On the Drawing Board

Yahoo chief data officer Dr. Usama Fayyad believes Google's advantage in keyword search will last only until Internet users find something better. "I refuse to believe that the way we will interact with the world's knowledge is to type 2.8 keywords into a search box and look at the top 10 results out of 600,000, on average," he said in a speech to the Association for Computing Machinery's Knowledge and Data Discovery special interest group, which presented him with its Innovation Award in August.

In that talk and in an interview with Baseline, Fayyad outlined how the Yahoo Research organization he oversees is tackling the scientific and economic challenges he hopes will lead to the next search breakthrough. Yahoo has invested in improving its own search technology to the point where, in blind tests, consumers judge Yahoo's search results to be just as good as Google's, Fayyad says. Still, being just as good isn't enough when it comes to building search market share, he concedes.

"There is a brand advantage, where Google is definitely benefiting from being earlier to market, at least with the technology of today," he says. "I would agree that if we don't offer something that's substantially better, there is little reason to switch."

While searching for a substantially better way to search, Yahoo Research is also looking for ways to capitalize on the company's related expertise. "Search is an example of an interactive application," Fayyad says, and it's intriguing to marketers because it gives a glimpse into what users are interested in at a given moment. But it's not the only such application.

When users at Yahoo Travel search for a flight between two cities, for example, they provide a host of clues about their preferences. So, in principle, a properly targeted ad presented at the end of a travel search could command a premium. "The market hasn't figured this out yet," he says, so Yahoo has to work on making the advantages more obvious.

Yahoo's research also suggests it should be able to deliver better results by having a longer memory for user behavior. "Just by looking at data from Yahoo Autos, we can identify 300,000 consumers per month whom we know, with 75 percent confidence, will buy a car within 90 days," Fayyad says.

If you're an automaker, every ad you show to one of those people within those 90 days is a chance to sway that person in your favor, he says.

Some numbers he cites regarding the effectiveness of these tactics, Fayyad cautions, are based on tests, not on operational results. Still, he talks about driving up conversion rates by 900 percent when users are shown a behaviorally targeted ad rather than a generic display ad. Further, a study Yahoo conducted with Harris Direct also suggests that display advertising in conjunction with search advertising is more effective than search advertising alone. Yahoo found that members who'd seen a brand ad for financial services firm Harris Direct were 60 percent more likely to search related to the ad's market category, 140 percent more likely to click through to its Web site from the search results, and 250 percent more likely to click on a "sponsored link" search ad for the company. The bottom line was a 91 percent better return on investment based on the business generated for the online brokerage.

"With search today, you get one shot and it's over," Fayyad says. "Well, if you're in a place like Yahoo, it's not over." SmartAds, the behaviorally targeted display ad format Yahoo announced in July, exemplifies this approach. By creating what it calls a "creative assembly platform," Yahoo lets advertisers design ads around interchangeable content modules—graphics, text blocks and real-time data feeds— so they can be delivered in thousands of variations. Initially, this capability is being targeted at the travel, auto and retail markets. In travel, for instance, it means the ability to display current deals for flights between two cities not just during flight search and not just on Yahoo Travel but for days afterward, anywhere on yahoo.com.

"It's actually very interesting," says Martin Laetsch senior director of search strategy at SEMDirector, a marketing automation firm that helps advertisers design campaigns and place ads. Creating these customizable display ads may be too complicated for some advertisers, particularly the smaller ones who like the simplicity of the textbased ads typically used for search marketing, he says. "But the more sophisticated advertisers will have an appetite for it if the ROI is there—and the people in the pilot programs say it's there."

Another important step Yahoo took recently—one Laetsch believes was long overdue—was to integrate the search and display ad sales teams. "It's been very difficult to purchase with them if you wanted to do an integrated media plan for both display and search," Laetsch says.

Yahoo must move quickly if it wants to capitalize on the synergy between display and search advertising, Laetsch adds, given Google's pending purchase of display advertising giant DoubleClick. (The $3.1 billion deal was announced in April, but as this issue went to press it was still being held up by U.S., European Union and Australian regulators.)

Yahoo also needs to hold onto its audience, which by ComScore's count dipped to 479 million unique visitors in August, compared with 482 million a year earlier.

What's more, time spent on some Yahoo sites is sliding. In September, JPM Securities analyst William S. Morrison warned that, according to ComScore's "minutes invested per user" metric, Yahoo Games had suffered a 47 percent drop, Yahoo Sports had lost 11 percent, Yahoo Mail nine percent and Yahoo News six percent. There were some big gains, too—time spent on Yahoo Answers rose 322 percent, Flickr was up 198 percent and Yahoo Messenger increased 36 percent—but while these are all important parts of Yahoo's social networking strategy, they're not as important to Yahoo's ad revenues as the areas on a downslide.

And, ultimately, even if Yahoo can get its audience numbers up, what matters most to its income statement is how well the company can convert that traffic into ad dollars.

If the answer continues to revolve around keyword search, that's bad news for Yahoo, which got just more than 23 percent of the total U.S. search traffic measured by ComScore's survey in August, compared with 56.5 percent for Google. "What we've been seeing is a slow but steady share increase for Google at the expense of everybody else in the industry including Yahoo," says ComScore analyst James Lamberti.

Google has established itself as the premier brand for keyword search, while people go to Yahoo for other reasons and sometimes use its search engine while they're at it, Lamberti says. Yahoo is seen as "an older brand, easy to use" but not necessarily appealing to younger consumers.

Yahoo still has a chance to redefine what it means to be an Internet media company, but first it has to learn to get out of its own way, according to John Zapolski, who quit as a director in Yahoo's user experience group in 2004 to found the Management Innovation Group consultancy. "In general, I think Yahoo's still caught between different business models and conceptions of what it should be as a business," says Zapolski. "And it's unclear to people inside the company what Yahoo should be." Yahoo has "not all that consciously" put itself in a position to become "a hybrid between a social media company and an old-world media play," but it also must follow a "shrink-to-grow" strategy that would mean cutting headcount and eliminating duplication.

If it can do all that, Yahoo has a chance to raise its value to advertisers through products like SmartAds that take advantage of the company's intelligence about its members. "There's definitely potential," Zapolski says, "but the window is closing more quickly every day now that it's up against Facebook and others who are also building strong user profiles."