Advertising's revenge of the nerds
Slogan-spewing creative directors once ruled Madison Avenue, but the ability to track which ads actually swayed consumers may give the number crunchers the upper hand.
NEW YORK (Fortune) -- Dimitri Maex, a tall, lean Belgian who runs Ogilvy & Mather's analytical team, remembers the day his world changed. It was 2004 and he had been invited to join a group of Ogilvy's top executives at Cisco's San Jose, Calif., headquarters. The meeting was a big deal for Maex. As a number cruncher, he had rarely been asked to participate in such a high-level pitch in the past. Almost as an afterthought, his presentation was scheduled last. And he had a tough act to follow.
The creative team, which is usually the highlight of such meetings, showed some initial ideas for the coming year's campaign. But it wasn't until Maex, who has a master's degree in econometrics, started pulling out charts and graphs that Cisco's chief marketing officer got excited. What had the CMO on the edge of his seat were a series of statistical algorithms that Maex said could predict when Cisco's clients were going to increase or decrease their spending on Cisco products. Maex's presentation helped Ogilvy cement its relationship with Cisco (CSCO, Fortune 500) and suddenly his place in the food chain had changed dramatically.
A similar power shift is playing out at other agencies. After years of calling the shots, the traditional Mad Men of advertising -- the creative types who cooked up memorable sell-lines like "the ultimate driving machine" -- are increasingly sharing the spotlight with, you guessed it, the nerds. Or as Jon Bond, a co-founder of Kirshenbaum Bond + Partners, which has done work for Target (TGT, Fortune 500) and Panasonic (PC), says, "If we were in India, it would be as if the untouchables had suddenly become the ruling class."
What has allowed the lowly quants to sit at the same table as the advertising Brahmin is a new way of thinking about the creation of desire. No longer is purchasing simply considered an emotional act, but rather one that can be measured with scientific precision. Our synapses, it turns out, are hardwired to respond to certain types of messages. And newly available numerical analysis is taking the guesswork out of deciphering which ads will best turn consumers into Pavlov's dogs -- purchasing not by will but by reflex.
Correlating ads and sales has been the Holy Grail of marketing ever since department store pioneer John Wanamaker quipped, "Half the money I spend on advertising is wasted; the trouble is, I don't know which half." But a lot has changed since Wanamaker uttered those words more than a century ago. Marketers have spent the past 25 years collecting heaps of data, mostly by tracking bar code purchases. But until recently, mining that data was costly and inefficient.
The main tool advertisers had to work with were focus groups, but while these groups were good at predicting attitudes -- as in, Mrs. Smith has a favorable opinion of Brand X -- they were lousy at forecasting actions. (Just because Mrs. Smith likes Brand X doesn't mean she'll buy it.) Focus groups could tell you which women ages 18 to 45 were likely to watch the Oprah Winfrey Show. But, says Curt Hecht, who runs Publicis' VivaKi Nerve Center, a new media hub, "we need to go a step further and instead target dog-owners, or diaper-buyers or moms who do laundry." Now, thanks to digital media, advertisers are getting better at making that leap. And these new techniques are migrating fast to print and TV.
Fueling the rise of this quantocracy is the recession, which has made advertisers hungrier for proof that their multimillion-dollar campaigns are working. For some companies, it's simply a matter of scarce resources. General Motors, which is the second largest advertiser in the country after Procter & Gamble (PG, Fortune 500), slashed spending 15% in 2008 to $2.1 billion, according to Nielsen. The carmaker said it plans to slash its marketing budget, including incentives, by a further $800 million in 2009. This bottom-line focus will only accelerate as the Obama Administration is called upon to justify the billions of taxpayers' dollars it has spent to prop up corporate America.
On the plus side, this show-me mentality has the potential to modernize what is an outdated agency compensation model. Coca-Cola (KO, Fortune 500) is leading the charge. Instead of paying a flat fee based on the hours worked, as is the industry norm, Coke is starting to grade its agencies based on four criteria, including sales and market share gains. The new compensation model will be introduced in 35 markets this year and is expected be rolled out to all of Coke's agencies worldwide by 2011. Agencies that perform have the chance to pocket 30% above their costs -- far more than they make under the old fee system. But those that disappoint could walk away with nothing. Procter & Gamble is testing a similar approach and most observers expect this Darwinian model to be the way of the future.
Some agency executives say they like the idea. "We never felt we should be paid by the hour," says Jeff Hicks, the CEO of Crispin Porter + Bogusky, which works on the Coke Zero account. But others worry about the downside of measuring what is essentially a creative process. "Coke sales go up and down for a lot of different reasons," says Mike Hughes, creative director of the Martin Agency, which counts Wal-Mart (WMT, Fortune 500) and Geico, but not Coke, as clients.
Whether or not the creatives like the reordering of their world, there is little they can do about it. Economic downturns have a way of shaking up industries. (Just ask your cranky friends who work at newspapers or investment banks.) Even more seductive than the promise of a fatter paycheck for agencies that perform is the idea that this left-brain power grab can somehow save advertising from a painful downward spiral.
You need to spend only a few minutes with two very different men -- a creative like Richard Kirshenbaum and a quant like Maex -- to understand how and why this power shift is underway. Kirshenbaum and Maex work only a few dozen blocks away from each other in Manhattan, but their mindsets are light years apart.
When you walk into the West Village office of Kirshenbaum, a co-founder of Kirshenbaum Bond + Partners, the agency that gave us Wendy, the pigtailed redhead who is the face of the Wendy's fast food chain, you know you are in the presence of a creative mind. His walls are lined with sketches and photographs. Brazilian Bossa Nova music plays softly in the background. Kirshenbaum, 48, wears a Cartier Tank watch, plays squash at the University Club and has a standing table at Da Silvano, the celebrity-riddled SoHo eatery.
It's not that Kirshenbaum is against quants. In fact, his firm has a whole team of them. "Quants are valuable, especially in this economic environment," he says. "But the creative lightning still has to strike."
The target the lightning has to hit these days is smaller, though. The rise of new media like the Internet, iPods, and BlackBerrys -- not to mention those advertising unfriendly digital video recorders -- is splintering our attention, which was once captivated by the 30-second television spot or the glossy magazine ad.
And that's where Maex comes in. In a huge Midtown loft overlooking the Hudson River, Maex overseas a team of 60 Ogilvy statisticians, a group that has mushroomed three-fold since 2005. If Kirshenbaum is a Mac, Maex is all PC. He wears a black, button-down shirt, khakis, and a rubberized sports watch. Lunch consists of a turkey sandwich at his desk. But although he works under fluorescent lights surrounded by white boards, Maex may well be the future of advertising.
The hope is that the quants like him can help the creatives hit that ever-splintering target audience. "They can make data tell its story," says Google's chief economist Hal Varian, who predicts that the role of the statistician will be the sexiest job of the next decade.
Unearthing a coherent story out of a pile of numbers was exactly the quantative marketing challenge that BF Goodrich faced when, in 2006, it solicited help from the Martin Agency, a creative shop based in Richmond, Va. Like most agencies, Martin had previously relied on a research department staffed by former account executives. But more recently, Martin had begun hiring Ph.D.s who specialized in statistical regression modeling. BF Goodrich had identified 47 million potential customers, but because of a limited advertising budget, didn't have the resources to go after them all. The tire-maker wanted Martin to help identify "the big fish," those who were most likely to buy its products.
"We had a ton of data," says Tom Jupena, a BF Goodrich brand communications manager. "But we needed help translating it into something we could act on."
Using statistical modeling, the Martin Agency's Dr. Lauren Tucker and her team of consumer forensic specialists spent a year crunching numbers. What they came up with were two types of big fish.
One was the typical BF Goodrich customer, which Tucker called "Mike." This group of 17 million people was not affluent, but spent whatever money they had on automotive products. BF Goodrich had been doing a good job of reaching the "Mikes" with its print ads in automotive trade magazines.
But there was another group they had overlooked. "Ty," as Tucker named this group, consisted of 12 million people. Ty was a bit more upscale than Mike and liked small, fast cars and anything digital. Ty spent most of his time online or watching sports channels like ESPN -- two areas in which BF Goodrich hadn't been advertising.
First, the Martin Agency created a series of TV commercials that featured all-terrain vehicles riding over rugged terrain or speeding through the desert. The idea was to show the toughness of Goodrich tires. Then, to zero in on the Tys, the agency developed a voice-activated, computerized driving game. Later this summer, the company plans to launch an online community that it hopes will be a hive for the Tys by allowing members to map their road trips and upload photos. Since the new TV commercials debuted in 2007, BF Goodrich says it has seen double-digit increases in brand awareness and purchase intent.
How did Tucker and her team identify the Mikes and Tys? They input lots of data -- including "hard" sales and profit figures, but also "soft" numbers that track brand preferences -- into a regression model. "Basically," Tucker says, "it's one big hellish equation." What does the equation tell her? "You know how there are girls who are attracted to a certain type of guy?" Tucker says. "Well, there are people who are predisposed to certain brands. The people who are attracted to BF Goodrich are different than those attracted to Michelin." The numbers flag those with a preference for BF Goodrich in their DNA.
Before advertising's revenge of the nerds plays out in full, there are some big challenges ahead, including technological and cost hurdles, as well as concerns about creativity and privacy.
Targeted (also called addressable) advertising suffered a big setback this spring when a much-touted new product from Canoe Ventures, formed a year ago by the nation's six largest cable operators, was pulled from the market over technical glitches. The product was supposed to allow advertisers to identify households with annual incomes that exceeded $100,000, but it ran into problems trying to "insert national ads into the local cable infrastructure," says Canoe spokeswoman Dana Runnells
Meanwhile, some companies that have tested addressable advertising are deciding to pass. BMW, for instance, discontinued a test with Visible World, in which very high-income households saw ads for its 7 Series and slightly less affluent households got the 3 Series commercials. Creating different versions of the ads "wasn't cost effective," says Patrick McKenna, BMW's manager of corporate communications.
And there's another worry: Too much data, say creative types, will make advertising boring. "If we're all using the same science, it's going to lead to a lot of really common themes," says Mark Fitzloff, executive creative director of Wieden+Kennedy in Portland, Ore., where he has worked on the Nike (NKE, Fortune 500) and Procter & Gamble accounts. That is a problem, he says, that has plagued direct mail, which is where he got his start. It's no coincidence that direct mailers all look the same. Marketers spend a lot of time testing which buzzwords generate the biggest response. As it turns out, "free" and "guaranteed" top the list. No surprise there.
Before we charge off into this Orwellian future, privacy advocates are determined to have their say. Leading the charge is Rep. Rick Boucher of Virginia, head of the House Subcommittee on Communications, Technology, and the Internet, who is pushing for safeguards. "I would much prefer to receive Internet advertisements that are relevant to my interests," he told a subcommittee joint hearing in June, but added that he wants to ensure there are rules to protect consumers, including allowing them to "opt in" before their information is shared with third parties.
None of these mine fields have stopped companies from trying to get rich off collecting and analyzing data. Which company will be the Nielsen of the modern age remains to be seen, but there are lots of horses in the race. The cable companies own the set-top box data that track viewing habits. Google (GOOG, Fortune 500) and Microsoft (MSFT, Fortune 500) own the click streams. And Facebook and Twitter are closing in. But don't count Nielsen out yet. The company has devised a new type of measurement called BuzzMetrics, which tracks Internet chatter to determine what the blogosphere is saying about a brand.
As seductive as data is, no one wants to live in a world ruled by numbers. Advertising has always been a mix of art and science. And the best quants know this. But as Albert Einstein once said, "technological progress is like an axe in the hands of a pathological criminal." Before the Nerds are finished transforming Madison Avenue, there are bound to be a lot of bodies.