Measure for Measure

Online advertising has reneged on its promise. Banner ads were supposed to give marketers a cheap, easily measured way to lure hordes of consumers: Count the number of people who clicked on an ad and you'd know whether or not it was effective.

At first, people clicked, but then they stopped—response rates plummeted from around 7 percent in 1996 to about 0.5 percent today. Disillusioned marketers made like Bart Simpson in summing up their feelings about banners: "They suck."

What really sucked were the metrics they were using to measure the effectiveness of those banners. Marketing researchers say that, for most campaigns, clicks are lousy gauges of success or failure. "If you're advertising Crest toothpaste, that campaign is not going to get many clicks. That's not the goal,” says Young-Bean Song, director of analytics for Avenue A, a Seattle-based marketing agency. The mantra for online marketers these days is "brand awareness."

Of course, that's nothing new. The marketer's first goal in any medium is to make consumers aware of the brand. The second goal is to get them interested in the brand. The final step—getting them to buy—comes after those first two. Banners, it turns out, are best for the first step. "Just because people don't click through on an ad doesn't mean they’re not exposed to it," notes Russ Winer, marketing chair at UC-Berkeley's Haas School of Business and co-editor of the Journal of Interactive Marketing. "People underestimate the exposure value of that ad tremendously."

That theory was tested by University of Southern California professor Xavier Dreze in a study released in December 1999. Using eye-tracking technology, he concluded that the likelihood of a busy Web page visitor seeing a banner ad was only about 50 percent. Still, a full day after visiting the page, 11 percent of surfers could recall the ad and name the brand without assistance. An additional 19 percent could identify the brand after seeing the same banner with the brand name omitted. A recent study sponsored by DoubleClick, the Internet Advertising Bureau, and MSN found that, while clicking on a banner did less to boost brand awareness than simply seeing one did.

Those findings were buttressed by Avenue A, which used cookies to track the responses of millions of real-world Internet users to a client’s banner campaign. Sales and site visits, the agency says, were 10 percent higher for the group exposed to client's ads than for a control group shown unrelated ads. Of the subjects who saw the ads and then visited the client’s site during the six-week study, 80 percent never clicked on a banner. The message: Online ads drive traffic and sales, but do so indirectly.

Such studies are changing attitudes in the marketing world. "We're trying to train clients not to just look at click-throughs," says Chris George, interactive media supervisor for FCBi, the Internet division of Foote, Cone and Belding. "I think people are realizing what the medium can do and not try to push it to do things it can't."

But it's hard to charge advertisers based on something as nebulous as brand awareness. Page views are still the dominant metric when it comes to pricing online ads: According to the IAB, 43 percent of ad revenues are based on CPM (cost per thousand impressions). But pay-per-click is growing in popularity. The proportion of online ad revenues priced solely on clicks grew from 8 percent in the fourth quarter of 1999 to 12 percent in last year’s fourth quarter, says the IAB; during that quarter, 57 percent of revenues came from contracts with at least some cost-per-click component.

Pay-per-click deals are bad for Web publishers; they stand to lose money if an advertiser’s banners are poorly designed. But with the increasingly soft ad market, advertisers have the leverage to demand click-based payment. "There's no question" advertisers are using pay-per-click deals "as a negotiating tactic to get lower CPMs," says Song.

While clicks can be valuable, especially for promotions