Mark Tack, vice president of marketing for Vibes, a mobile technology marketing company, put his point bluntly: “For us personalization isn’t a nice-to-have in the mobile channel,” he said. “It is absolutely required.” Vibes’ clients include some of the biggest retailers in the business—Home Depot, Sears, Men’s Wearhouse, Gap, Old Navy, Pep Boys. In mid-2015 he estimated that 90 percent or more of the work he does for those chains centers on the physical store. He said that personalization is “definitely” a topic of discussion among his clients, and that the smartphone “has become arguably the most important marketing channel.” Consumers are “living on their phones,” he said, and experience a “panic attack” if they forget them at home. “Our phones actually become extensions of ourselves,” he continued, and that level of personal connection means that people expect everything on their phones to relate to them individually. “Consumers are much more likely to turn off marketers from their phones than they are other channels,” he said. “Consumers’ tolerance for nonpersonalized messages on their phones is extremely low.” In other words, he said, a retailer that wants to attract the attention of smartphone users must “personalize and tailor every mobile message to the individual.”
Tack’s perceptions were closely echoed by each of the twenty-one executives I interviewed on this topic. Mark Miller, strategy director for the Catapult brand-marketing consultancy, said, “Locating individuals is a Holy Grail. Speaking one to one is a Holy Grail. . . . It’s all about one-to-one communication.” But, he asked, “Can we always get there?” He flatly answered his own question: “No.” Most of the other business leaders made the same point when describing their attempts to link technologies to data with the aim of tailoring messages that would get individuals to take certain actions near, around, or in physical stores. The words “experimental” and “testing” were mentioned often. “That stuff is really, really nascent; it’s highly experimental at this point,” said Gary Stokes of the Signal Digital consultancy when discussing his clients’ approaches to in-store tracking. Randy Jiusto of the Outsell strategic advisory firm agreed, noting that retailers have been testing beacons in certain stores as they attempt to determine “the best combination of the right services to offer, and how it relates to their analytics.” Jason Goldberg, an executive at the Razorfish marketing communications agency, noted an “appetite on the part of retailers to say, ‘Let’s do some tests to try out some innovative self-service experiments using the customer’s phone.’” Bruce Biegel, managing partner of the Winterberry strategy consultancy, pointed to the broad uncertainty standing in the way of the personalization grail. “They are testing,” he exclaimed, referring to physical retailers and their digital-marketing agencies. “They are testing everything.”
To Biegel, the advantage of retailers in the current competitive environment would devolve to those who have “the most experience in CRM [customer relationship management] data management, data aggregation, campaign management, and analytics” because they can use that experience to “deploy a new technology, develop testing criteria, and roll it out faster than those that don’t have the experience or the culture.” He saw Macy’s as a frontrunner. “I believe that Macy’s, because they have the analytics bench, because they’ve got the database and maturity in their organization, and because they have analytics talent, they are better positioned to continuously test. And any marketer that we’ve worked with who’s really, really good and really analytically driven, they will use anywhere from 15 to 30 percent of budget for testing. They’re trying to beat that control all the time. And so a Macy’s should always be testing. The whole job is to beat the control.”
Razorfish’s Goldberg agreed that Macy’s has implemented lots of testing to encourage optimal customer in-store involvement as well as to gather data toward understanding shopper patterns and personalization. He noted that the retailer was testing a number of loyalty programs, including Plenti, a trading stamp–like loyalty plan from American Express. He said that Macy’s was one of the first to enable customers to scan QR codes with their smartphones to obtain information about specific products. It was also one of the first to invest in mobile visual search, which enables a shopper to submit a photograph of an item to the Macy’s mobile app to find out if that product or a similar one is available in the store. “They are doing the beacon stuff,” Goldberg said. “They are doing GPS [tracking]. They are . . . partnering with shopkick. They’re doing some in-store analytics . . . to understand what the customer flow is, how frequently the same customers come back, and what the conversion rates are. They have a whole host of initiatives like that.”
But one executive, who asked for anonymity, was not nearly as enthusiastic. He gently scoffed at the image Macy’s promotes in the trade press and at retailing industry events. “Macy’s is a very good tech company,” he emphasized. “They are willing to try any ideas or concepts that come into vogue. But they are far less organic about it than they would have you believe.” He described the retailer’s activities as fragmented and noted that even though Macy’s hired “the first omnichannel officer in all of retail,” in actuality it’s “really an honorary position with no specific responsibilities and no real leverage within the company.” He said that the retailer assigns “a pizza box full of employees—you know, six employees”—to pursue disconnected company initiatives and “launch them all off and learn from each in their own little vacuum.”
This same assessment could be applied to many physical retailers in the mid-2010s. Despite a generally agreed-upon understanding of personalization and its relation to segmentation, there currently exist quite different perspectives on how best to use customized messages to reach shoppers. Moreover, the industry is struggling with key questions about what should be conveyed to shoppers via the smartphone: should the messages be mostly tailored salutations welcoming them to the store they’ve just entered? Should the communications mostly focus on pointing out certain products that fit a targeted customer’s profile? Should they present price adjustments aimed at offering just the right discount that entices a particular shopper to buy? Should other subjects—for example, loyalty, personal hobbies, provocative food choices—be included in the messages?
One aspect seemed clear: the data customers gave up, either knowingly or unknowingly, contributed to a data fest of experimentation. These trials created profiles of individuals that, when wedded to new targeting technologies, pushed the impulse toward discrimination to high gear. Shoppers might not be able to quite understand the hidden curriculum that urged them to give up information through various forms of bait. But they would increasingly feel how it would affect their movement through the twenty-first-century marketplace.
Joe Stanhope of Signal Digital mirrored the comments of most of those interviewed when he defined personalization as a “kind of data management where we collect data from lots of online and offline sources and then bring the data together for a single identity, and then make that data available to retailers many times.” For most of his clients, he added, personalization also involves creating a constant experience no matter the means a retailer uses. “So what I [as a customer] see in my laptop on my tablet and in my mobile, in the store, when I get mailings from the loyalty program there is some degree of continuity and consistency.” While not disagreeing with Stanhope’s broad description of personalization, Jason Goldberg of Razorfish believes that the concept is too vague to dictate specific activities that will solve specific retailing problems. He said merchants using the term are instead just reflecting panic over a new environment rather than an actual common understanding of how to carry it out. “If you interview certain retail execs,” he said, “[personalization] always shows up as the number one or number two point on the road map—‘Oh my God, I need more personalization online and in the store.’” But, he argued, “what that aspiration is varies widely from retailer to retailer. So for one retailer personalization is automated product recommendations, and for another it equals a 360-degree view of the customer with client-selling tools for their sales associates. And for another, personalization exclusively means customized pricing promotion. So you can look at the studies and say, ‘Oh my God, personalization is hot,’ but the word is so amorphous that it almost isn’t a useful data point to talk about personalization in the abstract. You can talk about individual tactics, and some of them have a more demonstrable return on investment than others. But it’s difficult to just talk generically about personalization.”
These two descriptions reflect the complexity involved in attempting to define and then institute personalization as a technique for reaching customers persuasively. For one thing, the word has charismatic charm within the retail industry; if done well, many see it as a key to meeting retailing’s marketing challenges in the twenty-first century. For another, the word is kaleidoscopic; the meanings retailing executives and industry consultants attach to it vary depending on their business needs. When it comes to enacting personalization, some executives focus on store aisles and checkout areas as the ideal locations for sending messages, while others emphasize means far broader than the physical store. Company reports, industry trade-paper discussions, and executives interviewed for this book go back and forth between these approaches, depending on the topic at hand. One aspect remains clear through it all: new norms are coming into being regarding the ways customers are viewed in the new shopping world.
The Neustar analytics consultancy tried to canonize some of those norms in a 2014 report titled An A–Z Glossary of Personalized Marketing. Noting that “many marketers are struggling” to accomplish “personalized digital marketing,” Neustar said the business “has grown so quickly that it’s developed its own vernacular.” It added, “This guide has been created to give you a better understanding of the components of scaled, personalized marketing, so that you can give your customers the one-to-one dialogue they’re waiting for.”1 Significantly, only four of the forty-three glossary entries mention physical retail in their descriptions. And the report has no entries for beacon, proximity (NFC) chip, lock screen, clientelling, geolocation, and a host of other terms crucial to understanding personalization in the physical retailing context. Still, Neustar made it clear it saw the physical store as a crucial part of the new personalization ecosystem.
The Janrain retailing consultancy took a similarly broad view of personalization. In 2014 and 2015 reports, the consultancy nested its understanding of the term within a model of marketing continuity and identity-driven marketing. It described marketing continuity as “a framework for thinking about and planning for an increasingly fragmented marketing landscape.” Identity-driven marketing, in turn, involves creating a “comprehensive, centralized and universally-accessible customer identity that informs every marketing message or offer an individual receives.”2 This continually refreshed “identity management system should become the driver of segmented and personalized messages.”3 Janrain defined segmentation as “a way to get a ‘compelling message’ in front of a smaller or more targeted group of customers who share similar characteristics.” It was, the consultancy said, a pragmatic waystation between the supposedly earlier “spray and pray method of marketing” and highly specific messages based on detailed knowledge of an individual.4 Recall from Chapter 5 Janrain consultancy’s proposition “Not all customers are created equal, and not all segments are, either.” A report from the firm advised marketers to divide target audiences into groups based on psychographic as well as demographic information. “Beliefs, values and attitudes are far superior predictors of response, affinity and loyalty [than] the typical age, gender and location data on which many marketers rely,” it asserted.5
As useful as discrimination through segmentation can be, Janrain said, it is actually a relatively weak form of personalization. The stronger type, “truly one-to-one marketing . . . goes deeper to enable customized experiences based on individual-level data, such as self-declared affinities and interests or past transaction history.”6 True personalization is “perhaps the most difficult identity-driven marketing tactic to put into practice.”7 It requires “real-time responsiveness and relies on highly accurate and comprehensive customer data.” Because of the difficulty of collecting the identity-linked information and setting up technology capable of analyzing and applying that information to messages in real time, “Most marketing organizations have really yet to implement systems and processes that enable true one-to-one marketing experiences.” Nevertheless, Janrain reported, “a majority of marketers believed [in a 2014 survey] that personalization would become the most important capability for their teams in coming years.” The ones that do it early, it claimed, “will be far ahead of the competition in terms of differentiation and added value.”8
Joe Stanhope emphasized an additional continuity between pursuing narrow segments of customers and specific individuals. “A segment,” he said, “is really any kind of rule that’s applied to the universe of customers.” It might be based on the day’s weather, or on whether the merchant had designated a shopper a gold or a platinum customer—or a lapsed one. An “awful lot” of his clients also “score” shoppers; they place them in categories that indicate their value to the store or the store’s loyalty program. The distance between segmentation and personalization can sometimes be quite slim. Asked, for example, whether “scoring” the value of segments typically affects the messages his clients send customers, Stanhope slid into a perspective that sounded very much like personalization. “It should affect them,” he began, and then added that “given what retailers know about their customers, if they have a way to authenticate a customer through a loyalty program or something like that, they should be able to get pretty granular because they know your purchase history, they know where you live, they know your interests, they’ve probably enhanced that data with third-party information already. They might’ve scored it to figure out what you’re going to want next.”
Andy Chu of Sears elaborated on Stanhope’s points as he described his company’s classification of its shoppers for its Shop Your Way loyalty program. Sears divides customers into various segments, and then sends specific customers within each segment individualized messages depending on the person’s profile and value score. The initial, broadest segments primarily reflect shopping frequency and the “way they engage with us across channels.” Shopping frequency is determined according to “someone who shopped with us over a year ago, someone who shopped with us in the last six months [but not recently], the frequent shoppers.” Sears then further segments these “loyalty” cohorts based on the amount spent during each visit (their “average basket size”) and by demographic information—income levels, ethnicity, home address, and the number of people in the household, for example.
Gathering that kind of information about a family, he noted, requires a way to identify a particular household—which the loyalty program provides when it asks for names and addresses. Sears also buys additional information from a list of companies including data provider giants Acxiom and Experian. Then, “we really mash those [pieces of information] together, if you will, to get a sense of who the customers are.” As an example he pre-sents “a Hispanic household that falls in the age group of thirty-five, has three kids, and income level of more than $75,000, they shop with us more than three times a year, and their basket size is XYZ dollars. So we have different attributes and then we look at different sources that those attributes potentially come from.” Of particular interest are “high-value customer” groups within the segments. Chu calls one such subsegment Four-Plus Trippers. “Those are customers who shop with us very frequently, and within those we look at how much money these people also spend with us.” Based on its statistical modeling, Sears places a certain value on one or another subsegment. Then, based on what it knows about the group’s purchasing habits in terms of products and the channels it uses for retail purchases (the Web, the phone, the store), the company will approach individual customers with particular messages and offers tuned to their group. “If we determine you have an interest in the tools category and the last time you bought tools is three or four months ago, and we also saw you looking online at lawn and garden stuff, we might push you a notification” to encourage a tools purchase. The communication could be a text message or an email, he said, depending on what statistically had worked best to lead people like “you” in the segment to purchase the item.
While such increasingly narrow segments come close to one-to-one personalization, Ryan Bonificino, the marketing head for the Alex and Ani jewelry business, asserted his company actually carries out the real thing. He said his firm’s one-to-one involvements are with those who already have bought from the chain in the physical store or online and whose identities the analytics team therefore often knows. The key to successful personalization, he said, involves optimizing the use of the company’s data management platform. It contains information about all the firm’s identified purchasers, including “every single digital touch point.” He added that the firm has an average of six hundred strands of information on each person and “upwards of twenty-eight thousand data points so far on our oldest customers.” That includes bits about activities on desktop computers, laptops, and mobile devices.
Identifying high-value customers is an important part of the customer analysis for future tailored interactions based on those data. “We’ll look at a couple of different dimensions,” he noted. “One is obviously the length of time they have been connected to us as a consumer. That could be as simple as signing up for email or as particular as purchasing a product online versus in-store. Making a new account, or maybe phoning [a] call center. All those little time stamps when they start their relationship with us. So that’s more of a lifetime-value play. Then we have highest spenders, most frequent spenders, and we actually do most influential as well.” Influential people are those whom the company has tagged as having not only many social contacts but also sway in relation to those connections. The company tracks shopper comments through a social listening product called Radian6, which traces millions of discussions about brands in the digital environment. Bonificino said Radian6’s parent company can often help his company’s email provider match email addresses with commenting individuals. This information enables Alex and Ani to link specific opinions to individuals on the firm’s customer list as well as to note the discussion topics and their frequency.
To assess the influence of those customers on others, Alex and Ani turns to social sign-on data, which is obtained from people who log into various websites using their password from Facebook, Twitter, Google+, or other social media sites. The advantage is convenience, as the individuals can use the same username and password instead of having to create a sign-in for each site. However, they must typically accept (through a pop-up window) that the site will receive data about them and the “friends” they have on the social media site they use for logging in. Retailers often hire a technology firm (such as Janrain and Gigya) to help implement the sign-on as well as to collect and analyze the data coming from the social media sites. What they learn can be well worth the effort. “It will be the email, the networks of friends, things that they ‘like’—and you can pull in a lot of other stuff,” said Bonificino. The more people in the target person’s network, the more likely that person will be deemed influential, especially regarding discussions captured by Radian6. All this information determines the influence score Alex and Ani gives its target customers and, in turn, the treatment they receive from the retailer.
Alex and Ani uses the data it can gather on an individual to tailor what the person sees on its website and app. When the firm has enough information to form a profile—usually after about a month once an individual has begun visiting the site—said Bonificino, “we start developing some logic around who you are and what you want to see. What keeps you on the site longer? What makes you share more on social, that kind of stuff?” As the retailer’s profile of a customer grows, what the shopper sees online and on the app becomes increasingly personalized. For example, the firm may show different people different photos of the same product, and it may present them with different product details. Personalization is not quite as focused in the case of advertising messages online or via mobile devices. “We don’t use all six hundred data points while personalizing” the ads, Bonificino said. In fact, he said, when pursuing customers for the jewelry charms it sells through partnerships with Disney, Team USA, or the National Football League, some of the personalization comes closer to segmentation. Many shoppers are unaware of these connections, and they may never even have shopped at Alex and Ani. The company has been attempting to identify people who are fans of Major League Baseball’s Philadelphia Phillies and who had purchased charms from the team or might buy them in the future. Bonificino said he has access to the Phillies’ records of charm purchases, and he also turns to data suppliers such as Datalogix, eXelate, BlueKai, and MasterCard Advisors to learn what Alex and Ani customers purchase from other merchants (both online and by other means) in order to identify shoppers who have a predisposition for Phillies-related items.
Industry consultants stress that the success of personalization is limited by the quality of the data. To them this means accuracy and relevance, and they argue shoppers want the same when they receive personalized offers—hence the widespread and continual collection of data about individuals from as many sources as can be identified. Consultants consider anything less than that unacceptable. A 2014 Janrain poll, for example, found that 71 percent of U.S. consumers and 61 percent of European Union consumers say they have received offers which “clearly show they [marketers] do not know who I am.” The poll also found that 51 percent of U.S. consumers and 36 percent of European Union consumers have received inconsistent information among various channels—online, apps, mobile devices, in-store—from the same brand.9 “Less than perfectly accurate” data was the source of much of this problem, according to the company.
To avoid such issues, industry analysts advise retailers to focus on amassing just first- and second-party data. First-party data comes from the retailer itself, while second-party data is obtained directly from other retailers or publishers (Facebook or Google social log-ins, for example). In one of its reports Janrain noted that third-party data (from firms such as Datalogix and Acxiom) may not be current, as an individual’s life events typically unfold so fast that it’s impossible to keep pace with them.10 Consultants suggest that regular triangulation of first-, second-, and third-party data is optimum. They say that to ensure the most accurate and actionable customer profiles, marketers should use multiple sources and compare “declared, first-party information directly from the consumer or identity provider, against purchased customer data.” Along with other analytics firms, Janrain recommends continually tracking individuals across as many platforms as possible. It suggests to retailers that they should engage shoppers when they visit the merchants’ websites or access their apps by posing questions, encourage shoppers to participate in a poll, and have their in-store cashiers ask customers for their email address and phone number. This “progressive profiling,” Janrain promised, would yield increasingly specific information about individual customers over time.11
Janrain was among the analytic and consultancy firms that wrote only in passing about brick-and-mortar venues as places for collecting and using data for message personalization. Despite their omnichannel rhetoric they had become comfortable with the idea of gleaning data for personalization from purely digital areas, especially websites. With the growth of mobile devices, though, an increasing number of consultants and agencies, along with retailers, emphasized the use of data for tailoring messages in the physical store. “Offline Personalization Matters Just as Much,” was the headline of an early 2015 marketing industry article. “There’s been plenty of talk about personalizing online communications,” the article noted. “But recent research suggests it’s equally important to tailor offline customer experiences as well.” A survey cited in the article found that “just 23% of client-side marketers worldwide personalized offline channels, compared with 88% who used email personalization and 44% who did so for websites. Agency professionals were even less likely to tailor offline efforts, at just 17%.”12 For eMarketer, the survey findings were “another reminder” of the need for retailers to provide a genuinely multifaceted experience. “Most final buying decisions still happen in-store,” it pointed out. “Retailers who can tie all the data collected on a customer [across all channels, while the customer is in the physical store] stand a better chance at closing the deal.”
Some interpreted the call for in-store personalization to mean bringing back the traditional one-to-one contact between merchant and customer. The publisher of online publication Furniture Today, for example, exhorted her readers not to “underestimate the power of connecting with customers within your very own four walls.” Arguing to “Let Amazon be Amazon,” she said local retailers should not overlook their biggest strength: the opportunity to connect with their shoppers and to remember that people are social by nature. Her suggestions: interpersonal conversations “that genuinely engage with customers about their day,” and to provide exceptional customer service and to inspire shoppers by creating a visual counterpart to social sites such as Pinterest—but adding smell and sound. “Making the shopping experience inspirational is a winning combination with shoppers; many prefer to have the touchy, feely experience. Again, local retail wins.”13
Others argued that by combining quality data with appropriate in-store technologies retailers could eventually accomplish the same goals—only more efficiently, at greater scale, and perhaps even more successfully. “Marketers Are Showing Customers Love,” read an April 2015 headline from the trade magazine DM News. The article said the “love” emanated from accurate, customer-relevant data leading to personalization.14 While DM News wrote broadly about tracing “multi-touch customer journeys” as a source for personalization, a Forrester Research report in 2014 heralded “new key metrics that retail store analytics will unleash” for “insights that will drive increased engagement with customers as well as greater efficiency in retail store operations.” The report went on to say that “retail stores have been living in the analytical ‘dark ages’ in comparison to digital channels,” and that recently “technologies such as Wi-Fi, Bluetooth, GPS, and video” have allowed retail store analytics to gain a foothold among digital retail businesses, although “the depth of these analytic tools is lacking and immature.”15
Forrester also noted that physical stores were beginning to view personalization in and around the physical sales arenas on par with that for the Web. Retailers’ experiences with this varied; not only were they targeting different audiences, but each consultancy and marketing agency had its own philosophy and so glommed onto different technologies to carry out the targeting goals. The mobile technology company Thinknear, which approaches shoppers between home and store, hyperbolizes on its website that its technology is “so accurate, it’s insane. . . . We’ve combined location technology, a deep knowledge of consumer behavior, and first-party data from our parent company, Telenav, to give you the ability of a mind reader—or some crazy shaman—so that you find your audience and finally deliver those mind-blowing ideas.” Telenav, a wireless location-based services company, offers an app featuring maps, navigation support, points of interest information, real-time commute times, and weather forecasts. The app is free; the company makes money capturing, storing, and analyzing information based on individuals’ use of the app so they can be targeted with commercial messages. According to Daniel Mahl of Thinknear, a special benefit of the app for his firm is that it continually updates latitudes and longitudes of restaurants, stores, train stations, airports, and other points of interest. Thinknear uses those locations in two ways. It can target people who are in certain areas, and it can target people who have been in those places.
Say a client wants to reach travelers who are near its stores. Thinknear goes after those types of people in two stages. First, Thinknear computers link to mobile-advertising exchanges that auction individuals based on their mobile phones’ current latitude and longitude. Thinknear bids on locations frequented by the types of people its clients might prefer—John F. Kennedy International Airport, for example, to reach “travelers.” If Thinknear wins the bid, it can serve its client’s ads to the particular phones at the airport. Just as important for Thinknear at this stage are the data that the company gleans from the phone receiving the ad. If the ad lands on a mobile browser, Thinknear includes what’s called a tracking pixel, which can record the Web page on which the client’s message was presented. Tracking pixels have less utility on phone apps, but in these instances the company can record the personal information people provided when they registered for the app—gender, age, and home address, for example. The most important value of this targeting stage to Thinknear is that it provides the company with the unique identification numbers of the devices it targets. Possessing those numbers enables Thinknear to track the individuals via GPS, cellular triangulation, or Wi-Fi as they move across the landscape mapped by Telenav. The company then executes its second stage of reaching “travelers” via mobile devices. It goes back to the mobile exchanges and has them find the travelers’ phone IDs among the phones they put up for bids. Thinknear then bids to serve ads to those phones, either for the initial retailers who reached the travelers at the airport, or for other retailers who want to reach travelers wherever they may be.
Thinknear’s specialty is determining the devices’ real-time locations and targeting them for clients who want to reach people in those specific areas. A typical geofence (meaning a defined geographic area) around a retail establishment, Mahl notes, is about a quarter- to a half-mile. Generally such a geofence is wider in the suburbs than in a city to account for cars being a more predominant mode of transportation in the former. The company uses the phones’ location to interact with people when they are in or near a competitor of its client in the hopes of steering the people toward Thinknear’s client instead. This can involve inducing the phone’s owners to click a button that will take them to a landing page that, in turn, might offer a simple click to call the advertiser as well as directions for the client’s store.
Mahl says this landing-page technique “works well,” though not necessarily better than simply sending ads for his clients. Although the company can target men or women separately as well as both sexes together, he preferred instead to discuss Thinknear’s focus on an individual’s location as the pivot around which all persuasive activities turn. The company’s software can continually remind the phone user how far the individual is from the advertised retailer. It also can change the message based on the weather, pollen count, or even UV index in the phone holder’s location. “I’m working on a campaign right now [for sunscreen] that we’re only serving ads to people when the UV index is 6-plus,” Mahl said. He went on to say that Thinknear is working with a location-analytics firm called Placed to determine whether the app-directed mobile device display messages lead to sales even though his firm is not involved in reaching shoppers in the store. He also said that Thinknear wants to be able to match behavior on the Web and other media to the owners of the phones they target, but the company has yet to fully achieve that goal. “It’s definitely getting better,” he said. Thinknear’s ability to identify individuals they can pinpoint geographically has increased, as has the firm’s ability to gather data about those individuals from various sources. Yet, he added, “it’s still a ways away from being the industry standard and even then it’s still tough to get it down to an exact science and prove that it’s completely correct.”
In contrast, the Vibes agency downplays the value of delivering ads on mobile browsers and apps. “When you think of mobile messaging and mobile marketing and communication on the mobile phone,” said vice president of marketing Tack, “we like to recommend [that] the three biggies are text messaging, ‘push,’ and mobile wallet.” Text messaging needs no special technology, but federal law requires that marketers offer a user the choice to opt in to receive these messages. Consequently, Tack said, because the users choose to receive such messages, they likely pay more attention to them than they would to random email messages. Moreover, he continued, those who opt in to receive texts likely are more loyal, and therefore more valuable, to the retailer. “Push” messages and the mobile wallet help ensure that these messages do not go unnoticed, Tack said. By push he was referring to messages that pop in front of the targeted person on the phone’s top screen—its lock screen—at a preset geographical location. The push notification is tied to the presence of the client’s app on the phone. Many of their customers download those apps but don’t use them. Push-notification technology exploits the app’s presence without relying on the shopper to do anything with it. It is activated by cellular technology, Wi-Fi, or store beacons based on the shopper’s location. If a shopper doesn’t have the retailer’s app (and many do not) there is another way to push a message to the person in a particular location: send a coupon as a text message and encourage the person to save it in the phone’s wallet. By placing location tags on the coupons, marketers can instruct the wallet to send reminders about them to the device’s lock screen when the person is in a particular place—a specific aisle in a specific store, for example.
Tack emphasized that these messages must be personalized if they are to be successful. “If you’re sending a text message to someone and you just do a spam one-size-fits-all, you’re going to have a pretty low response rate,” he said. “What a lot of our more sophisticated clients do—like a Sears, a Home Depot, a Pep Boys, a Men’s Wearhouse, as a few examples—is they’ll integrate their mobile database with their CRM [customer relationship management] systems” and their loyalty information. The customers are then placed into segments and sent messages “that include the customer’s name, category of products that they prefer to buy or they have purchased in the past—offers that have proven to be effective for that particular customer.” He continued, “In the mobile marketing channel, for retailers to be effective in what they care about most, which is driving in-store foot traffic, increasing basket size, and improving loyalty, we have found that using a personalized approach just delivers significantly better results.”
Shoppers who come to Vibes’ clients via their smartphone’s Web browser rather than the app are assigned a lower priority, as the agency believes browser use means that shoppers are not yet sure what, and from whom, they want to buy. Tack said the company prefers to try to learn about people who are already “engaged with the retailer, with the app installed, texts requested, or coupons downloaded.” Such people merit personalization, he said, a task that’s feasible “because now you actually got the data.” His company advocates text messaging, push notification, and wallet-generated lock screen views as the preferred means by which to reach those shoppers.
inMarket has yet another take on the technologies retailers should use to pursue shoppers. The company is struggling to marry two businesses, outdoor geofencing and in-store beacon messaging. The separate operations frustrate inMarket executive Jeff Griffin. “I work with retailers, and retailers care about two things,” he said. “One is driving trips to their store. Geofencing does a really good job of driving trips to a store. . . . [But] it’s not very good for driving the second thing that retailers care about, which is conversion—making a sale. Because you can’t penetrate the building. Beacons are the thing that allow retailers and brands to do that. And when you’re in a store to buy something, you can receive that message, whatever that motivational message might be.” He gave an example: “Hillshire Farm, which is a really good customer of ours, they would deliver messages to buy Hillshire Farm sausages within one mile, for instance, of every Walmart in the country. And that message, frankly, can’t penetrate the inside of a Walmart. You only get it when you’re in the parking lot.”
To link the two businesses and expand inMarket’s marketing possibilities, the company has been experimenting with targeting shoppers outdoors based on its successes with the beacons it has placed inside stores. For example, as a result of ads the company has delivered for Hillshire Farm within supermarkets that carry its beacons, the “interaction with the brand” within those stores, as well as the number of shoppers installing and using the Hillshire app, increased substantially.16 The next logical step, Griffin said, is to note via beacons where people linger in a store, infer the products [Hillshire’s or others] they might be looking at, and pursue them through geofencing with ads about the items after they’ve left. This practice, he contends, will give the physical retailer the same power as Web merchants to personalize ads to shoppers beyond the store’s borders.
He offered the example of a man pausing to look at a lawn mower in a store. The beacon would register that he lingered for a substantial time, indicating an interest in the product. Griffin said the retailer should be able to send a message to the person about lawn mowers when he leaves the store, much as a person who viewed a mower on a website might find himself encountering lawn mower ads while browsing elsewhere on the Web. inMarket’s location-specific ad business could do just that, said Griffin: “Once that person leaves the establishment and gets out of the range of our beacons, the fact is we can at that moment begin to drive retargeted messages to that phone that have to do with [the product].” Yet he acknowledged in mid-2015 that “we don’t do that much yet.” Because inMarket is “only in a couple of thousand stores”—and on only a dozen apps—it has been challenged to gather sufficient information on what shoppers are doing inside stores to enable it to engage in large-scale retargeting outside stores. He also said that many of his retailing and advertising associates have yet to understand outdoor retargeting: “I sit down with retailers who’ve been in the business twenty, thirty years, and they say, ‘Get out—you can’t do that!’ And I say, ‘Yeah, absolutely you can. It’s not that hard.’”
The various challenges notwithstanding, inMarket has been able to connect at least some of its outdoor and instore activities. Using its beacon data to analyze the shopping cycles of the millions of shoppers who hold its app, the company concluded that it could determine when particular customers would be due for a return visit to a particular store and send ads to those people’s phones, especially when they are near the retailers, to entice them to stop in at the retailer. In late 2015 the company reported that sending such ads on behalf of an unnamed client resulted in an 8 percent increase in store visits and a 14 percent increase in the amount of money spent per visit over the period of a year. The online media and marketing publication MediaPost recognized the synergy: “While most consumers may never be aware of beacons,” it said, “they may find themselves receiving more ads that matter at the right time.”17
inMarket wasn’t alone in attempting to profit from sending personalized messages to shoppers as they move toward, through, and away from a physical store. By mid-2016, Swirl, another location tracking company, was working with retailers including Lord & Taylor, Urban Outfitters, and Alex and Ani to combine beacons with Wi-Fi and GPS tracking for anonymous indoor and outdoor tracking. Swirl could identify phones and follow them across different locations, drawing inferences about the users’ interests and sending them messages sponsored by one or another merchant. For example, reported MediaPost, “a retailer could send a message in a parking lot when within a certain number of feet of a store offering a deal at the store. When that shopper enters the store and signs on to Wi-Fi, another location point is captured. And when that shopper nears a beacon in a department, yet another data point is captured. All that data could be used to trigger an instant, target message or offer or simply filed away for later use.”18 “Later use” is a particularly attractive feature, MediaPost noted. Although Swirl’s basic tracking would be anonymous, it did tie the data to the phone ID. It would then hand the tracking information, along with linked IDs, to the retailers where the smartphone owners shopped. The retailers could match the phone IDs with personal data they had stored via loyalty programs and other means. Anonymity would dissolve, and a new trove of material to build ideas about whom to target, when, and why, would come into play.
Twitter was one of the companies investing in Swirl’s new enterprise. In 2015 the social media giant was itself working hard to show that tailored ads placed alongside its social messaging could be a boon to retailers as shoppers moved through various daily experiences. In earlier efforts to increase revenues, Twitter had integrated the ability to place commercial messages carrying photos, videos, and links to advertisers’ websites into the flow of a subscriber’s feed. In 2014 it incorporated a Buy Now button to enable mobile users to purchase an item directly from the merchant whose product was being advertised, sometimes even using a credit card number they had stored with Twitter, without leaving the social media app. Twitter considered the development so important that it designated staff to work directly with very large retailers, including Walmart, Macy’s, Best Buy, and Walgreens.
Twitter’s Patrick Moorhead was assigned to work with Walmart. Part of his job is to identify people on Twitter beyond the currently active Walmart followers on the social network. “I’m probably reaching twenty million consumers today on Twitter with messages about Walmart’s fresh groceries that have nothing to do with the six hundred thousand people who follow the Walmart account on Twitter,” Moorhead said. Twitter’s agreement with Walmart incorporates a revenue sharing model that encourages Twitter to work toward achieving sales for the retailer online, via mobile devices, and in the store. Twitter does not charge the retailer directly for ads; rather, the social messaging site receives a portion of the sales revenue either when a person purchases a Twitter-advertised product using the Twitter purchase button on the Web or the mobile app, or when a shopper who received Twitter’s Walmart ad for a product subsequently purchases that product in-store (database coordination between the two operations identifies the link). Moorhead sees the program mainly as a way to reach consumers outside the physical store rather than in it (or rather than those who might travel there because of the Twitter program). “It’s less about Twitter being present at the shelf at Walmart and more about creating the ability for Walmart to export the shelf to consumers wherever they are,” he said. At the same time, he noted, “over 70 percent of Twitter users use Twitter while they are shopping, looking for product information, often reading the conversation about products even if they don’t participate in the discussion.” Consequently, he acknowledged, the Walmart ads could spark purchases in the physical store, and he noted that Twitter has the ability to use location as a consideration when it targets individuals with ads.
Moorhead believes that Twitter is uniquely positioned to reach large numbers of individuals based on their stated interests and pursuits. “We have very robust data science organization within Twitter,” he said. “On Twitter, people tell us what they’re interested in and they tell us what they feel all the time when they tweet. So we have built technology behind the scenes to compile all that information and operationalize it as, ‘Hey, find me everyone that via their Twitter activity—you know, who they follow, what they read, and what they tweet about—[who] are highly likely to be interested in fresh, cheap groceries from Walmart, either for same-day delivery or in-store purchase.’” He said that Walmart and Twitter mesh well. “When we look at Walmart we know their strategic growth customer is the millennial mom—younger mothers with kids in a household.” He explained that Twitter can analyze people’s tweets to unearth all kinds of information about them, such as whom they follow and what they read. Armed with all these data, the company can then zero in on specific demographics according to personal interests. For example, he said, he can target those who indicate by their tweets that they are interested in looking for recipes, or those tweeting about healthy food for their kids.
Walmart, for its part, has a world-class customer tracking and analytics division it calls WMX. According to Moorhead, it knows which products huge numbers of specific Americans and American households put into their shopping baskets, and in which combinations. Walmart tracks its individual customer purchasing habits by connecting them via their use of credit cards at the physical store’s register as well as their habits on Walmart.com and its related app. In their collaboration with WMX, Moorhead said, Twitter’s software experts built an infrastructure to enable Walmart to find its customers in Twitter’s vast database of registered users. An outside matching service then takes the data from WMX and matches it to Twitter users based on email addresses and other user account information. Once these connections are made, Twitter can analyze specific populations and separate them further based on its own data on the individuals and their interests before targeting them with personalized messages. For example, Walmart can identify Twitter users who are millennial moms who shop at Walmart for soy milk and gluten-free bread and who may be good prospects for a certain yogurt. Twitter takes this information and conducts further analysis on this group: Moorhead suggested hypothetically that “they all tweet about whatever product Oprah gives away each week, they all tend to be interested in yoga, they all tend to be over thirty, they all tend to be interested in fashion, and they all like Taylor Swift.” Twitter uses all this information to target Walmart ads to Walmart customers as well as to non-Walmart customers with similar characteristics (for example, Twitter users who don’t shop at Walmart but who have profiles that parallel those who buy soy milk and gluten-free bread).
Soon, Moorhead went on, the Twitter-Walmart connection would extend into an additional venue: commercial television. Many people tweet about the programs they are watching, and Morehead said that Twitter can monitor that activity and correlate it with Walmart commercials airing on television. Once Twitter identifies the person tweeting about a television program during which a Walmart ad airs, that individual would likewise receive a Walmart ad on Twitter. “With the buy-now functionality on Twitter,” Morehead said, “we would essentially have the ability to know that Joe [is] sitting in his living room with his phone in his lap watching CSI on CBS, and at 6:19 p.m. Joe saw a Walmart television commercial featuring Weber grill tools, and then at 6:23 p.m. we would serve Joe a promoted tweet from Walmart [inviting] him to buy those tools on Twitter.” The seemingly personalized message could lead “Joe” to purchase the tools in a Walmart store—and the chain could track the transaction and pay Twitter for its involvement. “The whole scenario that I just painted for you . . . sounds like it’s an episode of [the surveillance-themed movie] Minority Report,” he exclaimed enthusiastically, “but that will be a reality by the end of November [2015].”
None of Walmart’s messages offers a discount. The merchant, Moorhead says, instead wants to underscore that it offers low prices as a matter of course. (“Rollbacks,” which may seem like discounts, are not the same thing: Walmart offers them when it persuades a supplier to provide a product at a lower wholesale price so the retailer can pass the reduced cost on to the shopper.) For many other retailers discounts are a part of life even though they’re not necessarily profitable. Executives acknowledge that they have been feeding Americans’ virtual addiction to the appearance of a “sale” for decades through a variety of means—especially coupons in the grocery, drug, and chain store arenas. Even brand manufacturers as powerful as Procter & Gamble haven’t been able to break the habit. The mobile phone hasn’t changed the situation, though it has introduced new challenges, such as the need to calibrate the number of messages, and the mix of different message types, that retailers and their consultants send to shoppers’ phones inside and outside the physical store. In addition, when retailers choose to use price as an incentive, they must decide whether it should be personalized or not—and if so, how.
Jason Goldberg of Razorfish is among those who urge caution regarding discounts in the digital mobile world. “I am not a huge fan of the ever-increasing ways to deliver discounts and promotions,” Goldberg said. “It’s mostly a race to the bottom; we’ve mostly taken all the margins out of the stores already. The only way that I can come up with a sustainable win is to develop unique shopping experiences that the customers value beyond price. And so I’m much more interested in figuring out ways to use this technology to figure out a differentiated shopping experience that customers actually appreciate and will seek out than I am just using it as another delivery vehicle for promotion.” Mark Miller, an executive vice president at Catapult Marketing, agreed. “I’d rather not lead with price,” he said, arguing instead for the traditional formula to convert shoppers to a brand: a great product, great service related to the product, and a competitive price that together ignite great word of mouth. Miller added that in the new digital retailing arena, helpful “content” or other “enriched” experiences connected to the product and its use is a stand-in for great service. One such example would be a clothing store’s app that offers ensemble suggestions from different departments based on the customer’s shopping history and an initial choice of items.
Abhi Dhar, Walgreen’s chief technology officer, noted that Walgreens does not guarantee the lowest prices, nor does it match prices offered by other stores. Instead, his company’s marketing efforts are aimed at getting customers to focus on the chain’s service message—convenience and a positive shopping experience—as opposed to prices. Personalized messages tailored to a person’s shopping interests and background can reinforce this objective, Dhar said. Miller says his company offers a more quantified way to take a shopper’s mind off price: an armamentarium of customer-analysis tools that Catapult has assembled for such clients as Mars Incorporated, Frontier Airlines, and Diaggio, so that it can decide whom to target and which messages to target them with. In particular he pointed to Catapult’s access to corporate parent Epsilon’s “massive database” of people’s purchases—“100 percent household penetration of data with thousands of attributes”—that provides demographic and psychographic traits along with individuals’ shopping journeys to enable Catapult to infer “some of those micro needles in a haystack, insights that will drive consumer conversions toward a brand.”
For example, Miller said that to better understand different types of feline owners on behalf of Mars (which manufactures cat food in addition to candy and many other products), Catapult analyzes customers’ entire shopping basket when they are buying cat food. He also pointed to the firm’s additional capabilities resulting from Epsilon’s purchase in 2014 of the digital-marketing agency Conversant, which continually buys data relating to the transactions of millions of individuals at more than four thousand retailers.19 The company contends these people are “anonymized” because it doesn’t record their names or addresses. Yet this anonymity may be irrelevant because Conversant has so much information about each person—“7000+ dimensions,” the company asserts—that by using statistical techniques the company claims 96 percent accuracy in identifying an individual’s presence “across display, video and mobile.” Conversant assigns every shopper a unique ID that is the basis for tracking “how they spend their time and money—online, in-store, or both.” The company claims that “nearly everyone in the US” has been assigned an ID, which remain in place for “a lifetime so your conversation with each consumer never misses a beat.”20 Catapult’s Miller noted that using Conversant to match his clients’ offline customers with their digital presence and follow them puts his work at the leading edge of “micro-targeting.” These state-of-the-art capabilities (presumably along with services from Epsilon’s Agility/Harmony email targeting company)21 enable him to answer the ultimate personalization question for his clients: “How do I deliver the right message to the right person across devices at the right time?”
Miller recognizes that retailers and brand manufacturers offer people coupons based on their characteristics and in-store behavior. He sees his role, however, as helping his clients maintain their profit margins by working with them to craft messages that would be so persuasive at the individual level that discount coupons become unnecessary. He says he tells his clients they shouldn’t react to shoppers leaving their website without purchasing by automatically sending coupons to them in the hopes that they’ll change their minds and buy something. “That’s not good business,” he insists. His goal instead is to help clients “crack that [discount] addiction” by arming them with knowledge regarding types of shoppers, and even individual shoppers, so that they can steer customers toward those “two top levers” that encourage purchasing: a great product and great customer service, “without diluting price.”
In contrast, Sears goes with the addiction flow. Andy Chu, who oversees mobile management for the retailer, explained that the company carries out strong price-promotion efforts aimed at the company’s Shop Your Way loyalty program members, who constitute approximately three-quarters of all Sears shoppers. Price promotions are often designed so that a shopper’s overall purchases at Sears add up to an acceptable profit margin from that individual. Both virtual and physical transactions are factored into this calculation. A major concern for Chu is that some Sears customers targeted for discounts visit the store only to buy sale items—a practice that is destructive to profit margins. “As an industry, I would say we do not do a good enough job to identify who are the good and the bad customers,” he said.
In the meantime, Sears continues its general efforts to identify the good customers and incentivize them to be better ones. “On a regular basis we send email [and text messages] to our customer segments,” Chu stated. “We also send them loyalty points and rewards points. And all that information is based on our internal modeling. So we know that if you have affinity for the tools category and the last time you bought something is three or four months ago, and we also see that online you are looking at lawn and garden stuff—we might push you a notification [that states,] ‘Hey, here are some surprise points. . . . Come spend XYZ dollars.’” He added that the offer can be geared toward fairly specific discounts, such as $5 off on apparel or items purchased in the lawn and garden department, or to more general ones, such as $10 off for spending $75 or more. “It really depends on the different modeling that we do,” he said.
Chu went on to say that loyalty members don’t get individual pricing, but they do receive discount offers via their mobile devices, adding that members may receive personalized reinforcement “at the points level. We might give a person additional points—effectively it’s a discount. You might get $5 and I might get $10.” As of 2015 these blandishments were showing up on the mobile browser and in the Sears app, but not in the physical store. Still, things are changing. For one, Chu said, prices in different Sears stores as well as online are adjusted constantly based on Sears’ price monitoring of its competition. For another, Sears has also been testing “e-ink” paper-like electronic shelf-price displays, which the retailer can change instantly depending on various competitive exigencies (Chu noted that Kroger and Whole Foods supermarkets have also been moving in this direction). Still another change is evident in salespeople’s interactions with shoppers in the physical aisles as a result of the retailer’s digital personalization efforts. “As an example,” he noted, “in the appliance area our associates . . . have been carrying iPads for about two years now. And when you are interested in a refrigerator, they will walk you through, they will show you on the iPad all the different specs, they will help you compare refrigerators. And you might say, ‘Hey, I’ve just been looking at Best Buy. . . . This product is similar, but I can get it for X price.’ And it’s up to the salesperson to say, ‘You know what, we can cut you a deal.’ If you identify yourself as a member, they can actually look at your level—how many points you have, all that information. . . . We have something called member pricing. And if you’re a member you get special discount.” Chu said that Sears isn’t yet able to arm its salespeople with loyalty member value-level information so they can offer individualized pricing but added that “we are starting to develop that capability.”
Chu acknowledged that price incentives could be wasteful if presented to customers who entered a store already having decided to purchase a particular item. The dilemma, he said, was determining shoppers’ exact situations so that Sears would know whether incentives were advisable and, if so, what those incentives should be. “All retailers struggle with this, including Sears,” he noted. “Are you leaving money on the table if someone’s willing to buy at the [posted] price? You don’t [know]. That’s the biggest challenge right now.”
Chu’s approach notwithstanding, most of the executives interviewed believe that merchants should focus on nonprice messages when communicating to shoppers by smartphone, at least for the near future. Mark Tack of Vibes stated that some of his customers, such as Saks Fifth Avenue, do not send out any mobile coupons. And yet, he said, as frequently as any of Vibes’ clients Saks engages in mobile messaging as well as notifications through the mobile wallet. Instead of discounts, Saks sends its customers “exclusive, insider only information,” such as alerts about invitation-only events. Wintergreen consultant Bruce Biegel mirrored the prevailing view within the industry, stating that companies are trying to find their way with mobile messaging in the aisles, often without making discounts stand out. He was optimistic. “I think we’re going to do a better job of engagement, and engagement doesn’t just have to be price,” he said. “It could be product, it could be loyalty, it could be content, it could be an event, it could be sampling.” Sampling occurs when a store arranges with brand manufacturers to distribute free versions of their products, such as pizza slice samples in a supermarket, to attract customers. Biegel said that databases and in-store messaging technologies should be able to identify shoppers who might want to sample the item so that a phone message or a store clerk can invite them individually to do so. Further, he noted, merchants should position all personalized communication as a reward, an indication of privilege. That reward, he said, could be many things besides a discount. But whatever the perk, “the experience should be personalized. . . . You’ll come in and they will recognize you and they will say ‘Hey, based on [your] prior [purchases] this is what’s going on today. We are out of this but did you try that?’”
Biegel went on to say that in-store personalization efforts are constantly evolving. “There’s just all sorts of crazy stuff going on,” he said. Jason Goldberg of Razorfish pointed to Walgreens’ in-store partnership with Google to offer augmented-reality maps—video screens installed on shopping carts that present information about products as shoppers walk by them—for the chain’s loyalty program members along with messages based on their loyalty program data. Although determining the best mix of discount and nondiscount communications might be part of the motivation for the initiative, he said, “I honestly think that at this point it’s more about testing the mapping and augmented-reality technologies than it is the actual offers that they deliver through those technologies.” By contrast, Mars Advertising’s Ethan Goodman said his clients were in fact trying to learn “the right level to which you can do personalization with some of these in-store location-based technologies” such as beacons. He wants to be able to help clients with products in the frozen food aisle to “confidently deliver a frozen food–specific message” to shoppers going down that aisle. His company has yet to find the right content that will cement the sale. “Is it a telling of how the product works, is it testimonials, it is a video, is it [a discount] offer? That’s largely going to depend on the category; it’s going to depend on the shopper; it’s going to depend on the retailer. I don’t think anybody knows [more than that] yet.”
The executives also believe that shoppers need to get used to the new technologies and data-collection regimes associated with personalization. Jeff Griffin of inMarket takes a particularly conservative position in this regard, saying that his company recommends to retailers that they start by installing just one or two beacons. Of course, he said, merchants with large stores and those with more than one level, such as Macy’s, should place an additional number of units, but he disagreed with Macy’s decision to place beacons in multiple store departments. “We don’t believe shoppers are ready for that level of bombardment,” he said, noting that inMarket’s analysis concluded that people who receive only one message per store visit will use the app linked to the beacon more often and keep it longer than if they receive two or more messages per visit. That’s “a very bad thing,” he said, because retailers want shoppers to use their apps, though he quickly predicted that this low-messaging rule would not be permanent. “Long-term, we will be delivering, and frankly [will] recommend, [in-store messaging] as an incredibly powerful advertising tool” that retailers will use to deliver both personalized discounts as well as tailored information about the store and brands in the aisles. “But early on,” he said, “we absolutely believe that the shopper has to be handled with care” in order to create acceptance for the new activity.
Andy Chu of Sears was likewise reluctant to proceed with a different endeavor: sending messages (including discounts) to shoppers that Sears identifies as having inspected a competitor’s product, in either the merchant’s physical store or on the Web or a mobile device. “In the mobile site we have the capability to geofence against competitors,” Chu said. “So if you go to Home Depot three times and then you come online [to Sears and] look at a bunch of refrigerators, we might say, ‘Hey, you might go to Home Depot. We should go and give you an offer.’” To help identify these connections, Sears can purchase data pertaining to what their Web visitors view on competitor sites. Sears may also analyze its own data on people’s activities in either a brick-and-mortar Sears or on the company’s website and make assumptions about what they might look at when they visit competitors’ physical locations. To identify when those visits are made, Sears can triangulate the location of a shopper’s smartphone when the person is determined to be near a competitor store. This technique is not infallible, he acknowledged, but added that third-party marketing firms such as xAd and Place IQ have the technology to pursue it.
Despite its potential, Chu said that reasons such as “the creepy factor” are holding back this kind of preemptive personalization, as the ads might reveal to shoppers a level of tracking that they find distasteful. Moreover, some shoppers could conclude that to receive a discount offer from Sears all they have to do is shop at Home Depot, a way of thinking that Chu doesn’t want to encourage. He also was uncertain about the optimum moment for delivering such messages. “Where in the purchasing decision do you want to inject those types of trigger points? Do we just send it to everybody? How do we know that the customer will be receptive to these quote-unquote coupons versus other things we could give? It’s a slippery slope. A lot of retailers like us—I’m sure Home Depot and Walmart and everyone else—have been thinking about these notions.” Despite the uncertainties, he said, Sears was going to proceed with pilot studies.
Chu, along with virtually everyone else associated with retail marketing, believes that shoppers will ultimately move beyond the “creepy factor.” He noted that consumers who have been surveyed agree that “privacy is a big concern. In reality, what they do with their privacy setting in Facebook or in Google or everything else, I bet you a very small percentage of people go in and change the settings. So it’s just a matter of time. The younger demographics already know that everything they do online is tracked. So to them it’s just business as usual.”
The same overall attitude will eventually extend to personalized prices, all those interviewed agreed, even if shoppers still don’t like the practice. Razorfish’s Goldberg pointed out that Safeway supermarket’s mobile Just for You app already does “a little bit of that. You earn discounts on specific products based on your shopping behavior and so the pricing can seem completely unique to you.” The store encourages every loyalty program member to use their mobile device to aim the app screen at special store displays, whereupon the screen then displays an individual member’s pricing for particular products. He said the amounts are based on a shopper’s previous purchases, with more lucrative offers tendered to customers considered to have a higher lifetime value to the retailer. A mobile device is a particularly valuable tool in carrying out such discrimination. Goldberg noted that tailored discounts appearing on mobile devices are less likely to cause shopper pushback than if they were displayed more publicly, such as on e-ink shelves or video screens. Ethan Goodman of Mars Advertising agreed that “doing one-to-one dynamic pricing is certainly possible.” For most retailers “it probably requires a little bit more sophistication than is available now. But I don’t doubt that it could happen.” One executive, who requested anonymity, said his firm was already engaged in the practice and was offering customers whom its data suggested were in danger of abandoning the store higher discounts than those presented to reliably loyal customers.
In fact, some of those interviewed predicted that shoppers may enter certain businesses with the expectation that their phones will continually light up with discounts as they move through the store. To forestall unwanted bombardments, suggested Randy Jiusto of the Outsell strategic advisory firm, people may choose to turn on and off apps depending on the stores and the product brands they want to interact with while shopping. Such a situation may itself challenge retailers and brands to offer shoppers certain levels of privilege for simply keeping their apps turned on. In any event, stores will need to find ways to respond to desirable shoppers who, armed with product data, comparative prices, and a strong sense of their worth to certain brands, will demand certain prices as they stand in the aisles in front of products they want to buy. “I do think we’ll have more” in the way of personalized discounts, said Signal Digital’s Joe Stanhope. He predicted a range of discounts: one day some people will receive 10 percent off while others get 20 percent off on a particular product in a particular store; on another day some will see $10 off, while others are offered a 10 percent discount. Decisions on individual discounts will involve direct knowledge about each shopper as well as information obtained from look-alike modeling—arriving at conclusions about individuals whose backgrounds and behaviors suggest similarities to customers who have already purchased an item at a particular price. Marketers will also show similar types of people different prices and then use various statistical procedures to determine the optimal discriminatory approach. The need to rethink discounts through the lens of personalization is crucial, said Stanhope. “So many brands have trained their customers to shop [for] sales. That may not be sustainable, so [marketers] will start trying to get smarter to get out of that hole. Retraining customers to look for certain kinds of promotions right then [in the aisle], and to find out which customers really need those promotions and which don’t. You don’t have to give someone a promotion if they don’t need it yet. If they are already a loyal customer, why would you charge them less?”
This approach—awarding less for loyalty—flips the traditional understanding of loyalty on its head. Yet it does make sense in a shopping world where loyalty is a vehicle for data-driven, personalized discrimination. All the industry voices in this chapter strive to find the best ways to act on this realization in the hypercompetitive and multipronged retailing environment. At the same time, they acknowledge that what they are doing marks only the beginning; they see many more transformative changes to come.