Aardvark

» see vark.com as it lived, on the wayback machine

~ the goodbye ~

aardvark's peak years overlap with my late twenties on ttnet broadband, but the product itself never reached turkey in any visible way. it was a social q&a engine for the english-speaking web, the kind of thing tech blogs liked to write about while the rest of us kept asking our questions of plain old google search. when google itself bought aardvark in february 2010 for fifty million dollars, that acquisition headline was the first time a lot of us outside the bay area had even heard the name.

the pitch was strange and clean. you typed a question into a chat window and aardvark figured out which person in the extended network was most likely to know the answer, then routed your question to them as a live message. it plugged into google talk and the major im clients of the era. you got a human answer instead of a results page. that was the whole product.

it worked when the question was the kind google could not really handle in 2009. taste questions and local recommendations. the kind of question where you wanted a human in the loop rather than ten blue links and a sponsored result on top. aardvark was useful in a narrow strip and the founders, all ex-google, were betting that strip would widen.

~ the chronicle ~

~ how it landed ~

the team came out of google in 2007 under the name the mechanical zoo. max ventilla led it. they raised six million in series a from august capital after a 750k convertible round. the public version went live in march 2009. by late 2009 the product had a small but very tech-press-friendly base of users in san francisco who liked the gimmick of typing a question into google talk and getting an answer back from a stranger an hour later.

google bought it on february 11, 2010 for fifty million dollars. the obvious read on the deal was that google wanted the team back, with the product as a useful proof-of-concept rather than a real revenue line.

~ the google fade ~

inside google, aardvark stopped being a startup and became a feature waiting for a home. the home never appeared. by september 2011 google had decided to consolidate its non-core products and aardvark was on the list, alongside other things the company had quietly absorbed and quietly buried. the team scattered. ventilla later went on to start altschool, a different kind of failure entirely.

the funny thing is that the aardvark idea was not wrong. you can see it living on inside quora and inside the practice of just dm-ing the friend who might know. routing a question to the right human is still a real job. aardvark just tried to do it as a product, in 2009, on top of instant messengers that were already dying.

~ what got buried with it ~

the bigger thing aardvark hinted at was the idea that search did not have to be a page of links. it could be a conversation. fifteen years later that idea came roaring back as ai chatbots, except the human in the loop was now a model. aardvark had the human and was missing the model. it ended up on the wrong side of the timing curve and got fifty million dollars and two years of runway for the trouble.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »