CNET también está disponible en español.

Ir a español

Don't show this again


Kango: Semantic search for travel

A new site searches thousands of travel sites to generate personalized recommendations for you.

Kango, launching in private beta today, is yet another travel recommendation site. This one's special sauce is its "understanding" of the travel lingo, so it can turn out better advice for you.

For example, if you're searching for a family-friendly hotel, Kango will return results that mention "kids." It will also give you hotels with pools, since it knows that families often go swimming on vacation. This type of word-mapping is becoming common in new specialized commerce sites (see also Snooth for wine, and Retrevo for tech products). The site also displays abstracts based on these word maps, and it extracts numeric or star ratings from sites that have them.

Kango collects ratings and reviews on travel destinations. In this example, I asked for information on romantic getaways in Maui, and the service extracted highly relevant information from several sites.

Kango is not a booking engine. For that, it redirects users to other sites, such as Travelocity, and snags referral fees as part of its revenue strategy. The site also makes money by selling ad inventory based on the search results.

The service scans about a thousand sites for content, Kango president Yen Lee told me. And it scans information on campgrounds--data not often in travel sites. In the future, Lee hopes to pull in information from various printed travel guidebooks, which would make the service uniquely useful. It would also help Kango to sell the guidebooks themselves.

I used an early version of the site (which covers California and Hawaii) that was a bit too beta for me to form a reliable hands-on opinion. I found the search results pages excellent. I also like the Item pages (for hotels, beaches, and so on), which display useful information as well as abstracts from several sites that mention the property in question. However, the searching experience isn't quite there in this version. I couldn't get hits on a search for surfing in Hawaii, for example.

Early snags aside, Kango's product philosophy makes sense. So does the revenue model. But where will the customers come from? This is hardly a green field market. Lee expects to get a lot of traffic from search engines, and that's a big bet. All you have to do is Google "family hotels San Francisco" to see the competition this site already has.