The last time you looked for something on the Web, chances are, you typed a request into a Google search bar to find it. If people like Max Ventilla have their way, you'll soon be using other tools to unearth information online.
Ventilla is a co-founder of Aardvark, a startup that's hard at work on software that helps people find what they want online by asking others who know the subject matter best—and who are likely to weigh in with a helpful response. The 16-person San Francisco startup is testing a tool that lets users pose questions by instant message or e-mail to people they know.
Aardvark is among a crop of businesses specializing in social search, a technology at the crossroads of two of the Internet's hottest trends—social media and Web search. "If someone's looking for a recommendation on 'great music' or a 'hotel room in London', not even 20% of people are going to be satisfied with a search result" from traditional engines, says Ventilla, the 29-year-old ex-Google employee who is also CEO of Aardvark. Ventilla was chosen by BusinessWeek.com as one of the most promising young entrepreneurs in tech this year.
Recommendations from Those in the Know
Here's how it works. Add Aardvark to a PC or smartphone buddy list, and then send it questions about restaurant, travel, and product recommendations. Aardvark sorts questions into categories, then combs through data culled from other users' Facebook profiles in search of the half-dozen people who might know the topic best. Answers come back within a few minutes. A private test version of Aardvark currently has more than 10,000 users.
There are other tools that fill a similar role, such as Twitter, which lets you send an information request to those who follow your short messages, or tweets. But unlike with Twitter, Aardvark's users don't broadcast their questions to hundreds or thousands of followers; the software breaks down questions and sends them to a small subsection of a user's circle of acquaintances. Subscribers can also specify times when they don't want to be bothered. "This is not a broadcast channel," Ventilla says.
Investors have clamored to buy stakes in Aardvark since the focused conversations that take place on the service provide a medium for companies to reach consumers when they're ready to buy. The startup has collected referral fees from sites including Amazon.com (AMZN) and online shoe store Zappos.com that those companies pay when Aardvark users embed a link to their sites in the answer to question. "Advertisers would pay a lot" to appear next to relevant questions, says David Hornik, a general partner at August Capital, leader of a group that invested $5.5 million in Aardvark in October.
Augmenting Traditional Search
Venture capitalists' interest in Aardvark demonstrates some investors' willingness to fund promising ideas and savvy entrepreneurs despite the drought in venture investing.
To live up to the high expectations, Aardvark needs to avoid being eclipsed by Facebook, Twitter, or Google itself, all of which are interested in augmenting traditional Internet searches with information supplied by users in nearly real time. "The reality is, anytime we fund something we think is interesting, there's someone else out there who thinks it's interesting," Hornik says. "If Aardvark is successful, there are companies that will try to emulate it, or acquire it, or crush it."
Socially oriented search engines also need to fight what Internet marketing consultant Andy Beal of MarketingPilgrim.com calls "participation fatigue" by users peppered with questions from people they may barely know. Social search companies such as Eurekster haven't caught on among large numbers of Internet users.
Plans to Make Aardvark More Effective
Some aren't as effective as traditional search engines like Google and Yahoo (YHOO) at finding answers that already exist on Web pages—say, a simple street address or a list of U.S. Presidents. "When we use Google for our searches, we're tapping into a vast array of computer servers and complex algorithms that offer an infinite willingness to provide answers to our questions," Beal says. "Search engines that rely on crowdsourcing run the risk that the crowds in question will eventually run out of enthusiasm to constantly help others."
Ventilla, who has an MBA from Yale University and spent two years at Google (GOOG) before departing in 2007, is paying careful attention to Aardvark's software, making sure users have enough controls to avoid getting bombarded with questions. In addition to mining Aardvark users' Facebook profiles for information about their interests and expertise, the company plans to incorporate information about users' Amazon purchase histories and Twitter activity to make the system even more relevant.
VCs waged a bidding war last summer to fund Aardvark. At least a half-dozen top-flight funds, including Redpoint Ventures and Sequoia Capital, were interested in striking a deal with Ventilla and his three co-founders. Eventually, Aardvark took a $5.5 million Series A round from August Capital, Baseline Ventures, Harrison Metal Capital, and technology entrepreneur Marc Andreesen. Aardvark issued a previous $2 million round of convertible debt to angel investors, including Google executive Joe Kraus, Stanford University professor Rajeev Motwani, and Napster founder Shawn Fanning.
Opening Up to More Users
What excites investors is the prospect that Aardvark is ferreting out a new kind of online search for information. The bet is that some questions—such as the best restaurant in a city for a romantic dinner, or the best camera to take photos of a kid's soccer games—lend themselves better to a straw poll of acquaintances than a full-fledged Web search. "There are lots of interesting questions that Google and [other] search engines do an increasingly bad job at answering," says August Capital's Hornik. "If you ask a person that question, they'd have that answer for you instantly."
As it prepares to open up to more users this summer, Ventilla and his co-founders will need to ensure their software can handle the task just as well, if not better.