Sunday, January 06, 2008

Kango Sessions: Interview with with Kango CEO Yen Lee

Back in September I reported on the rumours of Kango having raised money to set up a company to "help family's plan travel through organisation tools, reviews and editorial content". Just before the end of the year we received confirmation that Kango had in fact raised a very respectable $4mm in series A funding from the web 2.0 loving Shasta Ventures (also backers of Techcrunch40 winner Mint, Turn and Flock).

We also saw the beta for Kango and found out it is meta-search product but from a different angle from what we have seen. In the past we had two types of meta-search:
  1. the free form search players such as the now combined Kayak/Sidestep where the interface looked like an online travel agent but with the back end fulfilled through aggregated search; and
  2. the advertising/deal hunter driven model such as Travelzoo and Cheapflights.
Kango's angle is to be the aggregator of content, reviews and travel information - not deals. Destination, hotel or activity searches on Kango produce a result that takes all of the unstructured data out there from web based sources (TripAdvisor, Travelpost, Expedia, Orbitz, Travelocity, etc) and combines it into a structure data result. In other words aiming to be a contextual summary of all of the editorial and user generated content relating to a travel term. Is a dramatic contrast to the manual page build approach of the high profile Mahalo.

I had a chance just before the end of 2007 to talk with Kango President and Founder Yen Lee about Kango, meta-search, raising money in online travel and how to deal with "too much information". Here are some of the themes and ideas from that conversation.

How did you come up with the idea?

Yen and his co-founders are search junkies. They love search, need search, trust search. Yen told me that the idea for Kango came from a desire to develop a targeted search product for a sector with lots of unstructured data, a sectoral ontology, vertical structural options and a need for subjectivity (ie editorial contribution). That is a sector with too much information - hence they naturally ended up in the travel sector.

What is the background of you and your founders?

Alongside Yen (ex-Yahoo), the other founders are Elliot Ng (ex-Necentives) and Gene McKenna (ex-
Acxiom Digital and Bluedot Software). Experience in search, customer loyalty and product. More on the founders here.

How hard was the product to build? How much human intervention is needed to make the unstructured data structured?

The content is accessed through the same methodologies as general search - crawling sites and normalising/structuring the data. At the content access and search level they operate just like the traditional search market. It is at the display and refinement level that Yen claims Kango is different. Instead of relevance (and therefore ranking) being based on inbound links it comes from a number of weighted dimensions based on the link between keywords. For example "spa", "private", "hideaway" in hotel reviews are linked to searches for "Romantic". Similarly "kids club" etc is linked to "Family Friendly". The relevance and ranking comes from the quality of this semantic analysis, not inbound links.

Needs editorial input to provide the knowledge that links search terms to themes, destinations and activities. All about going after what Yen put as 10 billion travel related searches a year.

How hard was it to raise money? How are you going to make money?

Yen made the raising of $4mm sound easy. Says they were able to raise the money before they had a product - that is based on the team and idea only. This is an uncommon story in travel 2.0. As to making money - Yen and his team are hoping that by giving an meta-index to travel content will give Kango ownership of the long tail of travel search. From this the cash will flow from travel advertisers.

What do I think

Talking to Yen was like lifting the bonnet on a search engine and looking under the hood, seeing how the words, algorithms, spiders and links combine to generate the answers that Google and Yahoo churn out. The search knowledge of the Kango team is clear. Also the product looks good. Search and search refinement are easy - modelled more like an OTA than a search engine. The big challenge is that to build on the travel ontology and develop the semantics takes time and manual energy. This why the beta is limited to only two destinations. This is why they need a round of $4mm to fund the building of more destinations and product lines. This is a case where if they can get the product out right, the traffic will flow (easy) and the money will follow the traffic (even easier). Very impressive stuff.

1 comment:

Anonymous said...
This comment has been removed by a blog administrator.