Soso

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

~ the obituary ~

Tencent launched soso.com in march 2006, on a domain it had been sitting on since 1998 for an earlier attempt at the same idea. The plan was straightforward and looked unbeatable on a slide. QQ was open on essentially every desktop in China and Tencent controlled what that window pointed at. Search was where the advertising money went, so put the box in front of the audience and the audience will use the box.

For a while it worked well enough. Soso reached third place in Chinese search by 2008, with one Nielsen reading that year putting it briefly at second behind Baidu. It grew a full stack around itself. There was a question-and-answer service called Wenwen and a user-written encyclopedia launched on 30 march 2009, both aimed straight at Baidu's equivalents, plus a discussion search over Tencent's own forums. In december 2011 it added a street-view map product that was the first of its kind on a Chinese platform. By october 2012 soso.com was the eleventh most visited site in China and the thirty-third in the world.

None of that turned into the thing Tencent needed, which was people typing queries into Soso when nobody had put Soso in front of them. Baidu had the habit and habits in search are close to unbreakable, because the cost of trying the other one is nothing and the reward for trying it is also nothing unless it is clearly better. Soso was not clearly better. It was clearly there.

On 16 september 2013 Tencent put $448 million into Sogou for just over a third of it, and handed across the Soso search business as part of the same deal. The brand did not get a shutdown notice so much as a slow relabeling: by 19 november the results on soso.com carried a line saying they were provided by Sogou. The domain redirects now, and the company that owns the destination is the same company that owned the origin.

I have never typed a query into a Chinese search engine. This is the largest part of the web I have no standing to describe from the inside, and the honest version of that is to write down what happened and leave the texture to people who were there.

~ the case file ~

~ the biggest funnel in the country ~

The Soso story is the strongest available evidence that distribution and habit are separate things. Tencent had distribution that nobody in the industry could match, a messenger installed on almost every machine with a social network attached to it, and total control over what the default pointed at. What it did not have was a reason for anyone to prefer its results, and search is a product where the user re-decides every single time.

A default gets you the first query. The second query is a choice, and by 2006 that choice had already been made across an entire country. Every large platform that has tried to enter search from an adjacent position of strength has run into the same wall, and the ones that spent the most on it usually got the least back.

~ what tencent did instead ~

Buying into Sogou was the correct call and it took seven years to reach it. Sogou had an input method sitting on hundreds of millions of keyboards, which is a genuinely different kind of distribution, and it had a search product people had chosen on purpose. Tencent kept increasing its position until it took Sogou private outright in 2021, with the regulator approving the deal after blocking others.

So the ending is not really that Soso failed. It is that Tencent eventually worked out which part of search it actually wanted, and the answer was the traffic rather than the engineering. Paying cash for somebody else's index was cheaper than another decade of losing to Baidu with its own.

@}-->--

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

« previous  ·  random  ·  next »