Delivering personalised digital experiences to … that simply show the user’s name in header as a soft nod to pretending they ‘know’ them, these experiences are 100 per cent

CUSTOMERS EXPECT A LOT FROMDIGITAL

Ask a hundred people what their favouritedigital experience is and a lot of familiarnames start popping up — Facebook,Twitter, Google, Amazon and the like.Now ask yourself what these experienceshave in common? What makes them soconsistently amazing to most of the planet?What keeps people coming back to them?

In a word: personalisation. Unlike somewebsites that simply show the user’s namein header as a soft nod to pretending they‘know’ them, these experiences are 100per cent personalised to the user. TheTwitter, Facebook or Google experiencerendered at a specific moment in time isunique to the person viewing it,

impossible to replicate by any other user,and 100 per cent personalised based onhundreds of variables (some well knownand others a little surprising).

This is an important insight for anyonemanaging a web property because whenall the techno babble about ‘predictivesearch’, the ‘social network effect’ orAmazon’s ‘recommendation engine’quietens down, what is left is the simplereality that each of these websites is adigital window into the soul of theconsumer. Furthermore, like all thingsfrom the heart, the more personalised it is,the more relevant it is; and — setting asidethe wizardry it takes to deliver theseexperiences — when it comes to winningover customers, relevance is king.

� Henry Stewart Publications 2050-0076 (2014) Vol. 2, 2 131–138 Journal of Digital & Social Media Marketing 131

Delivering personalised digitalexperiences to customer segmentsof one: Case study of SAPReceived (in revised form): 27th May, 2014

Shawn Burnsis the Global Vice President of Digital Marketing at SAP, responsible for web strategy, design,development and operations for SAP’s flagship web property, www.sap.com, and its 72 country sites in40 languages. Winner of two Atticus Awards for thought leadership in digital marketing, Shawn is arecognised expert in the digital community and a regular keynote presenter at industry conferences ondigital and demand generation.

SAP, 32 Rue de Monceau, Paris 75008, France E-mail: [email protected]; Twitter: @shawnpburns

Abstract This case study describes SAP’s efforts to deliver a personalised digitalexperience to every customer visiting its flagship web property, SAP.com. The paperdiscusses the key components required to deliver the personalised experience, and theresulting performance improvements measured as a function of online engagement andsoftware revenue. Key benefits and challenges of the programme are also covered,providing a view into the deep insights SAP has realised as a result of the programme.The paper concludes with a glimpse at how SAP is planning to evolve these efforts.

KEYWORDS: digital marketing, big data, online personalisation, web analytics, predictiveanalytics, data modelling, marketing skills

Burns:JSC page.qxd 22/08/2014 11:20 Page 131

HOW IS THIS DIFFERENT FROMSEGMENTATION?

Marketers used to deliver relevancethrough basic segmentation as campaignmessages were tailored to any number ofcustomer segments. Looking at customersthrough simple filters of gender orgeography or buying cycle stage has alwaysimproved results compared with traditionalmass market approaches. By definitionthough, such an approach is finite; finite inthe number of segments that may betargeted, finite in the number of campaigniterations that may be produced, and finitein the diversity of media channels availablefor distribution, which is to say, what isthe point in having different messages ifthey are being distributed through thesame pipe?

Now, with digital personalisation — thecornerstone of world-class web experience— every marketing preconception hasbeen blown away. Digital personalisation isthe nirvana of marketing segmentationbecause there are no segments any more. Theperfect blend of big data and high-endstatistics (or, ‘we have all your informationand we know what to do with it’) enablessegmentation to happen at a unit of one.Now, every single customer is a uniquemarket segment and expects relevance tobe crafted solely around them. This hasbecome so common in the consumerspace that most people are unable toimagine not having fully personalisedexperiences. Imagine Facebook withoutyour wall; if Google suddenly stoppedfinishing your search queries for you; or ifTwitter started showing you generic feedsof people you had never followed.

IS THIS RELEVANT TO B2B?Digital personalisation has gone frominnovation to the stake required to play inthe consumer space, but does the samehold true in a business-to-business (B2B)environment? Like so many in the B2B

space, SAP has always found that resultsimprove when clients are targeted byindustry and line of business. This is hardlyrocket science: if you talk to people aboutwhat they specifically do, then results willimprove. The big question is whetherrobust digital personalisation cansignificantly improve results above basicsegmentation.

SAP’S EXPERIENCETo explore this question, SAP started withactive customers first. This provided twoimmediate benefits. First, these customersare already deeply engaged online withSAP, so any personalisation approachwould reach a statistically significant groupof buyers in the quickest timeframepossible (a key consideration in B2B testswith long sales cycles to validate results).The second benefit was behind the scenes,but arguably even more important. Bystarting with active customers, it waspossible to ask them to drive digitalpersonalisation through their own activeparticipation in the system, such asconfiguring their preferences, interests,future business outlook etc. Suchself-reported ‘interest data’ augmentedwhat was already known about them inorder to personalise the experience, andmade it possible to temporarilycircumvent the need for new fuzzy-logicor web-rendering systems, which can takesignificant budget and time to implementproperly.

The experience was driven from asubset of SAP.com called ‘Your SAP.com’,and invited active customers to register forthe personalised experience and provideadditional data points around businessobjectives, needs or pain points andspecific solution areas of interest, such asmobility, cloud, big data etc. Whenaugmented with known variables aroundthe customers’ industry, geography, line ofbusiness and the specific SAP solutions

Delivering personalised digital experiences to customer segments of one

132 Journal of Digital & Social Media Marketing Vol. 2, 2 131–138 � Henry Stewart Publications 2050-0076 (2014)


they were already running, it was possibleto develop content streams personalised tosingle individuals. As an example, thiswould mean a small business ownerinterested in entry-level business softwarewould see an entirely different websiteexperience, content, imagery etc thanwould the chief financial officer of aFortune 500 company interested inspecialised analytics software for hisindustry.

The core of the personalisation enginewas a two-stage process focused onanalytics first and content delivery second.The analytics engine was continuouslyrunning a multivariate data model takingin all the data points from theabovementioned sources (a combination ofSAP customer relationship management(CRM), SAP.com and third-partyaugmented data sources such as Googlesearch terms used, public profile data fromLinkedIn etc) and calculating a ‘bestcontent recommendation’ and ‘lead score’for each individual in the programme.Content delivery was an automated e-mailand webpage rendering system which senta regular stream of new contentrecommendations to each individual withlinks to their personalised version of ‘YourSAP.com’. Thus, a retail-industry managerinterested in supply chain would receivean entirely different series of content thanwould a banking manger interested inmobile security solutions (see Figures 1and 2). The more the user engaged withthese pages, the smarter the predictionwould get, and the user would always seecontent that was new to them.

To meet the volume requirements ofthe personalisation programme, contentspanned every available asset includingvideos, customer references, white papers,blogs, solution demos etc. To ensure anappropriate evaluation of the programmeon a global level, it was deployed inGermany, Brazil, USA, India and theMiddle East region. All deployments were

in local language and housed on the localversions of SAP.com dedicated to thesegeographies.

PERFORMANCE AND THEN SOMEThe business-to-customer valueproposition around driving a personaliseddigital experience held perfectly in theB2B space. Indeed, the aggregated resultsof the personalisation pilot exponentiallyoutperformed anything SAP hadpreviously put into market. That theimpact was greater than forecast was likelydue to a lack of digital innovation in theB2B tech space, that is, the novelty factorplayed a role, as customers were used toseeing the ‘same old thing’ in the B2Bspace.

The following results illustrate theimpact nicely:

• click-through rate on personalisatione-mails: 350 per cent;

• interaction on ‘Your SAP.com’ website:67 per cent;

• recommended assets downloaded fromsite: 28 per cent;

• return rate to ‘Your SAP.com’ website:40 per cent.

To understand these numbers in context,one needs to understand that the typicalconversion rates for a non-personaliseddigital campaign and web experience inthe B2B space average 2–5 per cent(according to both SAP efforts andindustry averages). Simply put, the webexperience layer has outperformed everydigital marketing effort SAP has previouslydeployed.

The final performance indicator is ofcourse revenue. In this regard, theperformance gains experienced at the weblayer continued to be seen through thesales process, with a full 33 per cent ofregistered customers of the personalised webexperience converting to new sales


Burns


revenue as a result of the programme. Theresults outperformed the company’shighest estimates and set in motion anentirely new strategic direction for theweb properties, content and platforms atSAP.

LESSONS LEARNEDSAP’s ambitions were bold when itembarked on the personalisationprogramme and by nature, so were thechallenges and the eventual learnings ofthe effort.



Figure 1: ‘Your SAP.com’ built around a predicted interest for wholesale distribution


Components required to drive digitalpersonalisation

SAP identified the following componentsas essential to driving digitalpersonalisation:

• Data: The information sources neededto drive relevant personalisation

spanned company and sales data fromCRM (industry, location, purchasehistory, purchase dates and renewaltimeframes), personal data on jobposition, line of business, interest areas,stage in buying cycle and role in buyingprocess (eg decision-maker) in additionto third-party data across public profiles


Burns

Figure 2: ‘Your SAP.com’ built around a predicted interest in the banking industry


(eg LinkedIn), cookie data onpreviously visited websites and Googlesearch strings.

• Analytical model: A predictivemultivariate statistical model is requiredto take the above data sources and theavailable content and makerecommendations to the user basedentirely on their likelihood of engagingmore deeply and their propensity tobecome a sales lead.

• Deployment landscape: A new digitalplatform is required to run theanalytical model continuously, adjustingfor new data sources, content availabilityand stage in the customer buying cycle.Customers have different contentappetites depending on whether theyare early or late in their buying cycle;for example, customers who have justbought new software do not desireweekly content updates.

• New marketing talent: Across datasourcing, model development andtechnical deployment, an entirely newbreed of marketing talent is required.This new and growing breed of talentlooks nothing like a traditional‘marketing communications’ skill set.The dearth of these skills in the marketmeans that success depends on acombination of smart recruitment andleveraging agency partners.

ScaleWithout fully realising it at theconception stage, the efforts to deliver apersonalised digital experience forcustomers cut across almost everyorganisational silo in the company. Fromsourcing content, to deploying theanalytics model, to the necessary ITdevelopment changes for existing weband e-mail platforms, to design, toreporting, to integration with the salesteams through SAP CRM etc, it took avillage of people to build and deploy theprogramme at this level.

Readers will need to set realisticexpectations on their own attempts in thisspace as even simple personalisation effortsrequire many disparate teams to worktogether. Success requires executivesponsorship, buy-in across marketing andsales leadership, and budget commitmentsspanning multiple planning cycles. As aside note, it is easy to see why one-shotcampaigns, with lower performanceresults, will continue to be the normacross marketing organisations. The easewith which small teams can execute suchcampaigns makes them tough to beat forcompanies chasing short-term quarterlyresults.

Content, content and more contentIt is impossible to overstate the contentrequirements of a programme designed todeliver a personalised experience to everyunique customer. Marketing organisationstypically think in terms of single messagesupdated gradually over time — consider,for example, a product overview, a newproduct announcement or promotionalmaterial for a specific product. Each ofthese is created once and traditionallywould get refreshed on an as-needed basis,typically resulting in a slow-moving annualcontent update cycle. This meansmarketing material across key customerinterest areas, such as the automotiveindustry, analytics solutions or clouddeployment, may change significantly onlyonce a year at SAP. When comparing thiswith SAP’s new customer valueproposition — continual new contentpersonalised to the customer’s interests —the disconnect is immediately evident. Thepersonalisation programme created a needfor new content like nothing the companyhad seen before.

Closing the content gap is anever-ending task that requires entirelynew approaches to sourcing and creatingcontent. Depending on one’s point ofview, it also creates a vicious or virtuous




cycle, where more content allows fordeeper levels of personalisation, whichdraws in more customers, which in turncreates the need for yet more content. Tomeet SAP’s growing need for content, afour-prong process was employed to create10–15 new pieces of content daily insupport of digital personalisation:

1. Evolve the current content model to become‘modular’. Focusing on smaller,web-ready material as opposed to thetraditional print-ready brochure-styleapproach, made it possible to have moremarketing content available on a morefrequent basis. This new approachcreated a regular stream of newlyavailable content that worked inmonthly and quarterly cycles, in lieu ofthe historical, annual refresh timeframes.

2. Embrace user-generated content. Socialmedia, including the SAP CommunityNetwork, became a ripe source of newand dynamic content (eg comments,ratings, shares). This required theorganisation to get comfortable withnon-SAP authored content on thewebsite, but with appropriatemonitoring, the perceived risks of suchcontent can be almost entirelymitigated.

3. Leverage physical events for dramatic increasesin rich media. Customers find video to besome of the most engaging content onthe web, yet marketers regard it as one ofthe most expensive and difficult forms ofcontent to develop. Solution? Recordphysical events. Through recording themost significant SAP events (eg theannual SAPPHIRE NOW conferencebrings together 20,000 participants across300+ different topical sessions) SAP cancreate massive increases in the amount ofrich media available to support onlinepersonalisation efforts for very little cost.

4. Expand into alternative content. It soundsalmost heretical to ask an organisationto spend time and money creating

content that is not specific to thecompany or the solutions it develops.However, in the digital space, thecompetition for SAP’s customers’ timeis not other software companies offeringsimilar products (eg Oracle, Salesforce).The real competition is Wikipedia,Forbes and every other website focusedon content aggregation. To win in thisspace, it was necessary to get topical,meaning devoting time to creatingcontent around the broader subjectsthat our customers were looking for,such as, ‘the power of big data to drivedeeper customer relationships’,‘leveraging cloud computing to lowerCAPEX’ or ‘performance benefits ofmobilising the workforce’.

The power of automationAfter getting content sorted out, thedigital platform developed, the analyticalmodel and all the other elements requiredfor the successful launch of theprogramme, something amazing happens:the programme runs — continuously.New customers are continuing to sign upand old customers are continuallyreturning to engage with the content thathas been recommended to them. Thisbecomes a true always-on marketingengine that requires only the fuel of freshcontent to run forever. While this seemslike an obvious observation in hindsight,the power of automation was never trulyunderstood until we kept seeing more andmore sales leads and revenue coming fromthe programme without any additionaleffort.

Thus, one can make two comments onthe return on investment (ROI) of digitalpersonalisation, both of which can be usedto justify marketing investment in a similarprogramme. In the short term — as borneout by the results of the pilot programme— digital personalisation demonstratessignificantly better results than anytraditional marketing campaign.


Burns


Considering the upfront costs, however,fence-sitters may be unconvinced thatdigital personalisation is the better choice.However, when looking at multi-yearROI, the programme becomes ano-brainer for marketing departmentseverywhere. While traditional campaignsrequire the same level of fixed investmentyear after year, digital personalisationrequires this one-off up-front costfollowed by budget for ongoing contentmaintenance only. Compared withannualised campaigns that require budgetfor new creative designs, media andimagery every year, the cost ofmaintaining content is minimal, and ROIcan be exponential over a 2–3 yeartimeframe.

THE FUTURE: FROM KNOWN TOUNKNOWNSAP’s personalisation programme wasoriginally targeted to active customerswho were willing to register on SAP.com

and provide additional data pointsregarding their interests. While thisapproach provides the benefits outlined inthis paper, it suffers from a significanthandicap to growth: it is focused on theminority of SAP’s website visitors — thosewilling to register. Across all of SAP’s webproperties, the highest registration rateseen thus far is approximately 10 per centof visitors. This means that forpersonalisation to grow, the focus must beon the 90 per cent of visitors who are notregistering when they visit. This is theholy grail of the company’s efforts andwhere SAP is focused through 2014 and2015. With advances in technology,specifically in the area of cookies,retargeting and reverse IP lookup at anidentifiable company level, SAP is workingtowards providing a fully personalisedwebsite for users who have done nothingmore than visit SAP.com. The sky is thelimit when this begins and it is possible todeliver a personalised experience for everysingle person who comes to SAP.com.




Documents

Delivering personalised digital experiences to … that simply show the user’s name in header as a soft nod to pretending they ‘know’ them, these experiences are 100 per cent