Monday, April 6, 2009

A Pricing Revolution Looms in Online Advertising

Demographic profiling and behavioral targeting by such companies as Google, Quantcast, and ValueClick is slashing ad costs and threatening Web publishers

Look just to the right of this article. There, on your computer screen, lies a little, two-dimensional secret that will soon threaten major online publishers and their precious advertising revenue.

It's called the banner ad, and its cost may plummet.

For a decade, Web site publishers have relied on an old advertising model: Publishers provided advertisers access to readers, and the more desirable those readers, the more an online publisher could charge., for example, charges advertisers as much as $64.60 to show a banner ad to 1,000 viewers. (In advertising language, this is called CPM, or cost per thousand impressions.) In the past these fees made sense because The Wall Street Journal's readers are highly affluent, a perfect target for many upscale brands. The better the audience, the more advertisers are willing to pay for ad space.

But what if marketers could find new ways to reach the same audience—with ads on sites that won't charge nearly as much? What if those other ads cost as much as 95% less?

Profiling and Targeting Site Visitors

Online publishers face a big revenue squeeze as companies become more sophisticated in their ability to determine who is visiting what Web sites and when—just as marketers look to squeeze more from dwindling ad budgets.

The old online ad model is getting turned on its ear by such firms as ComScore (SCOR) and Quantcast. These and other upstarts specialize in such methods as so-called demographic profiling, which pinpoints the types of people visiting each Web site, and behavioral targeting, which helps advertisers reach a desired audience based on a person's past Web-surfing behavior.

Marketers can use these tools to reduce online ad costs dramatically. Say your company sells "Bidgets," a luxury product. Ordinarily you'd run banner ads on, which reaches your target audience of men and women who earn more than $150,000 a year. The ads are expensive—say $60 per thousand impressions—but they reach your ideal audience.

You might instead embed a snippet of code in the banners that run on This places so-called cookies on the computers of everyone who sees the ad so you can track them when they visit other Web sites. That's where retargeting kicks in. Every time a former reader who saw your ad visits other Web sites, your Bidget banner ads pop up again. The banner ads reappear because the cookie on that computer flags a retargeting "network" of thousands of sites, saying "This desirable reader is back." These new ads are cheap—$3 CPM—but they reach exactly the same audience.

Here Comes Google, via DoubleClick

Congratulations! You just used behavioral targeting to reduce your ad costs from $60 to $3 CPM, a 95% savings. (And yes, those cost quotes are based on real client experience.) Online targeting of individuals has been around for more than a decade, notes John Ardis, vice-president for corporate strategy at ValueClick (VCLK), another firm that specializes in online advertising. But interest has surged during the recession. "Obviously today's economy has advertisers looking to do more and more efficient things," he says. Not only are costs lower; results are often better. ValueClick, which provides retargeting, says click-through rates on such ads are 110% to 840% higher than average because they reach an audience more likely to be interested in a product or service.

Google (GOOG) is using its $3.2 billion acquisition of DoubleClick to get into the behavioral targeting game, using the data it monitors from millions of Web users to place more relevant ads. The size of the prize is significant; today Google has less than a 2% share of the $8 billion U.S. market for online display advertising, compared with its 63% share of the U.S. search market.

Quantcast is breaking some of the most interesting new ground in consumer targeting. Several services such as ComScore estimate Web audience demographics by using panels, or small, representative groups of people who agree to have their behavior tracked. Quantcast, however, makes 6 billion direct observations a day from cookies on computers that track the media habits of more than 900 million people worldwide. Quantcast then uses mathematical models to track the paths of consumers through millions of Web sites to determine, for instance, if you are a man age 35 to 44 with more than $100,000 in household income. The data is not individually identifiable—neither your name nor your address are collected—but the sheer size of the data set lets advertisers know exactly what type of people visit each Web site. This approach is relatively new. Quantcast launched its directly measured data service in 2006 and has been rapidly expanding it to new publishers.

And that's the rub. If you, an advertiser, know every Web site visited by your most desirable audience, you can find many ways to reach them other than through such marquee Web sites as (NWS), (MHP), or that of The New York Times (NYT)—especially if the cost is lower. "Why would I pay $60 if I can pay $4?" asks Ardis of ValueClick. "There is really going to be a backlash. Newspapers felt it. Magazines felt it. There is a reckoning coming in online channels, too."

Publishers Can Use Data to Fight Back

But online publications don't have to go the way of newspapers—if they, too, learn to use the new customer data. "It will give them new and more innovative ways to sell the audience," says Quantcast Chief Marketing Officer Adam Gerber. Consider a site whose visitors are 80% male. In the past, that site may have attracted only advertisers who wish to reach men, leaving some inventory unsold. But with better data about readers, the site can now sell to a new set of advertisers the 20% of its space that attracts women.

Christine Peterson, president of 212, New York's Interactive Advertising Club, agrees that publishers can defend themselves. For example, they could charge an advertiser a premium if they can predict that readers are in the market for that advertiser's services. "Just because I looked at a financial page doesn't mean I'm ready to hear from Charles Schwab (SCHW)," she says. "But if I've been back and forth to multiple financial pages in recent days, could they charge Charles Schwab more? Yes."

The new, behavioral-targeted data could spur demand for online advertising. "We had a marketer who was launching a low-carb bread when Atkins and the South Beach diets were all the rage," Ardis notes. "They were targeting women—but it turned out a huge portion of those who signed up were men in their 40s who were starting to develop a little paunch." The advertiser started to spend more on advertising to reach the newly discovered male audience, he says.

So hope remains for online publishers: Find unsold segments of your readers and sell them to advertisers that previously couldn't access them. Use data on past reader "click streams" from other Web sites to help advertisers reach the consumers who are most interested in their products.

Publishers, your audience data is waiting. Advertisers want it. But you'd better hurry, because the price of your old online business model is falling fast.

No comments: