Stealth Startup Kango Working on Semantic Search For Travel

Does the world really need another travel site? With Expedia, Orbitz, Travelocity, TripAdvisor, Farecast, TripHub, Yapta, and many more, prospective travelers already have more than enough to choose from. (With an estimated $87 billion in travel booked online, it’s no wonder why). Soon they will be able to add Kango to that list.

The startup, which has been in stealth-mode until now, does manage to add a new twist to search travel. It is indexing 18 million opinions and reviews across 1,000 travel-related sites to derive the best travel search results based on what type of trip you want to take. If you are planning a romantic getaway in Big Sur, you will get one set of results. If you specify that you are looking for a family outing instead, you will get another. Or you can look for pet-friendly hotels and activities. Of course, you can also search by price or amenity, like any other travel site. And you can see where each hotel or activity is located on a small Google Map.

picture-204.pngBut what’s promising about Kango is the way it slices up search subjectively. Kango is building a semantic search engine focussed narrowly on travel. It parses the language in all of those reviews and guides, and categorizes them by generating tags for them. “You cannot wait for users to add tags, you have to derive them,” says CEO Yen Lee. So hotels that have been reviewed across the Web (on sites like Yahoo Travel, TripAdvisor, or Yelp) with words such as “perfect,” “relaxing,” “couples,” “honeymoon,” or “spa” would rank higher in a search for romantic travel. Hotels associated with the words “kitchen,” “pool,” and “kids,” would rank higher in a search for family trips.

Whether this will be enough to draw people from other travel sites is hard to say at this point. But Kango’s executive team has an impressive pedigree. Lee is a former general manager of Yahoo Travel. His search architect, Huanjin Chen, used to be the search architect at eBay. His natural-language search scientist, Boris Galitsky, used to do work for the British government. And his head of marketing, Elliott Ng, headed up marketing for Intuit’s QuickBooks and is the founder of Netcentives.

Lee estimates there are 6 billion to 8 billion travel-related searches done every year, and he thinks Kango can help expose more of the hidden gems in travel that today don’t quite make it to the first page of most travel sites. He plans to make money on travel-specific search advertising, rather than on booking or listing fees.

Kango will be rolling out a limited beta in the next few weeks, and is reserving 100 spots for TechCrunch readers who sign up here.

picture-223.png

hotel-detail-page-cropped-sand-dollar-10607.jpg

Attendi Wants to Search Inside Your Head

As if the more than 20 billion Web pages out there aren’t enough, a new startup coming out of stealth mode today called Attendi has come up with a new twist on “people search.” This is not to be confused with the type of people search that Facebook is getting into (actually searching for people—see also Spock, Wink, Zoominfo, WikiYou and PeekYou), or the type of people-powered search results that Mahalo, Wikiasari, and others are exploring (also known as social search). Actually, Attendi could more aptly be called chat search because it wants to search what’s in people’s heads as expressed through online chats. Attendi is launching at DEMO fall.

Attendi is half a social network, and half a knowledge database. Here’s how it’s supposed to work. Members, known as “Attendis,” will create profiles on the site describing their areas of expertise, hobbies, and interests, as well as adding links to their blogs, social networks, or simply Websites they identify with. The site, which opens in beta today, dynamically creates tags that define what each person knows and cares about (they can also add their own tags). It is built on top of the Lucene open-source search engine, and the Jabber instant-messaging protocol.

When someone searches for a topic on Attendi, what comes back as results are profiles of other “Attendis” whose tags match the search request. And if they happen to be online at the moment, even on another IM system, the other person can initiate a chat discussion with one of them to ask questions about that topic. “Attendi will just be a way to broker your availability,” says CEO Drew Rayman. Every chat is archived, indexed, and becomes fodder for future search results.

Attendi is based in New York City, and Rayman is also the founder of an interactive ad agency called i33. He plans on selling search ad sponsorships based on Attendi topics, as well as a live chat ad unit that only pops up when a company’s online customer service rep is at the ready to do a hard sell through IM. It’s that kind of in-your-face advertising, though, that might drive people away from an IM-centric search engine and never give it a second chance.

Making topic-specific IMs searchable is certainly a novel way to create a knowledge database. But Attendi faces a huge hurdle in getting anybody to actually use its system. What’s the incentive? There is no existing network of super-smart Attendis anyone would want to tap into. One way around this chicken-and-egg problem, though, would be to take advantage of free advice that tens of thousands, if not millions, of people are already giving away for free online in the form of comments that people leave on blogs. There is no easy way to search across those. (Startups like Big Swerve, which was in the TechCrunch40 Demo Pit, are already onto this).

Bloggers today install search boxes from Google or Eurekster to allow readers to search through their posts. Why wouldn’t they also install a way to search comments if it were available. Attendi would be better off trying to build such a service to gain traction for its technology. It could offer a way to power comments for blogs that would make those comments searchable both on that blog alone and across all Attendi-powered blogs. That way, those people who leave their opinions across the blogosphere in the form of comments would surface in Attendi search results. Tapping into blog comments would be a great way to seed its knowledge network.

Real Time Search Soon At Twitter

We just got word from Twitter CEO Jack Dorsey that the company will be launching a real time search feature “very soon.” Type in a keyword or keywords and any time a Twitter is created that includes those keywords, you’ll be notified via IM or SMS. They’ll add the the ability to access this via their API in the near future, Dorsey says.

The feature hasn’t launched yet, and we’re unsure on some of the details. For example, I’m not sure if you can limit results to just people you are tracking, or if you can get results via RSS or email instead of or in addition to SMS and IM. Either way, Twitter is clearly adding features and functionality at a quick pace, which is good to see. Hopefully, the worst of the downtime is behind us, too.

The fuzzy image to the right is a sample user interface that we received from Twitter.

Update: Dorsey says in the next couple of weeks they’ll allow historical search as well. Also, for now, searches can only be done on all Twitter users. Over time, users will be able to refine searches by friends, geography, time and/or language. For now, results are only returned via SMS and IM - RSS, email and the API will come later.