,westill

have employees, butwe do not have public shareholderswho live and die oneverypieceof press.More important, atAndreessenHorowitz I amnot reallyCEO. We invest in companies that have CEOs. The burden of unrelentingconfidence lieswith them. I cannowsharemyweaknesses,my fears, andmyshortcomings. I can say what I want without worrying about offending thewrong people in the power structure. And it’s those fears and controversialopinions that hold the clues to dealingwith hard things.Hard things are hardbecause there are no easy answers or recipes. They are hard because youremotionsareatoddswithyourlogic.Theyarehardbecauseyoudon’tknowtheanswerandyoucannotaskforhelpwithoutshowingweakness.When I first became a CEO, I genuinely thought that I was the only one

struggling. Whenever I spoke to other CEOs, they all seemed like they hadeverything under control. Their businesses were always going “fantastic” andtheirexperiencewasinevitably“amazing.”Ithoughtthatmaybegrowingupin

Berkeley with Communist grandparents might not have been the bestbackground for running a company. But as I watched my peers’ fantastic,amazingbusinessesgobankruptandsellforcheap,IrealizedthatIwasprobablynottheonlyonestruggling.As I got further into it, I realized that embracing the unusual parts of my

backgroundwouldbethekeytomakingitthrough.Itwouldbethosethingsthatwouldgivemeuniqueperspectivesandapproaches to thebusiness.The thingsthatIwouldbringtothetablethatnobodyelsehad.ItwasmyborrowingChicoMendoza’sshockingyetpoeticstyletomotivateandfocustheteam.ItwasmyunderstandingofthepeopleunderneaththepersonaandskincolorthatenabledmetoputJasonRosenthaltogetherwithAnthonyWrighttosavethecompany.Itwasevenmybringing in to themost capitalisticpursuit imaginablewhatKarlMarxgotright.Onmygrandfather’stombstone,youwillfindhisfavoriteMarxquote:“Lifeisstruggle.”Ibelievethatwithinthatquoteliesthemostimportantlessoninentrepreneurship:Embracethestruggle.When I work with entrepreneurs today, this is the main thing that I try to

convey.Embraceyourweirdness,yourbackground,yourinstinct.Ifthekeysarenotinthere,theydonotexist.Icanrelatetowhatthey’regoingthrough,butIcannot tell themwhat to do. I can only help them find it in themselves.AndsometimestheycanfindpeacewhereIcouldnot.Ofcourse,evenwithalltheadviceandhindsightintheworld,hardthingswill

continuetobehardthings.So,inclosing,Ijustsaypeacetoallthoseengagedinthestruggletofulfilltheirdreams.

APPENDIX

QUESTIONSFORHEADOFENTERPRISESALESFORCE

Isshesmartenough?

Cansheeffectivelypitchyouonhercurrentcompany?Howarticulateissheonthecompanyandmarketopportunitythatyouarepresentingtohernow?Willshebeabletocontributetothestrategicdirectionofyourcompanyinameaningfulway?

Doessheknowhowtohiresalespeople?

Whatisherprofile?Askhertodescribearecentbadhire.Howdoesshefindtoptalent?Whatpercentageofhertimeisspentrecruiting?Howdoesshetestforthecharacteristicsshewantswithherinterviewprocess?Howmanyofhercurrentpeoplewanttosignup?Canyoureferencethemandvalidatethat?Couldyoupasshersalesinterviewtest?Shouldyoubeabletopass?Doessheknowhowtohiresalesmanagers?Canshedefinethejob?Canshetestfortheskills?

Is she systematic and comprehensive on how she thinks about the salesprocess?

Doessheunderstandthebusinessandthetechnicalsalesprocesses?Doessheunderstandbenchmarking,lockoutdocuments,proofofconcepts,demos?Doessheknowhowtotrainpeopletobecomecompetentintheprocess?Cansheenforcetheprocess?Whatisherexpectationofherteam’suseoftheCRMtools?Didsheruntheprocessatherlastcompanyordidshewritetheprocess?

There is a big difference between people who can write a game plan andpeoplewhocanfollowagameplan.Howgoodishersalestrainingprogram?

Howmuchprocesstrainingversusproducttraining?Canshedescribeitindetail?Doesshehavematerials?Howeffectiveishersalesrepevaluationmodel?Canshegetbeyondbasicperformance?Canshedescribethedifferencebetweenatransactionalrepandanenterpriserepinawaythatteachesyousomething?

Doessheunderstandtheinsandoutsofsettingupacompplan?

Accelerators,spiffs,etc.Doessheknowhowtodobigdeals?

Hasshemadeexistingdealsmuchlarger?Willherpeoplebeabletodescribethat?Hassheacceleratedthecloseofalargedeal?Doesshehavecustomerswhowillreferencethis?Doessheunderstandmarketing?Canshearticulatethedifferencesbetweenbrandmarketing,leadgeneration,andsalesforceenablementwithoutprompting?

Doessheunderstandchannels?

Doesshereallyunderstandchannelconflictandincentives?Issheintenseenough?WilltherepinWisconsinwakeupat5a.m.andhitthephonesorwilltheywakeupatnoonandhavelunch?

Cansheruninternational?Isshetotallypluggedintotheindustry?Howquicklycanshediagnose?

Doessheknowyourcompetition?Doessheknowwhatdealsyouareinrightnow?Hasshemappedyourorganization?

OPERATIONALEXCELLENCEQUESTIONS

ManagingDirectReports

Whatdoyoulookforinthepeopleworkingforyou?Howdoyoufigurethatoutintheinterviewprocess?Howdoyoutrainthemforsuccess?Whatisyourprocessforevaluatingthem?

Decision-Making

Whatmethodsdoyouusetogettheinformationthatyouneedinordertomakedecisions?Howdoyoumakedecisions(whatistheprocess)?Howdoyourunyourstaffmeeting?Whatistheagenda?Howdoyoumanageactionsandpromises?Howdoyousystematicallygetyourknowledge?OftheorganizationOfthecustomersOfthemarket

Core management processes—please describe how you’ve designed theseandwhy.

InterviewPerformancemanagementEmployeeintegrationStrategicplanning

MetricDesign

Describethekeyleadingandlaggingindicatorsforyourorganization.Aretheyappropriatelypaired?Forexample,doyouvaluetime,butnotquality?Aretherepotentiallynegativesideeffects?Whatwastheprocessthatyouusedtodesignthem?

OrganizationalDesign

Describeyourcurrentorganizationaldesign.Whatarethestrengthsandweaknesses?Why?Whydidyouoptforthosestrengthsandweaknesses(whywerethestrengthsmoreimportant)?Whataretheconflicts?Howdotheygetresolved?

Confrontation

Ifyourbestexecutiveasksyouformoreterritory,howdoyouhandleit?Describeyourprocessforbothpromotionandfiring.Howdoyoudealwithchronicbadbehaviorfromatopperformer?

LessTangible

Doesshethinksystematicallyorone-off?WouldIwanttoworkforher?Isshetotallyhonestorisshebullshitty?Doessheaskmespontaneousincisivequestionsoronlypre-preparedones?Canshehandlediversecommunicationstyles?Issheincrediblyarticulate?Hasshedoneherhomeworkonthecompany?

ACKNOWLEDGMENTS

Firstandmostimportant,Ithankmybeautifulwifeoftwenty-fiveyears,FeliciaHorowitz.It isa littleoddthankingher,becauseshehasbeensocentral tothestory that she is more properly the coauthor. She has been my number-onesupporter and her belief in me and in this book has meant everything. Therewouldbenobookwithouther,andthereisnomewithouther.Sheismypartnerin life and the love of my life, and I owe her everything that I have andeverything that I am. I do not have the words to express the magnitude ofgratitudethatIfeel.Felicia,IloveyouandIthankyou.Iamextremelygratefultothecountlesspeoplewhohelpedmethroughallthe

difficulttimesandtothepeoplewhohelpedmearticulatewhatallthatwaslike.Ihopethisbookpayssomeofthatforward.Next my mother, Elissa Horowitz, who always encouraged me to pursue

anything that Iwanted to do—fromplaying football towriting this book. Shebelievedinmewhennobodyelsedidandunderstoodmelikenobodyelseevercould.Thanks,Mom!Ialsothankmyfather,David,whoconvincedmethatwritingthisbookwasa

goodideaandthenputinlonghourshelpingmeeditit.Noneofthiswouldhavebeenpossiblewithoutmylongtimebusinesspartner

Marc Andreessen seeing things in me that nobody else did. Beyond that, it’sbeenamazing toworkwithhim for thepast eighteenyears.He’sbeenagreatinspirationtomeineverythingthatIdo.HewastheprimaryeditorofthefirstdozenblogpoststhatIwroteandaveryhelpfuleditorofthisbook.It’sagreatprivilegeformetoworkwithsomeoneofhiscalibereveryday.IthankmyfriendBillCampbellforteachingmesomanythingsabouthowto

survive thehard times.Sofewpeoplewent throughwhathewent throughandalmostnobody iswilling to talkabout it.Bill, thankyou foryourhonestyandcourage.MichaelOvitzhelpedmerewritetheendingofthebookandmakeittentimes

better.Beforethat,hedideverythingimaginabletosupportmeinmyimpossiblequest—uptoandincludingbuyingOpswarestockwhennobodyelsewould.He

isatruefriend.ToeveryemployeewhoeverworkedatLoudcloudorOpsware,I thankyou

from thebottomofmyheart. I still cannotbelieve that youbelieved inmeasmuchasIbelievedinyou.Onthatteam,veryspecialthankstoJasonRosenthal,MarkCranney,SharmilaMulligan,DaveConte,JohnO’Farrell,JordanBreslow,ScottKupor,TedCrossman, andAnthonyWright for being part of this book.Hopefully, I didn’t get anything too wrong. Thank you, Eric Vishria, EricThomas, Ken Tinsley, and Peter Thorp, for helping me remember whathappened.ThanksalsotoRaySoursa,PhilLiu,andPaulIngramforsavingthecompany. Darwin forever! Thanks so much, Shannon Callahan. I still cannotbelievethatIlaidyouoff.ThankstoDaveJagodafornotlettingmeforgetwhatmatters.I thankTimHowes,myLoudcloud/Opsware cofounder and confidant. I do

notknowifwemadealltherightdecisions,butIdoknowthatourconversationskeptmesane.Thankyouforbeingtherefromstarttofinish.WithoutCarlyeAdler,myeditorandcoach,IamnotsurethatIwouldhave

even started this book, let alone finished it. Nobodywasmore jazzedwhen Iwrote something good or sadder when I wrote something boring. Thank you,Carlye,formakingthisbooksomuchbetterthanitshouldhavebeen.Special thanks toHollisHeimbouchfor trackingmedownonFacebookand

gettingme towrite the book. I couldnot have asked for a better publisher.Athank-youtotheentireteamatHarperCollins.BinkyUrbanisthetopliteraryagentintheworld,andIamsoluckytohave

hadthechancetobeherclient.Itisagreatjoytoworkwiththeverybest.I thankmyfriendsNasirJonesandKanyeWestforbeingsoinspirationalin

their work and helping me articulate emotions that seemed impossible toexpress.Ialsoappreciatethemforlettingthisfanbackstage.SteveStoutehasbeenanoutstandingfriendthroughout thisprocess,helping

me find my voice and letting me know that the work that I am doing isimportant.Thankstomyoldestfriend,JoelClarkJr., forbeingagreatfriendforforty-

threeyearsandlettingmetellthestoryofhowwemet.ChrisSchroederhelpedmeedit thebookandmaintainedacrazyamountof

enthusiasmwhiledoingso.Chrisblowsmeawaywithhisinterestinthiswork.Manytimes,heseemsevenmoreinterestedthanIam.Thank you,HerbAllen, for being a great friend and lettingmewrite about

you.Iknowthatisnotyourfavoritething.

I thank all the partners and employees atAndreessenHorowitzwho put upwithmygrouchinessandincreasedprofanityasIwrotethis.Itwouldnothavebeen possible without you. Thank you for making the dream of a firm forfoundingCEOscometrue.A special thank-you to Margit Wennmachers for believing that I had

somethingtosayandhelpingmefindpeoplewhowouldhearit.Iamsoluckytobeworkinginthepresenceofsuchgreatness.GraceEllishasbeenbymysidethroughouttheprocess,handlingeveryweird

detailthatonecouldimagine.Duringthistime,Ihaveyettohearhercomplainaboutasinglething.Beyondthat,shehasgivenmegreatadviceandbeenagreatfriend.Thank you,KenColeman, for givingmemy first job and being a fantastic

mentorfornearlythirtyyears.Thanks to my brother in-law, Cartheu Jordan Jr., for being an important

character in this book and in my life. He is Branch Rickey to my JackieRobinson.Thank you, John and Loretta Wiley, for being so supportive of me in

everythingthatIdo.Thanks tomybrothersandsistersJonathanDaniel,AnneRishon,andSarah

Horowitzforshapingme.Loveyoualways,Sarah.I thank the late, greatMikeHomer for hiswisdom, help, and love. I thank

AndyRachleffforbeingagreatgentlemanandfriend.Thankyou,SyLorne,forkeepingmeoutoftrouble.Thankyou,MikeVolpi,forbeingontheboardofaveryscarycompany.Finally,thankyou,Boochie,Red,andBoogie,forbeingthebestchildrenthat

Icouldimagine.

ABOUTTHEAUTHOR

BEN HOROWITZ is the cofounder and general partner of AndreessenHorowitz, a Silicon Valley–based venture capital firm that invests inentrepreneursbuildingthenextgenerationofleadingtechnologycompanies.Thefirm’s investments include Airbnb, GitHub, Facebook, Pinterest, and Twitter.PreviouslyhewascofounderandCEOofOpsware,formerlyLoudcloud,whichwasacquiredbyHewlett-Packardfor$1.6billionin2007.Horowitzwritesabouthis experiences and insights from his career as a computer science student,softwareengineer,cofounder,CEO,andinvestorinablogthatisreadbynearlytenmillionpeople.HehasalsobeenfeaturedintheWallStreetJournal,theNewYork Times, the New Yorker, Fortune, the Economist, and BloombergBusinessweek,amongothers.HorowitzlivesintheSanFranciscoBayAreawithhiswife,Felicia.Visit www.AuthorTracker.com for exclusive information on your favoriteHarperCollinsauthors.

CREDITS

CoverdesignbyJarrodTaylor

Grateful acknowledgment is made for permission to reproduce from thefollowing:

“Gorgeous,” words and music by Malik Jones, Gene Clark, Jim McGuinn,Kanye West, Ernest Wilson, Mike Dean, Scott Mescudi, and Corey Woods.Copyright©2010UniversalMusicCorp.;JabrielIzMyne,TicksonMusicCo.;Sixteen Stars Music, EMI Blackwood Music, Inc.; Please Gimme MyPublishing, Inc.; Papa George Music; Chrysalis Music; Let the Story BeginPublishing;GeneClarkMusic; Elsie’s BabyBoy;Beautiful Sekai Publishing;and Chrysalis One Songs. All rights for Jabriel Iz Myne controlled andadministered by Universal Music Corp. All rights for Sixteen Stars MusiccontrolledandadministeredbyHoriproEntertainmentGroup,Inc.AllrightsforPlease Gimme My Publishing, Inc., and Papa George Music controlled andadministeredbyEMIBlackwoodMusic,Inc.AllrightsforLettheStoryBeginPublishingcontrolledandadministeredbyChrysalisMusicGroup,Inc.,aBMGChrysalis Company. All rights for Gene Clark Music controlled andadministered by BugMusic, Inc., a BMG Chrysalis Company. All rights forChrysalisOneSongscontrolledandadministeredbyBMGRightsManagement(Ireland), Ltd.All rights forElsie’sBabyBoy controlled and administered byKobalt Music Publishing America. All rights for Beautiful Sekai PublishingcontrolledandadministeredbyShellyBayMusic.Allrightsreserved.Usedbypermission.ReprintedbypermissionofHalLeonardCorporation,KobaltMusicPublishingAmerica,ShellyBayMusic,andAlfredMusicPublishing.

“WhoWeBe,”wordsandmusicbyEarlSimmonsandMickeyDavis.Copyright© 2001 Boomer X Publishing, Inc.; Dead Game Publishing; Fifty Four VillMusic, LLC; andKobaltMusic PublishingAmerica.All rights forBoomerXPublishing, Inc., controlled and administered by Universal Music Corp. Allrights for Dead Game Publishing controlled and administered by EMI AprilMusic,Inc.,andKobaltMusicPublishingAmerica.FiftyFourVillMusic,LLC,controlled and administered by theRoyaltyNetwork, Inc.All rights reserved.Usedbypermission.ReprintedbypermissionofHalLeonardCorporation, theRoyaltyNetwork,Inc.,andKobaltMusicPublishingAmerica.

“IWill Survive,” words and music by Dino Fekaris and Frederick J. Perren.Copyright © 1978 Universal-PolyGram International Publishing, Inc., andPerren-VibesMusic, Inc.All rights controlled and administered byUniversal-PolyGramInternationalPublishing,Inc.Allrightsreserved.Usedbypermission.ReprintedbypermissionofHalLeonardCorporation.

“On to the Next One,” words and music by Kasseem Dean, Jessie Chaton,ShawnCarter,MichelAndreAugeGaspard,andXavierDeRosnay.Copyright©2009UniversalTunes,adivisionofSongsofUniversal,Inc.;MonzaRonza,Universal Music Publishing MGB France, Options; Because Editions; WBMusicCorp.(ASCAP);andCarterBoysMusic(ASCAP).AllrightsforMonzaRonzacontrolledandadministeredbyUniversalTunes,adivisionofSongsofUniversal, Inc. All rights for Universal Music Publishing MGB France andOptions in the U.S. and Canada controlled and administered by UniversalMusic-MGBSongs.AllrightsforBecauseEditionscontrolledandadministeredbyBlueMountainMusic, Ltd. All rights on Behalf of Itself andCarter BoysMusic administered by WB Music Corp. All rights reserved. Used bypermission. Reprinted by permission of Hal Leonard Corporation and AlfredMusicPublishing.

“Screamon’Em,”byKasseemDeanandJayceonTerrellTaylor.Copyright©2006 Universal Tunes, a division of Songs of Universal, Inc.;Monza Ronza;Sony/ATV Music Publishing, LLC; Baby Game Music; and Pico PridePublishing. All rights for Monza Ronza controlled and administered byUniversalTunes,adivisionofSongsofUniversal, Inc.All rightsonbehalfofSony/ATV Music Publishing, LLC, and Baby Game Music administered bySony/ATVMusicPublishing,LLC,8MusicSquareWest,Nashville,TN37203.All rights forPicoPridePublishingcontrolledandadministeredbyCiw2, Inc.All rights reserved. Used by permission. Reprinted by permission of HalLeonardCorporation,Sony/ATVMusicPublishing,LLC,andCiw2,Inc.

“All Gold Everything,” words and music by Devon Gallaspy and NicholausJoseph Williams. Copyright © 2013 Sony/ATV Music Publishing LLC andTrinlantaPublishing.AllrightsonbehalfofSony/ATVMusicPublishing,LLC,administered by Sony/ATV Music Publishing, LLC, 8 Music Square West,Nashville, TN 37203. All rights reserved. Used by permission. Reprinted bypermissionofSony/ATVMusicPublishing,LLC,andTrinlantaPublishing.

“The Don,” by Salaam Remi, Dwight Myers, Marcos Palacios, Nasir Jones,ErnestClark,A.Marage,andMkrumahThomas.Copyright©2012EMIAprilMusic, Inc.; E-Z-Duz-It Publishing, EMI Blackwood Music, Inc.; LindenSpringfield, Universal Music-Z Songs; Nasir Jones Publishing Designee,Sony/ATV Music Publishing, LLC; Two Works; Viva Panama; and RufusMusic,Ltd.AllrightsforE-Z-Duz-ItPublishingcontrolledandadministeredbyEMI April Music, Inc. All rights for Linden Springfield controlled andadministered by EMI Blackwood Music, Inc. All rights for Nasir JonesPublishingDesigneecontrolledandadministeredbyUniversalMusic-ZSongs.All rights on behalf of Sony/ATVMusic Publishing, LLC, TwoWorks, andViva Panama administered by Sony/ATV Music Publishing, LLC, 8 MusicSquareWest,Nashville,TN37203.AllrightsforRufusMusic,Ltd,controlledand administered by Sunflower Entertainment Co., Inc. All rights reserved.Internationalcopyrightsecured.Usedbypermission.ReprintedbypermissionofHal Leonard Corporation, Sony/ATVMusic Publishing, LLC, and SunflowerEntertainmentCo.,Inc.

“Stronger,”words andmusic byThomasBangalter,Guy-ManuelDeHomem-Christo,EdwinBirdsong,andKanyeWest.Copyright©2007byDaftLife,Ltd.;Please Gimme My Publishing, Inc., EMI Blackwood Music, Inc.; EdwinBirdsong Music Publishing; WB Music Corp; and Zomba Music Publishing,Ltd.AllrightsforDaftLife,Ltd.,intheUnitedStatesandCanadaadministeredbyUniversalMusic-ZSongs.AllrightsforPleaseGimmeMyPublishing,Inc.,administered by EMI BlackwoodMusic, Inc. International copyright secured.All rights reserved. Contains sample of “Harder, Better, Faster, Stronger” byThomas Bangalter, Guy-Manuel De Homem-Christo, and Edwin Birdsong.ReprintedbypermissionofHalLeonardCorporation,AlfredMusicPublishing,andEdwinBirdsongMusicPublishing.

“Right Above It,” words andmusic by Aubrey Graham, Dwayne Carter, andDaniel Johnson. Copyright © 2010 EMI BlackwoodMusic, Inc.; LiveWrite,LLC; Artist Publishing GroupWest (ASCAP); Warner-Tamerlane PublishingCorp. (BMI); and Young Money Publishing, Inc. (BMI). All rights for LiveWrite, LLC, controlled and administered by EMI BlackwoodMusic, Inc. Allrights on behalf of Artist PublishingGroupWest administered byWBMusicCorp. All rights on behalf of Itself and Young Money Publishing, Inc.,administered by Warner-Tamerlane Publishing Corp. All rights reserved.

Internationalcopyrightsecured.Usedbypermission.ReprintedbypermissionofHalLeonardCorporationandAlfredMusicPublishing.

COPYRIGHT

THEHARD THINGABOUTHARD THINGS.Copyright©2014byBenHorowitz.Allrights reserved under International and Pan-American Copyright Conventions.By payment of the required fees, you have been granted the nonexclusive,nontransferable right to access and read the text of this e-book on-screen.Nopart of this text may be reproduced, transmitted, downloaded, decompiled,reverse-engineered,orstoredinor introducedintoanyinformationstorageandretrievalsystem,inanyformorbyanymeans,whetherelectronicormechanical,nowknown or hereinafter invented,without the expresswritten permission ofHarperCollinse-books.

FIRSTEDITION

LibraryofCongressCataloging-in-PublicationDatahasbeenappliedfor.

ISBN978-0-06-227320-8

EPubEditionMARCH2014ISBN9780062273215

ABOUTTHEPUBLISHER

AustraliaHarperCollinsPublishers(Australia)Pty.Ltd.Level13,201ElizabethStreetSydney,NSW2000,Australia

http://www.harpercollins.com.au

CanadaHarperCollinsCanada

2BloorStreetEast-20thFloorToronto,ON,M4W,1A8,Canadahttp://www.harpercollins.ca

NewZealandHarperCollinsPublishers(NewZealand)LimitedP.O.Box1Auckland,NewZealand

http://www.harpercollins.co.nz

UnitedKingdomHarperCollinsPublishersLtd.77-85FulhamPalaceRoadLondon,W68JB,UK

http://www.harpercollins.co.uk

UnitedStatesHarperCollinsPublishersInc.

10East53rdStreet

NewYork,NY10022http://www.harpercollins.com