Don’t Let Big Data Bury Your Brand
Deep into the second quarter, the chief marketing officer of a restaurant chain arrives at work to find that the CEO has dropped by. In this business, as in many others, “CMO” means chief revenue officer to the CEO, who’s here to talk sales. “There’s only a month left,” he says, “and I need a boost to compensate for what we lost because of the weather. The data analysts over in IT tell me we get the highest response to burger and apps offers. So, time for some coupons?”
The CMO’s plan was to use the coming months to strengthen brand equity, through ads promoting the quality of the food and the heritage of the brand, while moving away from the emphasis on discounts forced by a tough first quarter. But when the boss asks for something, what can you do?
This all too realistic scenario illustrates the classic tension for CMOs and, we might say, the main reason their high-level job should even exist: to achieve just the right balance between short-term revenue pursuit and long-term brand building. Advanced marketing analytics and big data make that job much harder today. If it was difficult before to defend branding investments with indefinite and distant payoffs, it is doubly so now that near-term sales can be so precisely engineered. Analytics allows a seeming omniscience about what promotional offers customers will find appealing. Big data allows impressive amounts of information to be obtained about the buying patterns and transaction histories of identifiable customers. Given marketing dollars and the discretion to invest them in either direction, the temptation to keep cash registers ringing is nearly irresistible.
The credit card giant Capital One is known for its pioneering use of marketing analytics and big data, so it might be surprising to learn about its recent realization: that too much reliance on those tools had left it without a meaningful brand. The authors explain how the number one job of the CMO—to strike the right balance between promotions that goose revenue in the short term and brand-building campaigns that support healthy margins in the long term—has become dramatically harder in the age of data-driven targeting. Capital One got its own wake-up call when its CEO, Rich Fairbank, commissioned a brand equity study. The research revealed that the company was overwhelmingly known by consumers for just one attribute: “They send me lots of mail.” Efforts to strengthen the brand and give Capital One a stronger foundation for future growth yielded five lessons—all learned and refined through conversations with other marketing executives, including Tony Pace, of Subway; Mark Addicks, of General Mills; Tariq Shaukat, of Caesars Entertainment; Russell Weiner, of Domino’s; and Jim Speros, of Fidelity, as they dealt with the same tension in their very different organizations.
We can speak to the power of analytics directly, because one of us led marketing efforts for many years at Capital One, and the other is a longtime adviser to that company and others. Capital One was “born analytical” in 1988 as a credit card company wholly invested in the data-crunching capabilities that were just then emerging. Recognizing how the credit card industry could be transformed by means of computer analysis and continual scientific testing, its founders outlined an information-based strategy that would not only give it a competitive edge as it went to market, but ultimately define its entire culture. The ability to gain granular insights into the behavioral economics of very narrowly defined segments allowed Capital One to make credit card offers to far more consumers at far more attractive rates with far less financial risk. Back then the offers weren’t electronic—but direct mail provided more than enough closed-loop feedback for analysis, and Capital One was able to out-innovate its competition by learning from it.
That history makes the direction that Capital One’s marketing has recently taken rather surprising. After a decade of promotions that drew on ever-expanding digital data sources, which resulted in growth that catapulted the company into the Fortune 200, Capital One’s CEO, Rich Fairbank, decided that the company should invest more in brand building.
Most companies have gone in the opposite direction in recent years. Inspired by success stories like Capital One’s, organizations around the world concluded that they were overinvested in brand-building campaigns with unclear returns and started pushing toward a greater reliance on promotional analytics. But Fairbank saw a troubling future for a company whose name held no relevance for customers and relied instead on the appeal of the Visa and MasterCard brands. Indeed, he commissioned a brand-equity study to understand how consumers saw Capital One, and it identified only one attribute the brand overwhelmingly owned: “They send me lots of mail.”
When sales can be precisely engineered, it’s hard to defend branding investments.
In the past few years we’ve learned a lot about how to strike the necessary balance—both through Capital One’s course correction and through insights gained from other marketers we respect. Here we offer some practices that amount to a playbook written by the smartest CMOs we know, whose companies target specific customers with short-term promotions but also depend on strong brands.
Many students of marketing are familiar with the classic case of the weekly newsmagazine that promoted its way to irrelevance. Time enjoyed a circulation of 4.6 million at its peak, in 1988. Despite a 31% increase in the U.S. population since then, that number today is below 3.3 million—a 30% drop. In 2014 Time was finally spun off by Time Warner, the company that grew from its once strong foundation. The magazine was no longer earning its keep.
The first challenge came in the form of all-news cable channels, which by accelerating news cycles made it difficult for a print weekly to carry sufficiently current information. Time’s marketers responded with a TV ad that made things much worse: “Hi, I’m Judy, one of the operators here at Time magazine. Remember, if you call now, you’ll get Time at almost half off the cover price. And this exclusive Time AM/FM walkabout, free. This offer ends soon, so call right now. Our operators are standing by.”
Desperate to replace the subscribers it was losing, Time ran blatantly promotional ads that were disconnected from the traditional image campaign it had under way—one that emphasized the heritage and quality of Time journalism and highlighted the brand’s personality. Thus the magazine unwittingly set up an experiment, and the results quickly flooded in. The promotional ads were so effective that Time soon killed off its brand campaign. A truly sad aspect of the case is that the editors had been working hard to create a better Time for a sped-up era. They knew they had one of the strongest brands for news in a world that had suddenly become news-obsessed. Their circulation-focused colleagues, by doubling down on the “call now” offers, not only failed to communicate that—they implicitly conveyed a lack of faith in the brand.
Time’s is hardly the only case that teaches a simple truth about maintaining balance. Michelob beer, for example, was for many years the premium brew that “weekends were made for”—until price promotions sent it spiraling down to the bottom of the barrel. Brand building is necessary to sustain the healthy margins that allow a business to keep fulfilling its brand promise in the long run. But because of the natural tension between brand equity and sales promotion, any pressure to spur revenue in the short term will threaten brand-building investments.
With that tension in mind, consider how great the pressure has become. Big data and analytics can put promotions on steroids. As a startling New York Times story revealed, they can enable a retailer to use a customer’s history of buying pregnancy-related goods to seize the opportune moment for promoting new-baby goods. As recent moves in the personal insurance market show, they allow the tailoring of price offers to individuals’ revealed habits to replace large-group risk pooling. In ways we encounter as consumers every day, they enable sellers to dangle offers with uncanny insight into what will make us bite.
Make Every Piece of Messaging Do Double Duty
At Subway, whose franchisees operate more than 40,000 restaurants in some 100 countries, CMO Tony Pace managed to get the balance right during his tenure: The Subway name stands for a set of brand promises and simultaneously drives sales to the 40 million customers the company serves every week. Pace told us his secret: Rather than try to balance promotions and image-building campaigns at the portfolio level, his organization strikes that balance in every element. For example, “Our five-dollar footlong promotion also features some of our ‘famous fans’ [such as the star athletes Robert Griffin III and Michael Phelps],” he says. “Their presence in the ads is for branding, not just sales.” In fact, what began in 2008 as a four-week sales-seeking promotion turned into a strategic $4 billion brand asset, complete with its own catchy jingle, logo, and hand gestures signifying “five dollars” and “footlong.”
Subway is a perfect example of a company whose marketing could easily be taken over by analytics-driven promotions. With a strong loyalty program and many cardholders subscribing to receive electronic promotional messages, it has an enviable ability to track and influence its customers’ behavior. Pace points out, however, that if analytics-driven promotions are programmed to maximize sales in the short run, they will always push to a given customer whatever that customer has ordered most often. (Other offers have a lesser chance of prompting immediate action.) That is destructive in the long term, because, as the company knows, the bigger a customer’s “repertoire” (sandwich types ordered more than once), the more loyal that customer remains to Subway.
Even worse long-term effects, Pace notes, would come from overly automated media buying and ad placement. “Programmatic marketing can lead to activities that are off brand strategy and brand-destructive,” he told us. “I’m not going to abdicate the messaging decisions.”
Yet if a marketing team stays focused on the right long-term goals, big data can often help pursue them. Subway’s new ability to uncover sales relationships among the 20 different sandwiches it offers, for example, means that it can see opportunities for growing a customer’s repertoire. The unprecedented quantity of data available to his marketers, Pace says, “does lead to enhanced precision of hypothesis making.”
Pull Branding-Level Insights from the Data
As we looked for other companies that had effectively balanced branding and sales promotion, it was hard to ignore what Mark Addicks managed to achieve, both for General Mills and personally. By the time he retired, in 2015, he had been with the company 26 years, making him, as Advertising Age reported, “one of the longest-tenured CMOs in the food industry.”
Today Addicks describes General Mills’ use of big data as evenly balanced between the pursuit of short-term sales and the building of long-term brands. But it wasn’t always that way, he says. “Initially, data was used simply to drive sales: advertising on the right day of the week, adding precision on when to engage, knowing what offer to put in front of whom, which pie to focus on for which part of the country.” Indeed, Addicks was more comfortable than many other CMOs with “letting the data tell us the consumer logic” for a promotion. He offered one example of a pattern revealed by web data: Many shoppers who looked at a certain kind of yogurt went on to consult certain chicken recipes. “It may seem odd, but we are now starting to follow the logic without questioning it,” Addicks says.
Over time, such assumption-challenging patterns have yielded insights that have helped General Mills deepen its brand relationships and generate content that is more relevant and more meaningful. In addition to enabling targeted offers, granular, data-driven understanding of consumer behavior and segments can reveal the shared concerns and underserved needs of subsets of customers, such as those who have young children or those who are responding to cholesterol guidance from a physician. “This is hard for people to get their heads around,” Addicks says. “People worry about taking their eye off the brand when you get that granular.” But when a marketer’s message can strengthen the connection between how consumers perceive a brand and the particular problem they need to solve, he says, it “drives sales, but also links to the broader brand positioning.”
For Addicks, perhaps, this was a natural direction for analytics to take. Somewhat atypically, when big data began to generate all the hype, he pushed hard to build the company’s capabilities with it. But when his team made its first pitch to the rest of the C-suite, its emphasis on the predictable sales payoff “really hurt our efforts to get the investment,” he says. The company “so wants to focus on building a brand and serving customers that a focus on data seemed inconsistent with that.” Now that it’s clear that data can be just as useful in brand building, the investment is less controversial.
Use Data to Make the Case for Brand Building
Mark Addicks was in a company that understood and loved brand building, and he had to advocate for data’s role in that. More often, as we’ve said, the brand builders are on the defensive, because others in their company want to see more sales emphasis. That’s why we were very interested by a company that makes a strong case for brand building with data.
Caesars Entertainment gained a reputation early on for data-driven decision making. Under the leadership of Gary Loveman, its culture became both sales-oriented and deeply analytical, with a general bias toward elements that are measurable and transactional. So when Tariq Shaukat became the CMO, he was prepared to drive even more accountability and granular analytics—but he was also determined to put them at the service of brand building. “With the precision available now in zip+4 TV ad targeting, we’re taking a direct-marketing approach to advertising,” he explains. That means the company can rigorously test and compare the effects of brand-focused versus promotion-focused advertising. And once customers attracted by either type walk in the door, transactions inside the casino reveal “how they behave.”
Given the nature of Caesars’s business, says Shaukat, the real brand building is done on each property, as guests experience what the company calls its “total service.” The big data Shaukat’s team uses includes not only transactions with guests, but also the answers to more than 500,000 feedback surveys every year. Together these offer deeply tactical insights as to how components of the experience drive brand perceptions and where investments will have the most impact. Caesars can also discover how customers’ brand perceptions correlate with their lifetime value to the company—and therefore can predict how an investment that increases brand ratings will translate into incremental revenue gains. In other words, Shaukat says, “if we can move from a B to an A rating of our brand, that will have x impact on results.”
It’s gratifying to see that interrogating larger data sets with better questions can lead to very different marketing decisions. Today Caesars’s TV ads are almost wholly branding-oriented, designed to heighten anticipation and emotional appeal. (A five-second promotional offer at the end allows them to do the double duty that Tony Pace advises.) Although the corporation has run into trouble from its debt load, the brand and core business remain strong.
Don’t Do It If You Can’t Defend It
We’ve mentioned Mark Addicks’s willingness to “follow the logic” of data analysis that makes surprising connections and suggests nonintuitive marketing moves. In, say, a grocery setting, the risk is small enough that it’s worth trying out a novel promotional idea. At Capital One, however, being part of customers’ financial lives means that computerized decision making could backfire if it produced outcomes that were inconsistent with the company’s values or brand.
Some CMOs apply the standard “Can I intuitively explain this before simply following the dictates of the data?” CMOs whose companies, like Capital One, are in regulated industries need also to ask, “Could I explain this to regulators, community leaders, and other stakeholders?” Taking a follow-the-data approach could lead to marketing initiatives that generate strong ROI but unwittingly expose the company to allegations of inappropriate targeting, or unfair exclusion, or using data-driven correlations that in hindsight appear discriminatory. In a human-driven, hypothesis-first model, this risk is minimized by training and awareness and oversight. In a data-driven, automated world, the risk of unintended missteps grows significantly in the absence of an appropriate judgment screen.
That is why Shaukat, even though he is thoroughly grounded in Caesars’s analytics-loving culture, insists that oversight by marketers who are comfortable using their intuition and judgment will always be needed. Surprising correlations are great for challenging a marketer’s preconceptions or the conventional wisdom; but in the end, the marketer must be able rationally to accept the logic. “If the data tells us to do something that doesn’t make intuitive sense, we don’t do it,” says Shaukat. “And if I can’t explain to a customer why we’re making a particular offer, we won’t do that one either.” This sounds like a good brand-protecting philosophy for any company.
Get Branders and Analysts to Collaborate
One last problem for CMOs who are attempting to strike the right balance between branding and promotion is that, increasingly, the two objectives are staffed by very different kinds of people. At the pizza chain Domino’s, for example, we heard from Russell Weiner, then the CMO (and now the president), that the complexities of the new data environment can’t be mastered by the jack-of-all-trades marketer of the past. “Big data requires such special skills,” he said, that it “attracts a different kind of person, who can’t rotate in and out of functions.” Just as the most creative marketers aren’t the best data people, analytic professionals usually lack the skills, the experience, and perhaps even the “internal wiring” to excel at brand, image, and creativity.
Even a coupon offer says something about the brand.
Like other CMOs we consulted, however, Weiner believes strongly that drawing on big data to create targeted offers can be instrumental to brand building. “Not showing a pepperoni pizza to a vegetarian is about both sales and branding,” he said. “A double good when you get it right, and a double bad when you get it wrong.” His company’s recent effort to encourage customers to order online rather than by telephone so that more and better data can be captured is part of helping his team deliver that double good.
But the bigger key, according to Weiner, is seamless working relationships between data-savvy creative marketers and consumer-centric big data analysts. “It’s like linking a quarterback with an offensive line, or Elton John with [Bernie Taupin] his lyricist,” he told us. The teamwork isn’t always harmonious, though. There can be “governance issues” about which side takes the lead and which plays backup. Weiner’s solution strikes us as one that other CMOs might borrow: He brought a top market researcher with him when he moved to Domino’s—someone whose skill set placed him right at the intersection of data crunching and customer understanding. This professional’s insights into the whys of what the data showed allowed him to create collaborative space for the rest of the team.
Weiner said of one campaign that drew on both sides’ strengths, “It’s not about big data. It’s about big marketing.” In the world of sports, he pointed out, many organizations have become more data-obsessed since Moneyball, the book (and movie) that celebrated the number-crunching sophistication of the Oakland A’s. But they don’t all play better ball as a result. Why is that? “Maybe the numbers get you only so far,” Weiner mused, “and after that it’s about the people getting it right.”
Between the Extremes
Another difference between marketing people correlates with age. Reliance on analytics and data-driven decisions may be second nature to the new hires in a large marketing organization, but very foreign to their veteran colleagues. In Mark Addicks’s words, big data “fundamentally challenges what they do and how they think.” Speaking from experience, he told us, “A junior person can look at the data and know more about the category than a senior executive. It can be very humbling when a junior associate in a meeting says, ‘I’m sorry, but that’s not how the category works.’”
It’s not easy to resist the pressure to go all-in on data-driven promotions. Even Rich Fairbank encountered (and encouraged) substantial pushback to his thinking about the importance of branding at Capital One. Extensive internal debates preceded the company’s strategic leap to make a significant multiyear investment in building the brand.
Yet much in the traditional marketer’s knowledge and skill set is not prone to obsolescence. Every marketing program has some degree of impact on both sides. Sales-oriented marketing influences short-term actions; branding prompts feelings and understanding of what a brand represents. But even a coupon offer says something about the brand, and even an ad with no call to action changes the consumer’s proclivity to buy. Good marketing might be defined as successfully navigating between the Scylla and Charybdis of the extremes. In the past marketers have sometimes disastrously pushed toward branding while neglecting sales (the infamous Pets.com comes to mind). Today many are steering just as frighteningly toward sales promotions at the risk of causing their brands to founder. The best way forward will always require a middle course.
Originally posted on HARVARD BUSINESS REVIEW with Robert Duboff.
Marketing strategies to help your team get inspired to make bold moves. Join me.