Efficient Frontier: Hacking Madison Avenue

They may not look so sexy, those four-line ads on the right side of the page when you do a Google search. But Rob Woods loves them. After all, they let Woods -- director of marketing and merchandising at building-materials wholesaler BuildDirect Technologies -- and other advertisers create their own minimarkets defined by the words searchers type in. But until last summer, Woods was spending several hours a day tracking and changing bids on 1,500 terms on Google and Yahoo! "It was very time-consuming," he says.


Enter Efficient Frontier Inc. This New Age advertising agency operates just like Madison Avenue shops, in a sense. On one side, it has ad clients, folks like Woods. On the other, it has access to a portfolio of ad space -- in this case, the Web search engines. It matches the two. But that's where the similarities end.

The 70-person Mountain View (Calif.) startup is the largest buyer of keywords on Google, some 15 million a day. Borrowing mathematical formulas developed for Wall Street portfolio managers, Efficient Frontier's computers sort through upward of 10 billion variables, such as the number of people who search using particular words and changes in what companies bid for their ads to appear on pages with those keywords. Then they spit out the optimal mix of keyword purchases and suggest adjustments throughout the day. Efficient Frontier claims that, on average, its method improves measures such as the rate at which people click on ads by 30%.

Woods, whose company operates over the Web, can buy 10,000 keywords on Google and 6,000 on Yahoo. Since he started using Efficient Frontier, he's seeing people on average click on twice as many of his ads. Woods's sales are up 35%, even though his marketing budget is flat. "The reason people spend on Google and Yahoo is not that it's sexy but because it gets results," says Ellen Siminoff, CEO of Efficient Frontier and a former Yahoo executive. "It's about visibility and control -- and they can get that on the Internet."

Even search ads may not be the endgame. John Hagel, a consultant and author of several books on the Net and business, notes that the problem with advertising of any kind is that it essentially aims to hijack people's attention. "You have to move from an intercept model to an attraction model," says Hagel.

Still, few expect search marketing to slow anytime soon. Siminoff believes that ultimately any kind of digital media, even TV, can be optimized the same way. Suddenly, ad agencies, which depend in part on the difficulty of marketers and publishers to measure results, are facing an onslaught of highly analytical competitors. They could be staring at the companies that will replace them as media goes digital. (Read on for the next installment, "BBC: Step Right into the Telly.")

By Robert D. Hof in Mountain View, Calif.

Before it's here, it's on the Bloomberg Terminal.