Frenemies: The Epic Disruption of the Ad Business (and Everything Else) (17 page)

While Irwin
Gotlieb
worked hand in hand with Martin Sorrell to transform GroupM and WPP into a global enterprise, the future challenge that consumed much of Gotlieb's waking hours was how to master big data. Digital cookies yielded data that allow marketers to better target individuals rather than amorphous groups, but also to measure the effectiveness of a marketing campaign. Big data was the promise Carolyn Everson and Facebook employed to entice advertisers. And big data did one other significant thing: it granted more power to media agencies who used data to shape marketing campaigns. But this was so only if the media agency was big enough to hire the data engineers and convince clients it could unlock consumer secrets.

“The kinds of stuff that we are focusing on, the technology, the data, all of these things benefit from scale, which is a real amplifier of those capabilities,” Gotlieb said one day from his Garment District office on Seventh Avenue and West Thirty-seventh Street. “The more touch points we have, the more volume we have. The more data we gather, the more we understand. Let me give you an example of something I could foresee: the GroupM shops are going to be the guys who do the major ideation. What I see us doing is that we are increasingly able to slice and dice audiences into microsegments, each of which will need customized communication. When you logged into a client's e-commerce site, we custom built an offer for you in real time. How did we do it? We know what you purchased in the past. We know what you own. And we know what the next thing you're likely to need is.
That's different from what I need. Each of us gets a highly customized offer.” To customize, however, will mean that the media agency may encroach on some creative agency's function, as Gotlieb concedes: “We might be doing creative executions” to customize messages.

Customized messages beg another question: How to define creative marketing? CMOs like Keith Weed like to say that the Holy Grail separating effective from ineffective advertising is a “big idea” executed by “good creative.” But when messages are customized to individuals and not aimed at a mass audience, does the individualized message shove aside some big creative idea? If it did, it would cement the rising power of media agencies like Gotlieb's GroupM. We saw this occur when one of GroupM's media-buying agencies, MediaCom, created the “Love Is On” campaign for Revlon, inducing women to carefully apply lipstick and fragrances to open up to love and romance. One could argue that this was a “big idea,” but the idea was conceived not by a creative agency but by a media agency and it was targeted at individuals. We would see a manifestation of the rising power of media agencies in 2016, when Omnicom won a bid to create a single agency dominated by its Hearts & Science media agency to represent AT&T, the fourth largest advertiser in North America. These “Mediapalooza” accounts, as they've been called, start with data, for the data shapes the narrative. “This is not brain surgery. You don't need a big idea,” Gotlieb says, choosing not to dwell on the classic big ideas crafted by those who'd qualify for an advertising Hall of Fame—Bill Bernbach, David Ogilvy, Mary Wells Lawrence, Lee Clow, Keith Reinhard, Bob Greenberg.

Big data is siphoned from three sources. Generally, first-party data is gathered by the companies that interact with customers directly and usually retain their credit cards; plus from department stores, credit card and car companies, Amazon, magazines and newspapers, and companies reliant on subscriptions. First-party data has the names of
individuals, which the company can use and can share. Second-party data is anonymous but also contains a rich trove of individual information from various sources, including clients, data companies like Nielsen, comScore, and a WPP-owned agency, Kantar. Anonymous third-party data is purchased from catalogues and stores. GroupM assembles all this data and strives to match the products of its clients with targeted consumers.

More than a few obstacles impede the data collection efforts of media agencies. Companies like Facebook and Google, who bank some of the richest first-party data, resist sharing much of it. Eric R. Salama, CEO of Kantar in London, its thirty thousand employees rivaling Nielsen and making it the world's second largest research company, expresses a common frustration with Facebook and Google among agencies. “We really have to get into these walled gardens to really understand what people are doing and how they're behaving,” he says. Mobile phones pose obstacles as well; marketing on mobile phones is complex. Salama cautions, “How do we test ads on mobile?” And if the mobile phone lacks a flash drive, they can't show the ad. Talent may be big data's most crucial impediment, Salama thinks. “Everyone wants to hire data scientists and engineers.” The supply is limited, the competition intense.

In the data arms race,
Irwin Gotlieb set out to build a state-of-the-art data weapon, known internally as the Secret Sauce project. Gotlieb was intent on building GroupM's own proprietary data system because, as he would publicly complain in 2015, Facebook and Google had their own ad tech vehicles to target ads and were, in effect, muscling agencies and clients by warning: “If you want to buy ads on our properties, you have to use our ad tech tools.” He saw other holding companies making inexpensive deals to utilize Google's DoubleClick
or Facebook's tech tools, and he thought they were crazy to become dependent on a frenemy. He liked Carolyn Everson, was proud to be listed on her long list of mentors, but was not seduced by her luminous smile. “They want to sell their impressions for the highest possible price,” he says. “I want to buy them for the lowest possible price. In order for me to accomplish this objective, I have to keep them in the dark as much as possible.”

With the support of Martin Sorrell, he says he spent a total of $2.5 billion to build GroupM's own tech stack, the unique operating system and supporting programming to run a complex application, in this case their secret sauce. One day in a GroupM conference room, Gotlieb and Harvey Goldhersz, CEO of GroupM Analytics, shared its ingredients. The initial ingredient came when WPP acquired digital marketing company 24/7 Media. Gotlieb says the two biggest attractions of 24/7 were its hard-to-recruit two hundred engineers, and WPP's desire to block competitors from acquiring the company. Earlier, WPP bought KBM, or KnowledgeBase Marketing, a company that had relationships with three thousand retailers and stored a cornucopia of what is known as personally identifiable information (PII), including the people's names and their behavioral traits. Now KBM gathered sales data from companies WPP either owned outright, like Kantar, or those in which it was a major shareholder, like Rentrak and comScore. In addition, KBM vacuumed additional data, providing GroupM with a range of about forty thousand personal “behavior attributes”—what people purchase, eat, read, watch, their interests and hobbies, along with their age, sex, and neighborhood. GroupM only has anonymized data, but it assigns each individual a digital identifier or persistent identifier. “We know you're number 001,” Gotlieb says. In all, of America's 200 million adult population, Gotlieb says Group M has anonymized data on almost all of them, the exceptions usually being those who don't have a bank account or e-mail address.

What's key, Gotlieb says, is the anonymized PII. “We rely on persistent identifiers, not cookies,” and convert them into cookies. He offers this example of how it would work: “Assume we have a client who wants to sell tomato paste.” From available data on people who buy lots of tomato paste and tomato sauce, GroupM then matches these people with its database and comes up with a universe of regular tomato paste and sauce consumers. GroupM then aims a marketing campaign at them and can tell whether they saw the marketing messages. Then they go back to the sales data to learn whether the targeted consumers purchased more. “That's one measure of whether the marketing campaign worked,” Gotlieb says. It also allows them to match this data with other data they've collected on the individual, Gotlieb continues. “This also allows an assessment of the same group's other attributes. Do they drive similar cars? Do they watch similar TV shows? Do they fit in the same income segments?” GroupM knows a tremendous amount. But, he insists, they don't know their actual identity. “We know the family. We know the income. We know the credit score. We just don't know the name and the address. KBM does. It's not the law in the U.S. that prevents us from knowing the names, it's WPP's internal policy.”

“Think of it as a spreadsheet,” Gotlieb continues, “with two hundred million rows and forty thousand behavioral characteristics.” To illustrate the difference targeted marketing can make, he turns to the consumption of rice: “Seventy percent of the U.S. population has not bought rice in more than two years. It doesn't mean they don't eat rice.” But the 30 percent who do buy it “tend to be ethnic—Asians and Hispanics—because rice is endemic to their cuisines. And they buy a lot,” usually in bulk. “Only ten percent of the population buys convenience rice products, like Minute Rice and Uncle Ben's.” These convenience rice products each only have a market share of 2 to 4 percent, and an American consumer typically buys only two boxes a year. If GroupM represented one of these 2 to 4 percent brands, what he
would do is create a marketing campaign offering enticing rice recipes. “If I can move them from two boxes a year to three boxes a year, you've got a fifty percent increase in volume.” An even bigger marketing opportunity for Minute Rice and Uncle Ben's is among those who buy rice in bulk. Since GoupM would be able to identify these consumers, for the 20 percent of them who might be potential customers, his marketing message would be, “‘One day you will come home and not have time to cook your rice.' Maybe they will buy one box. If I can penetrate five percent of that twenty percent, my convenience rice has gone from two- to three-percent penetration.”

Gotlieb believes they have reduced the guesswork. “Today we have evidence that this person saw two commercials and bought this product,” he declares. With clients increasingly asking agencies to prove their return on investment by demonstrating that their marketing campaign was effective, and with clients willing to pay them a bonus for success, Irwin Gotlieb is convinced his secret sauce will enhance GroupM's bottom line.

But not just yet. “It is going to take another couple of years for this to be fully deployed,” he says. “We haven't exposed it to all of our clients because we can't retrain our people quickly enough. We're changing out the engine while the plane is flying.” There is also a security concern. Media agencies usually share clients with other creative agencies outside of WPP, meaning that Gotlieb works with an Omnicom or IPG creative team. How much of his secret sauce does he want to share with competitors? Like Facebook and Google, he has reasons to wall off his data.

He also knows that the era of shared campaigns may be shrinking. Media agencies like his are changing in profound ways. If his secret sauce works, more clients will push for the media agency to serve as the pilot of the marketing plane. The definition of “what is creative will keep changing,” Gotlieb says, and “the creative function will
bifurcate,” with media agencies, as his friend Michael Kassan believes, planning and executing entire campaigns, perhaps even the creative.

A number of impediments may be thrown in the path of Irwin Gotlieb's future vision. He is aware that digital and consulting giants pose a competitive threat, but is he guilty of assuming that his holding company competitors are more inert than they are? Can he galvanize his own organization to discard the familiar legacy model they have relied on for decades and embrace a more scientific approach to targeting? Can media agencies retain the trust of their clients? But perhaps the primary impediment to Gotlieb's secret sauce centers on privacy.

What is the privacy line Gotlieb won't cross?

Although they could collect real names, he says, “We will not cross the line to know their name.” Of course, since they know “a tremendous amount” about each individual and have assigned a personal identifier to each, their name is not crucial for marketing purposes.

Did he worry that GroupM's abundant data on individuals is too much information and might be abused?

“No,” he instantly responded. “Could I get to a place where I have too much? It's possible, but unlikely. And to be blunt, we have nowhere near what Facebook, Google, Apple, and Amazon have.”

Gotlieb's statement won't discomfort the digital giants, for he is acknowledging their data dominance. But it would discomfort consumer advocates. They are shocked that privacy limitations are set by companies like WPP, not the U.S. government, which doesn't impose the same strict privacy laws of other countries, particularly Western Europe's.

Perhaps because he lives in Western Europe, Eric Salama is acutely aware of the threat stricter privacy laws pose to his data company. “I think the privacy issue is going to become a bigger issue,” concedes Salama. “In Germany, it's massive. In the United States, less so. I think it's going to be huge because I don't think most people have a clue
how their data is being used. There is an implied contract where people are getting all this content for free. But someone has to pay for it. And the pay-for-it comes through advertising. But that's an implied contract, not an explicit contract. And people are being targeted on the basis of their behavior, and we need to be much more open with people that when they sign up for something, that they know what they are signing up for. If I go on my computer or my mobile, I'm quite happy that a cookie has been stored there. It means that I don't have to reload everything, and I'm getting stuff that is relevant to me.”

What Salama does not say is that media agencies are wary of giving consumers the choice of opting in, meaning that instead of allowing consumers to follow labyrinthine instructions to opt out of sharing their cookies with companies, consumers would have to voluntarily opt in before their data could be shared. While consumers choosing to opt in would advance privacy protections, it would derail the targeting dreams of Irwin Gotlieb and media agencies.