Little Engines That Can; Even Google Can't Think of Everything. A Host of Start-Ups Are Working to Fill Niches and Capitalize on the Search Boom

Article excerpt

Byline: Brad Stone

A Google search for the phrase "apple tree" draws 2.4 million results, all tucked into an endless, impenetrable catalog of blue links. Entrepreneur R. J. Pittman thinks that's a few too many. "Traditional search engines don't solve the information-overload problem," he says. His Sausalito, Calif., start-up, Groxis, is working on a solution. Its downloadable software tool, Grokker, sits on the desktop, plugs queries into the major search engines and uses home-cooked algorithms to analyze the pages and organize them into categories. Then it renders those categories on the screen in an easy-to-parse, graphical display of circles and squares. Grokker is available on the Net for $50 while the company tests a free, ad-supported version. "Search on the Internet needs to graduate to the next level," Pittman says.

Groxis isn't alone in that endeavor. Over the past few years, dozens of start-ups have followed in the wake of the search giants like the pilot fish that travel with sharks, hoping to feed on leftovers. Thanks to the success of Google, the search ocean is now large enough to support many of these smaller life forms. Securities firm Piper Jaffray predicts global revenue from search engines will grow to $8.9 billion in 2007, up from $2.6 billion today. Though big players like Google, Yahoo and Microsoft will take the biggest bites, there's still plenty left for upstarts with unique technologies and fresh approaches. "The exciting thing for me is that this industry is so young. There's lots of innovation left to be done," says Eric Matlick, CEO of a two-year-old search-marketing firm called Industry Brains.

In fact, there are so many start-ups entering the search fray these days that to sort through them you almost need an, umm, search engine. Eurekster, launched in January, mixes search with social networking, where you make online connections to friends and business associates, and delivers results based partially on what those people found useful in their related searches. Another effort, Nutch, is an open-source search project; programmers around the world freely contribute to its code. One of its cool planned features is letting searchers tinker with the parameters of the search algorithm. For instance, they can tell the search engine to focus only on the number of times a search keyword appears inside Web pages and to ignore other, possibly irrelevant, factors. …