iving feedback to your em-ployees, particularly when their performances fall short of ex-

pectations, is one of the most criticalroles you play as a manager. For mostpeople, it’s also one of the most dreaded.Such conversations can be very un-pleasant – emotions can run high, tem-pers can flare. And so, fearing that anemployee will become defensive andthat the conversation will only strainthe relationship, the boss all too ofteninadvertently sabotages the meetingby preparing for it in a way that stifleshonest discussion. This is an uninten-tional – indeed, unconscious – habitthat’s a byproduct of stress and thatmakes it difficult to deliver correctivefeedback effectively.

The good news is that these conver-sations don’t have to be so hard. Bychanging the mind-set with which youdevelop and deliver negative feedback,you can greatly increase the odds thatthe process will be a success – that youwill have productive conversations, thatyou won’t damage relationships, andthat your employees will make real

improvements in performance. In thepages that follow, I’ll describe what goeswrong during these meetings and why.I’ll look in detail at how real-life con-versations have unfolded and what themanagers could have done differentlyto reach more satisfying outcomes. Asa first step, let’s look at the way bossesprepare feedback–that is, the way theyframe issues in their own minds in ad-vance of a discussion.

Framing Feedback

In an ideal world, a subordinate wouldaccept corrective feedback with an openmind. He or she would ask a few clari-fying questions, promise to work on theissues discussed, and show signs of im-provement over time. But things don’talways turn out this way.

Let’s consider the following example.Liam, a vice president at a consumerproducts company, had heard somecomplaints about a product manager,Jeremy. (Names and other identifyinginformation for the subjects mentionedin this article have been altered.) Jer-emy consistently delivered high-quality

To o l K i t

Critiquing weak performance is a job

nobody likes. But by taking a more open

approach, you can be a better boss –

and get a lot more from your team.

A Better Way to Deliver

by Jean-François Manzoni

BadNews

G work on time, but several of his subor-dinates had grumbled about his appar-ent unwillingness to delegate. They felttheir contributions weren’t valued andthat they didn’t have an opportunityto learn and grow. What’s more, Liamworried that Jeremy’s own career pros-pects would be limited if his focus onthe day-to-day details of his subordi-nates’ work kept him from taking onmore strategic projects. As his boss, Liamfelt a responsibility to let Jeremy knowabout his concerns. Here’s how the con-versation unfolded:

Liam: “I’d like to discuss your workwith you. You’re doing a great job, andwe really value your contributions. ButI think you do too much. You have somegreat people working for you; why notdelegate a little more?”

Jeremy: “I don’t understand. I dele-gate when I think it’s appropriate. Buta lot of people in this company rely onquality work coming out of my depart-ment, so I need to stay involved.”

Liam:“Yes, and we all appreciate yourattention to detail.But your job as a man-ager is to help your employees grow into

4 Copyright © 2002 by Harvard Business School Publishing Corporation. All rights reserved.

september 2002 5

the conversation escalates until a rela-tively minor difference becomes muchmore dramatic. (For a visual represen-tation of a deteriorating discussion, seethe exhibit “Scripted Escalation.”) Often,as Liam did in the preceding conversa-tion, one person or the other uninten-tionally says something overly critical.Of course, it may not get to that point–one or both parties may choose to givein rather than fight. But either way,escalate or fold, the subordinate proba-bly hasn’t accepted the news the bossset out to deliver. Managers tend toattribute such nonacceptance to em-ployees’ pride or defensiveness. Indeed,it’s not unusual for people to feel de-fensive about their work or, for thatmatter, to hold inflated views of their

new roles and take on more responsibil-ity. Meanwhile, you’re so focused on thedetails that you don’t have time to thinkabout the bigger picture, about the di-rection you’re taking this product.”

Jeremy: “That’s not true. I’m alwaysthinking about the future.”

Liam: “I’m just saying, you’d havemore time for strategic thinking if youweren’t so mired in the day-to-day stuff.”

Jeremy: “Are you saying I’m not astrategic thinker?”

Liam: “You’re so busy dotting everyi and crossing every t that I just don’tknow what kind of thinking you’re ca-pable of!”

This type of exchange is surprisinglycommon. Each side pushes his point ofview more and more aggressively, and

performance and capabilities. But moreoften than not, the boss is also to blame.Let’s examine why.

Whenever we face a decision or situ-ation, we frame it, consciously or not.At its simplest, a frame is the decisionmaker’s image of a situation – that is,the way he or she pictures the circum-stances and elements surrounding thedecision. The frame defines the bound-aries and dimensions of the decisionor situation – for instance, which issueswill be looked at, which components arein and which are out, how various bitsof information will be weighed, howthe problem might be solved or a suc-cessful outcome determined, and soon. Managers tend to frame difficult sit-uations and decisions in a way that is

Take a look at how quickly a minor point of difference

during a feedback discussion can turn into a major

disagreement. Jerry starts the conversation by noting

that he’d done a good job on his project. Beth, his boss,

is not in violent disagreement with his assessment and

acknowledges that “it wasn’t bad.” Jerry could reaffirm

his opening bid but instead tries to pull Beth’s view

closer to his own by overstating his initial point. Beth

disagrees with Jerry’s inflated statement, and instead of

reiterating her first comment, she yields to the tempta-

tion to pull Jerry closer to her point of view. Both present

stronger and stronger positions, trying to convince the

other, and a minor difference quickly becomes a major

point of contention.

Scripted Escalation

“I did OK.”

“What do you mean,it wasn’t bad? It was pretty damn good!”

“Come on,it was great!”

“It wasn’t bad.”

“But there wereproblems.”

“And the problems were pretty severe.”

“Come to think of it,it really wasn’t very good.”

“Listen, I did amazingly well!”

Gap at the end of the conversation

J4 J3 J2 J1 B1 B2 B3 B4

Initial gap

Jerry (Subordinate) Beth (Boss)

narrow (alternatives aren’t included oreven considered) and binary (there areonly two possible outcomes – win orlose). Then, during the feedback discus-sion, their framing remains frozen – un-changed, regardless of the direction theconversation takes.

In anticipation of the conversationwith Jeremy, for example, Liam framedthe problem in his mind as “Jeremy’stoo controlling.” This is a narrow fram-ing because it excludes many alterna-tive explanations–for instance,“Jeremywould really like to hand off some re-sponsibility but doesn’t know how andis embarrassed to acknowledge that.”Or “Jeremy is actually delegating asmuch as he can given his subordinates’current skill levels; they are frustratedbut really cannot handle more than theydo.” Or maybe “Jeremy is delegatingquite a lot, but Frank and Joan havesome other ax to grind.” Liam may bemaking matters worse without realiz-ing it by sending Jeremy mixed signals:“Empower your subordinates, but makeno mistakes.” We don’t know for sure;nor does Liam.

Operating from this narrow view,Liam also approached the discussionwith a binary framing that leaves bothparties with very little room to maneu-ver: “Jeremy must learn to delegate orwe’ll lose Frank and Joan – and mean-while, he’ll burn himself out.” Last butnot least, Liam’s framing remainedfrozen throughout the exchange despiteclear signals that Jeremy was not buyingthe feedback. At no point was Liam pro-cessing, let alone addressing, Jeremy’sobjections. It’s no surprise that the meet-ing ended badly.

The Dangers of Easing In

After they’ve had a few bad experiencesdelivering narrowly framed feedback,managers tend to fall back on the con-ventional wisdom that it’s better tosoften bad news with some good.

They try to avoid uncomfortable con-frontations by using an indirect approach:They make up their minds about anissue and then try to help their employ-ees reach the same conclusions by ask-ing a carefully designed set of questions.

At first glance, this type of “easing in”seems more open and fair than theforthright approach that Liam took,since the manager is involving the sub-ordinate in a conversation, howeverscripted. But like the forthright ap-proach, easing in reflects a narrow andbinary framing that typically remainsfrozen throughout the process. Indeed,there would be no need to ease in if themanager were approaching the conver-sation with a truly open mind. And eas-ing in carries an additional risk: The em-ployee may not give you the answersyou’re looking for.

For example, Alex, an executive at apharmaceuticals company, had somedifficult news to communicate to one ofhis subordinates, Erin. She was a middlemanager at the company and did anexcellent job handling her departmentbut was not contributing satisfactorilyto a companywide task force chaired byAlex. Erin was remarkably silent duringthe meetings, which led Alex to con-clude that she was too busy to partici-pate fully and had little to offer thegroup. Alex’s solution? Take her offthe task force so she could focus on herprimary responsibilities. But becausehe suspected Erin would be hurt or in-sulted if he suggested she step down,Alex hoped to prompt her to resignfrom the committee by asking her a se-ries of questions that would make hersee she was too busy to continue. Let’slook at what happened.

Alex: “Do you sometimes feel asthough you’re wasting your time in thetask force meetings?”

Erin:“No, I learn a lot from the meet-ings – and from watching the way yourun them.”

Alex:“But do you find that your mindis on your daily job when you’re at com-mittee meetings?”

Erin: “Not really. I hope I haven’tgiven you the impression that I’m notfully committed. I think this is impor-tant work, and I’m excited to be a partof it, and I think I have some good ideasto offer.”

Alex: “What if you could participatemore informally? You could take your-self off the team as a permanent mem-ber, but you could continue to receivethe agenda and minutes and contributewhen your particular area of expertise isrequired.”

Erin: “It sounds like you want me offthe committee. Why? I don’t think thecommittee work has undermined mycommitment to my real work. I’m mak-ing my numbers. Plus, it’s a learningopportunity.”

Alex:“No, no, I just want to make sureit’s something you really want to do.”

Erin: “It is.”As you can see, Erin didn’t play along.

Alex was not ready for a confrontation,so he folded–and lost. He didn’t get Erinoff the committee, nor did he commu-nicate his view that her committee workwas subpar, so he has no way to helpher improve her performance. What’smore, he introduced a source of stressinto their relationship: Erin is likely tohave been unsettled by the interaction,as Alex implied some level of dissatis-faction with her performance withouttelling her what it is.

As in our previous example, Alex’sframing of the issue was narrow: “Erindoesn’t talk at the meetings, probablybecause she’s overloaded, so the com-mittee is a waste of her time.” His fram-ing was also binary; the interactioncould be a success only if Erin agreed toget off the committee without losingher motivation for her regular work.And this framing remained frozen be-cause Alex was concentrating on askingthe “right” questions and couldn’t pro-cess anything but the “right” answers.

Meanwhile, Erin may actually benefitfrom being on the committee, even ifshe doesn’t say much. She learns a lot,and it gives her visibility. And if she can

6 harvard business review

T O O L K I T • A Better Way to Deliver Bad News

Jean-François Manzoni is an associate professor of management at Insead inFontainebleau, France, and the director of the Insead–PricewaterhouseCoopers Re-search Initiative on High-Performance Organizations. He is a coauthor, along withJean-Louis Barsoux, of The Set-Up-to-Fail Syndrome: How Good Managers CauseGreat People to Fail (Harvard Business School Press, 2002).

find a way to contribute more, the com-mittee may well benefit from her mem-bership. But by framing the issue theway he did, Alex excluded other possiblesolutions, any of which may have beenmore productive for all concerned:Maybe Erin would talk more in themeetings if Alex probed the reasons forher silence and helped her find a way tocontribute what may be very valuableinsights. And if overwork is indeed anissue, perhaps there are duties Erinmight give up to gain more time andenergy.

Easing in is a gamble. You might getlucky, but you have only half the cards.The subordinate may not give you theanswers you’re looking for, as we sawwith Erin, either because she genuinelydoesn’t agree or because she sees thatthe game is rigged and refuses to playalong. Or the subordinate may decide tostop resisting and pretend to go along

but still fail to believe the feedback. Andthere’s another risk, regardless of howthe conversation ends: The employeemay forever lose confidence in his or herboss. Erin may always wonder what Alexhas up his sleeve, having caught himbeing disingenuous once.

Indeed, that’s what happened toMark, a marketing director at a largeconsulting firm. His boss, Rene, hadcalled him into a meeting to discuss hisrole, and Mark left the meeting havingrelinquished control of his pet project,developing and implementing the com-pany’s first advertising campaign. Renehad asked him a series of seemingly in-nocuous questions, such as “Do you findendless meetings with different agen-cies to be a waste of your time?” and“Do you feel like your time would bebetter spent developing new communi-cations materials?” Mark eventually ac-cepted what was clearly the “right” con-

clusion from his boss’s perspective – tosurrender the project – even though hewanted to continue. Worse, he didn’tknow why Rene wanted him off theproject, so as a learning opportunity, itwas wasted. His relationship with hisboss is now tainted; Mark can no longertake Rene’s comments at face value.

Why Is It So Hard?

It’s very clear from a distance what wentwrong for Liam and Alex. Most man-agers today are well trained and wellmeaning; why can’t they see whatthey’re doing wrong? The tendency toframe threatening situations in narrowterms can be traced to the combinationof several phenomena.

First, research shows that when ana-lyzing others’ behavior, most peopletend to overestimate the effect of a per-son’s stable characteristics–the individ-ual’s disposition and capabilities – andunderestimate the impact of the spe-cific conditions under which that personis operating. So, for instance, a managerwill attribute a subordinate’s perfor-mance problems to his or her disposi-tion rather than to circumstances inthe workplace, leading to a rather sim-plistic interpretation. This phenome-non is known as the fundamental attri-bution error.

Second, people are more prone tocommitting the fundamental attribu-tion error when they operate under de-manding conditions. We can better dis-tinguish the impact of situational forceswhen we have time and energy to sparethan when we face multiple demandson our attention. Unfortunately, man-agers tend to be busy. Facing huge work-loads and tight deadlines, they have lim-ited time and attention to engage inexhaustive analyses of all the potentialcauses of the situations they observe orof the many possible solutions to agiven problem. So they settle on the firstacceptable explanation. “Jeremy’s toocontrolling” explained all the symp-toms, so Liam did not go further.

Research can also give us some in-sight into why bosses tend to framethings in a binary way. In particular,Harvard Business School professor

september 2002 7

A Better Way to Deliver Bad News • T O O L K I T

Research shows that people tend to be more willing to accept

feedback when they have the feeling that:

� The person offering the feedback is reliable and has

good intentions toward them.

� The feedback development process is fair – that is, the

person giving the feedback collects all relevant infor-

mation; allows the subordinate to clarify and explain

matters; considers the subordinate’s opinions; and

applies consistent standards when delivering criticism.

� The feedback communication process is fair – that is,

the person offering the feedback pays careful attention

to the subordinate’s ideas; shows respect for the sub-

ordinate; and supports the subordinate despite their

disagreements.

This short list makes clear the negative impact of approaching

a feedback discussion with restrictive framing: Narrow framing

tells the employee that the feedback wasn’t developed fairly.

And a boss constrained by a binary and frozen frame comes

across as biased, closed minded, and unsupportive – ensuring

that the subordinate will feel as though the feedback hasn’t been

communicated fairly.

Making Feedback More Acceptable

Chris Argyris’s work over nearly fivedecades has established that understressful circumstances, people behavein predictable ways. They design theirbehaviors, often unconsciously, to gaincontrol of a situation and to win–whichmeans, unfortunately, that the otherside usually has to lose. That’s binaryframing.

And why is it so hard for bosses to revise their restrictive framing mid-stream? For several powerful reasons.First, bosses don’t set out to frame situ-ations in restrictive ways; they do so un-consciously, most of the time, and it’shard to question a constraint that wedon’t know we’re imposing on our-selves. Second, humans tend to assumethat other reasonable people will seethe situation as they see it. That’s calledthe false consensus effect. Our framing ofan issue represents our view of reality,the facts as we see them. We are reason-able and competent people; why wouldothers see the situation differently?

Bosses can get past these hurdles byrecognizing them and becoming moreconscious and careful when framingdecisions. But then they have to beatanother cause of frozen framing: a busyprocessor. For instance, Liam becomesincreasingly stressed as Jeremy contin-ues to push back against his version ofthe facts, and both devote so much en-ergy to trying to control their growingirritation that they have few resourcesleft to listen, process, and respond con-structively.

Reframing Feedback

Let’s be clear: I’m not suggesting thatbosses systematically misdiagnose thecauses of their subordinates’ perfor-mance problems. Liam’s and Alex’s earlydiagnoses may well have been right.And even if their feedback discussionshad been more productive, their subor-

dinates may not have been able to suf-ficiently improve their performances tomeet their bosses’expectations. But Jer-emy and Erin will almost certainly fail to improve if they don’t understand andaccept the feedback. Restrictive fram-ing not only makes feedback conversa-tions more stressful than they need tobe, it also increases the likelihood thatsubordinates won’t believe what theirbosses say. Indeed, subordinates aremore likely to accept and act on theirbosses’ feedback if they feel it is de-veloped and communicated fairly. (Seethe sidebar “Making Feedback MoreAcceptable.”)

So, for instance, imagine how differ-ently Liam and Jeremy’s conversationmight have gone had the managerframed his concerns more broadly: “I’veheard complaints that Jeremy isn’t del-egating – and some of his employeesare feeling sufficiently frustrated thatI’m afraid we’ll start losing them. I’dlike to find out if Jeremy knows aboutthe complaints and get his take on thesituation.”

This frame isn’t narrow. Liam hasn’treached a conclusion about why Jeremydoesn’t delegate or whether, indeed,Jeremy is refusing to delegate at all. Noris the frame binary. Liam hasn’t fixedon a win-or-lose outcome. And becauseLiam hasn’t entered the conversationwith a preconceived outcome in mind,he has nothing on which to freeze. Now,Liam can open the conversation in amuch more open way. He might say, forinstance,“Jeremy, I don’t know if you’reaware of this – or if it’s true or not – butI’ve heard that Frank and Joan are anx-ious to take on a bit more responsibility.What do you think?” This can lead toa discussion of Frank’s and Joan’s capa-bilities, as well as Jeremy’s own role andaspirations, without locking Jeremyand Liam into a test of wills.

As for Alex, instead of approachingthe meeting with the goal of gettingErin off the committee with minimaldamage, he could have framed the in-teraction more broadly: “I have a greatsubordinate who doesn’t say much onthe committee. Let’s sit down and talkabout her work, the committee, her ca-reer plans, and how committee mem-bership fits in with those plans.”Becausethis framing doesn’t fix on a win-or-loseoutcome, Alex would have felt less needto control the discussion and hence lesscompelled to ease in.

• • •While most managers can easily seewhat they’re doing wrong when shownhow they’ve developed and presentedtheir feedback, restrictive framing re-mains a surprisingly persistent problem,even for seasoned managers who excelat other aspects of leadership. But giv-ing feedback doesn’t have to be stressfulfor you, demoralizing for your employ-ees, or damaging to your professionalrelationships.

Offering more effective critiques re-quires that you learn to recognize the bi-ases that color the development of feed-back. It requires that you take the timeto consider alternative explanations forbehaviors you’ve witnessed rather thanleaping to hasty conclusions that onlyserve to paint you and your subordi-nates into a corner. And it requires thatyou take into account the circum-stances an employee is working underrather than attributing weak perfor-mance to the person’s disposition.

In short, it requires a broad and flex-ible approach, one that will convinceyour employees that the process is fairand that you’re ready for an honestconversation.

Reprint r0209jTo place an order, call 1-800-988-0886.

T O O L K I T • A Better Way to Deliver Bad News

8 harvard business review