January 26, 2007 Searching for the new search by Carolyn Murphy If an Internet year is like a dog year, as one Mangrove Capital partner suggests, then Google Inc. is pushing 60, while Yahoo! Inc. and Microsoft Corp.'s MSN are in their mid-80s. By contrast, search is in its infancy and ripe for change, says Mangrove's Mark Tluszcz, an early Skype Technologies SA investor. Despite their dominance, today's major players in the search realm are nevertheless vulnerable to new players relying on fresh approaches to serving consumer needs. Tluszcz thinks recent portfolio addition Quintura Inc. will be among those to usher in a new era for search. It's one of a host of startups trying to innovate beyond their first-generation predecessors, targeting a growing market and banking on a few fundamentals -- semantics, visuals and verticals. A November 2006 JupiterResearch report pegs search as the largest online advertising category at 41.4%, expected to grow by 42.9% by 2011. Data shows it was just shy of $6 billion for 2006. Meanwhile, Google, Yahoo! and MSN are in the lead, with 49.5%, 24.3% and 8.2% of market share respectively, according to a NetRatings Inc. report. Hoping to change that, startups have largely been netting venture dollars on two fronts, says Foundation Capital general partner Charles Moldow: vertical and visually focused search sites. Vertical search is "a deeper effort on indexing and a better packaging of information," Moldow says. Sites include employment engine Simply Hired Inc. and travel startup Kayak.com, which both received funding last year. On the visual front are sites using previews, a la Snap.com, or clustering, the grouping of results by category, like Vivisimo Inc.'s consumer site Clusty. Moldow describes these innovations as incremental improvements. "We think search is a huge market and business," he says. "But we don't believe that search will improve through incremental changes to either the statistical approach to keywords or the link analysis. We don't think that incrementally improving those capabilities, which today is the underlying technology and foundation of search, is really going to be where we're going to see the breakthrough innovation." In November, Menlo Park, Calif.-based Foundation led a $12.5 million Series A round for natural language search startup Powerset Inc. of San Francisco. Still in semistealth mode, Powerset expects to launch full-scale by the end of 2007. This summer the company will likely be loking for more capital. Powerset's investors include Internet vets like PayPal Inc. co-founders Peter Thiel, also a Founders Fund LP partner, and Luke Nosek; Facebook Inc. and Plaxo Inc. co-founder Sean Parker; and eGroups founder Carl Page. The startup's chief is Barney Pell, a former entrepreneur-in-residence at Menlo Park, Calif.-based Mayfield Fund. Unlike a statistical search engine that relies on word frequency, semantic or natural language search focuses on the meaning of words, their representation and their relation to each other. Powerset has been developing its technology for nearly 2-1/2 years, and though natural language indexing has long been considered a good, if expensive, idea, Moldow says, the economics have shifted, so it makes sense. Like Moldow, meaning-based search startup Hakia Inc.'s CEO, Riza Berkan, has also bet on semantics. He doesn't believe other approaches represent a paradigm shift. "They don't attack the core problem of relevancy," he says. "The major search engines are really not going to disappear because of this fragmentation." Founded in 2004, New York-based Hakia drew an undisclosed, $5 million VC round, led by Noble Grossart Investments Ltd. in November, which brought its total funding to $16 million. Fully funded, the startup's site is in beta mode, set to launch before the end of the year. Hakia aims for profitably by 2008. But applying a semantic filter to search isn't necessarily groundbreaking, says SearchEngineWatch.com news editor Kevin Newcomb, adding that Google has been using its own form of semantic filter for a few years. If a startup such as Hakia was to come through with something groundbreaking in the form of an intelligent search enginge, it might change users' expectations of how search should work, he says, but adds it would be hard to unseat Google. A more likely possibility: Use it as an opportunity to showcase its technology for acquisition. "People don't really want to change their searching behavior -- just ask Microsoft or Ask.com," he says. "They've been fighting an uphill battle with no real gains against Google, a competitor with virtually unlimited resources that has proven to be surprisingly agile for its size." He says opportunity lies in vertical and social search. In local search, San Mateo, Calif.-based Zvents Inc. recently landed $7 million. Social networking for the over-50 crowd is the idea behind Boston-based Eons Inc.' Cranky.com, a vertical search site that Monster.com founder Jeff Taylor launched Jan. 9. "The ones that are trying to build a better Google and that kind of thing are not going to do it," Newcomb says. Foundation's Moldow acknowledges that Google isn't going away. He likens what could be on the horizon for search to how today's dominant player displaced mid- to late-'90s companies such as InfoSeek, Excite and Lycos, with statistical search and link analysis, which yielded better results. Most had a six- or seven-year run until Google powered into the lead, he says. "I think we're at another point at which consumers perfectly well understand the search model, understand what they're getting for their effort and their time, and if something better comes along, I think we might see a migration," Moldow says. For Mangrove's Tluszcz, all bets are on visual search. Alongside Russia's ABRT Venture Fund and Boston's OpenView Venture Partners, the firm has lined up a seed round between $1 million and $5 million in funding for Alexandria, Va. incorporated Quintura, whose technology stems from the work of 30 Russian engineers. Tluszcz calls the underlying technology "world class." Quintura plans a $5 million to $10 million Series A this spring. When engaged, the search engine returns a cloud of words related to a term based on connections between the words. The startup is building a neural map to construct a Web index with more relevant results. In addition to its search engine, Quintura features one for kids, while Clusty, too, offers users a search cloud. This approach to search is more intuitive, Tluszcz says, and he parallels the industry to telecom in 2001, '02 and '03: "We see some entrenched players who are simply not innovating, who become so big that they believe they're not beatable anymore ... You get that same behavior." But he believes comparing Quintura to, say, Skype is premature. "My hope is that it would be close, but I think it's too early, frankly, to draw any serious parallels there." Looking ahead, some obersers argue that search isn't a winner-take-all scenario because consumers continue to want options. There are many types of search and room for many different technologies. "I think that it can't end up one-size fits all. It just has to be the case that certain technologies are better suited for certain tasks," Moldow says. But, entering the fray is not without obstacles. "The challenge in the case of the search space is [that] Google's become a verb now. ... You have to be so much better than that experience," Tluszcz says, for users to change. But, he adds, the challenge isn't one to shy from. "Nobody can get [the entire Web] right and I don't think ultimately you're going to have just the same dominant players." Further, he says: "We've been told for too long by a small amount of players, 'This is how you can do things,' and I think search is right at that cusp."