Blog

Marketing & SalesThe future ofpersonalization—andhow to get ready for itPersonalization will be the prime driver of marketing success withinfive years. Here are the capabilities companies need to develop tostay ahead of the curve.June 2019by Julien Boudet, Brian Gregg, Kathryn Rathje, Eli Stein, and Kai Vollhardt© Getty ImagesThe exciting promise of personalization maynot be here yet (at least not at scale), but it’snot far off. Advances in technology, data, andanalytics will soon allow marketers to createmuch more personal and “human” experiencesacross moments, channels, and buying stages.Physical spaces will be reconceived, and customerjourneys will be supported far beyond a brand’sfront door.While these opportunities are exciting, mostmarketers feel underequipped to deliver. A recentMcKinsey survey of senior marketing leadersfinds that only 15 percent of CMOs believe theircompany is on the right track with personalization.But there’s a big incentive to figure it out. Today’spersonalization leaders have found proven waysto drive 5 to 15 percent increases in revenue and10 to 30 percent increases in marketing-spendefficiency—predominantly by deploying productrecommendations and triggered communicationswithin singular channels.Positioning businesses to win requiresunderstanding the three main shifts inpersonalization and building up the necessaryskills and capabilities to respond to them.Read on to learn what it will take to win tomorrow’spersonalization game.Three major shifts will makepersonalization more personalOver the next five years, we will see three majorshifts in personalization:Physical spaces will be ‘digitized’Fewer than 10 percent of the companies wesurveyed currently deploy personalization beyonddigital channels in a systematic way. That presentsa big zone of opportunity. One area where theimplications could be significant is in store visits.Our survey data suggest that “offline” personto-person experiences will be the next horizonfor personalization. Some 44 percent of CMOssay that frontline employees will rely on insightsfrom advanced analytics to provide a personalizedoffering; 40 percent say that personal shopperswill use AI-enabled tools to improve service; and37 percent say that facial recognition, locationrecognition, and biometric sensors will become morewidely used.Some retailers have already started down thispath to move beyond established, though stillrudimentary, personalization practices. At Covergirl’snew flagship store, an AI-powered program, enabledby Google’s conversational Dialogflow platform,directs customers, while augmented-reality glamstations let customers “virtually try” products—byaltering the customer’s image as if the product hasbeen applied. But this doesn’t mean the end of thesalesperson or stylist. These virtual experiences stillneed the human touch. Covergirl’s glam stations stillneed customers to tell stylists what products they’dlike to try. As AI evolves, systems can generaterecommendations based on analyzing a customer’sskin tone, facial features, and emotions in real timeto tailor what to recommend or avoid offering.Macy’s, Starbucks, and Sephora are usingGPS technology and company apps to triggerrelevant in-app offers when customers near astore. Other retailers have begun to provide salesassociates with apps that generate personalizedproduct recommendations for specific customersautomatically. One retailer found an app like thisgenerated a 10 percent lift in incremental sales and a5 percent increase in transaction-size growth.The next level of in-store personalization is likely toinclude providing these kinds of experiences to allcustomers as well as pulling in more advanced ARfeatures to help customers experience products andservices in different environments, such as tryinghiking boots on a “virtual mountain.”Empathy will scaleEmpathy—the ability to relate to and understandanother person’s emotions—is the basis of strongrelationships. Understanding social cues andadapting to them is how people build trust. That’s noteasy to do digitally or at scale.2 The future of personalization—and how to get ready for itMachine learning is changing that, or at least gettingmuch better at reading and reacting to emotionalcues. More sophisticated algorithms are allowingprograms to interpret new kinds of data (visual,auditory) and extrapolate emotions much moreeffectively than in the past. Amazon has patentednew features that will enable its Echo device todetect when someone is ill—such as nasal tonesthat indicate a stuffed nose. It will then offer asuitable recommendation, such as a chicken-souprecipe or cough drops, some of which could thenbe purchased over the device for at-home delivery.Other companies are getting into the game too.Affectiva, which spun off from work scientistswere doing at the MIT Media Lab, is usingmachine learning to develop emotion-recognitionalgorithms to classify and map facial expressions,such as anger, contempt, disgust, fear, and joy. Todate, the company has raised $53 million frominvestors including Kleiner Perkins, CAC Holdings,and the National Science Foundation.¹In time, these advances could help marketerscommunicate with customers in a way that’s tiedto specific moods, offering specifically curated¹ Julien Boudet, Brian Gregg, Jason Heller, and Caroline Tufft, “The heartbeat of modern marketing: Data activation and personalization,” March2017, McKinsey.com.2019Four success factors for workplace automationExhibit 6 of 6Assistants like Alexa are getting smarter over time as more “skills” are introduced,driving platform stickiness. Alexa learns quickly# of 3rd-party skills, 000sExamples of Alexa’s skillsShoppingInformation and entertainment 0Q1 Q2 Q3 Q4 Q1 Q213510151825Q3 Q42016 2016 2016 2016 2017 2017 2017 2017“Add oranges to the grocery list”OurGroceries“Ask Dominos to track my order”Dominos“Tell Seamless I’m hungry”Seamless“Tell Starbucks to start my order”Starbucks“Activate climate control”BMW“Where is the nearest gas station?”Car navigation“Play the book I was just listeningto at home”Audible“Start my car”Ford Sync“What’s the weather?”“What’s the news?”“Play Ed Sheeran”“Set a timer”“Turn down the temperature”Nest“Preheat the oven to 350°”GE“Start my coŠee”Mr. Co„ee“Lock my doors”VivintDriving Home controlExhibitThe future of personalization—and how to get ready for it 3promotions for music or movies, for example, thatmatch that mood.Brands will use ecosystems to personalizejourneys end-to-endDifferent providers jointly own the customerexperience. A mall operator, retail store, and brandproduct, for instance, all contribute to a shopper’sbuying experience. But each sees and affects onlya portion of the total buying experience. Creatingconnections between those points represents abig opportunity in the next level of personalization,as expanding partner ecosystems allow brandsto provide more seamless and consistentconsumer experiences across all stages of theirdecision journeys. As AI gets better at predictingconsumer needs—turning on the lights or turningup the heat shortly before someone comeshome—personalization programs can navigate thetransition from one system (car) to the next (homelights or home furnace).While the share of global sales that move throughthe ecosystems is still less than 10 percent, wepredict it will grow to nearly 30 percent by 2025.Perhaps the biggest frontier for consumerecosystem development is the home. As devicesproliferate, they will need to work with each otherand use platform standards. Consumer goods,home-mechanics systems, automobiles, and avast array of digitized devices will need to be partof a seamless experience for the consumer or riskbeing completely shut out.Industries as diverse as banking, healthcare, andretail are also forging ecosystems comprising avariety of businesses from different sectors toimprove customer service and expand the

qualityand array of solutions offered.How to turn the future into realityPersonalization tends to be thought of as animportant capability for marketing, but we believeit must become the core driver of how companiesdo marketing. Here’s where successful brandsneed to focus now:Invest in customer data and analytics foundations:Personalization is impossible if marketers don’thave the means to understand the needs ofhigh-value customers on an ongoing basis. Sotop marketers are developing systems that canpool and analyze structured and unstructureddata, algorithms that can identify behavioralpatterns and customer propensity, and analysiscapabilities to feed that information into easy-touse dashboards.Setting up a centralized customer-data platform(CDP) to unify paid and owned data from acrosschannels is essential to these efforts. Unliketraditional CRM solutions, CDPs providebuilt-in machine-learning automation that cancleanse internal and external data, connect asingle customer across devices, cookies, andad networks, and enable real-time campaignexecution across touchpoints and channels. Thebest ones are also easy to use.Individualizing outreach across channels alsorequires companies to develop and interactwith new sorts of data, from voice to visual. Thebest are already actively experimenting withthese technologies by developing use cases tounderstand how to best use them.Making this technological leap forward requiresmarketing and IT to join forces. A productmanagement team, with representation fromboth IT and marketing, should be established tobuild and refresh the organization’s martech roadmap, develop use cases, track pilot performance,and compile a robust library of standards andlessons learned. Martech engineering shoulddeliver needed capabilities to the team, includingcybersecurity systems able to keep pace with theexpansion of personalized experiences.Find and train translators and advanced techtalent: Personalizing spaces, moments, andecosystems will require very different skill setsfrom those of the traditional marketing operationtoday. In addition to data scientists and engineers,marketing organizations will need analytics4 The future of personalization—and how to get ready for itThe future of personalization—and how to get ready for it 5translators who can communicate business goalsto tech stakeholders and data-driven outcomesto the business. As data become more complexand personalization use cases more advanced,the need for these translators will intensify.We predict a significant shortage of analyticstranslators. Today, only 10 percent of US graduatesin business and STEM fields (science, technology,engineering, and mathematics) go into businesstranslator roles. By 2025, we estimate that figurewill rise to 20 to 40 percent to meet the expecteddemand. Needless to say, the ability to recruit anddevelop translator talent will provide a significantcompetitive advantage in developing cutting-edgepersonalization capabilities.²Not surprisingly, the battle for AI talent willescalate. While organizations will need to figure outwhich talent to hire (those with core capabilities todrive creative problem solving) and which to accessthrough an ecosystem of external talent, leadingcompanies are moving aggressively to acquirerelevant talent. Some 52 percent of the mostdigitized companies build AI capabilities in-house,compared with just 38 percent of other companies,according to a recent McKinsey survey.³The buildup of relevant tech talent will need tobe matched by improved training so that peoplethroughout the organization understand not justhow to use new personalization tools but also howthey can help them make better decisions. Thatmeans training call-center agents, sales teams,and marketers, for example, in how to use emotionand sentiment-analysis systems to better supportcustomers. Rotation programs, in which peopletrain in a variety of teams, will be particularlyimportant to help spread knowledge and deepenorganizational proficiency in key areas.Build up agile capabilities: Given the iterative,cross-disciplinary nature of personalization efforts,traditional siloed marketing teams won’t work.Instead, personalization requires a commitmentto agile management, including cross-functionalteams dedicated to specific customer segmentsor journeys with the ability to execute rapidly.Performance measurement must evolve similarlyand become more focused on testing velocity,success rates, and creative boldness.In addition to demonstrating needed expertise,the ability to collaborate and integrate informationwill be key to professional advancement. Equallyimportant will be the ability to collaborate andsolve problems with colleagues from across theorganization, in IT, analytics, product development,and legal.Annual budgeting and strategy processes mustalso become more flexible with frequent reviewsto assess current initiatives, chart new ones, andrealign funding and resources in support of keypriorities.Protect customer privacy: The nature of dataprivacy has rightly become a source of concernfor consumers and brands alike. In the wake ofrepeated data-privacy scandals, 46 percentof customers want increased governmentalregulation to protect data privacy.? Companies atthe cutting edge of personalization innovation aremore likely to—rightfully or not—trigger privacyconcerns among the customer base. As such,proactively managing customer privacy will beespecially important.Organizations need to go out of their way to makeclear that they take data privacy seriously, bybeing transparent about how data will be used,limiting processing of personal data to what isnecessary, protecting data against theft, andgranting customers the right to be forgotten. Datagovernance and privacy policy considerationsmust become part of a marketing organization’sstandard operating model rather than an incidental2“The Age of Analytics: Competing in a data-driven world,” McKinsey Global Institute, December 2016.³ Ibid.4April 2018 Reuters/Ipsos poll of 2,237 people in the US; Gallup poll of 1,509 individuals in the U.S. between April 2-8.6 The future of personalization—and how to get ready for itmanagement issue. Ongoing training on legitimatefirst-party data-collection methods and usesof data, including sharing data with partnerplatforms, are required to make this happen.Privacy issues lie at the heart of the trust betweencustomer and brand. Companies need to get itright.Get started. Developing a personalizationcapability is a journey to get to the full suiteof capabilities needed for true dynamicpersonalization: always-on, real-time, one-to-onecommunication across the consumer ecosystem.Often, the hardest part is just getting started. Butin our experience, most companies have morethan enough data and people to get value rightaway. The first step is to determine which usecases to focus on (converting new customers,increasing spend of loyal customers, etc.) and putan agile team on each of them to rapidly test andlearn which offers and interactions best deliver.This hands-on approach quickly helps peopledevelop real experience and expertise. Even asteams move quickly, however, it’s important tokeep one eye on the end state so that you caneffectively plan in order to make better decisionsabout people, technology, and investment.Personalization is becoming more pervasive.Only by acting today, however, can companieshope to be in a position to deliver value to boththeir customers and their brands.Designed by Global Editorial ServicesCopyright © 2019 McKinsey & Company. All rights reserved.Julien Boudet is a partner in McKinsey’s Los Angeles/Southern California office; Brian Gregg is a senior partner in the SanFrancisco office, where Kathryn Rathje and Eli Stein are associate partners; Kai Vollhardt is a partner in the Munich office.The authors wish to thank Jessica Zhang for her contributions to this article.The future of personalization—and how to get ready for it 7