fflick

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

~ the eulogy ~

fflick's whole life sat inside a five month window in late 2010 and early 2011. its public years overlap with the part of my online life that was already on twitter, but i was using twitter the way most engineers in turkey used it then: a follow list of foreign tech accounts and the occasional turkish journalist. movie recommendations were not something i was asking the timeline to solve.

kurt wilms and three other former digg engineers, including ron gorodetzky, started fflick in august 2010. the product was small and well defined: sign in with twitter, see a list of current movies, and read tweets about each film bucketed into positive sentiment and negative sentiment. the underlying piece was a sentiment classifier trained on twitter language about films. the founders said publicly that movies were just the first vertical and the classifier was the actual asset.

google announced the acquisition on 25 january 2011 for a price reported at around ten million dollars. the team moved into youtube. the fflick site itself was shut down shortly after. there was no period of fflick living inside google as a branded product. it was an acquihire end to end.

~ the dossier ~

~ how it landed ~

the launch coverage was friendly. fflick fit a specific 2010 moment where social-graph sentiment was a fashionable startup pitch and where twitter's firehose was still accessible enough that small teams could build on it without rate-limit pain. the design was clean. the founders were known from their digg years. the bet inside the company was that classifying short-form opinion language was a general capability with movies as a tractable demo.

~ the actual end ~

google announced the deal as a youtube move. the framing in press coverage was that youtube wanted the sentiment engine to help with comment classification and trending-detection inside its own product. there was no public retention plan for the fflick brand. the consumer site was end-of-life on the same announcement.

at the time it was unclear if any of the sentiment work shipped into a visible youtube feature. youtube did expand its trending and comment-ranking systems over the next year but never credited a fflick lineage publicly. the standard acquihire outcome: the team gets integrated, the brand disappears, and the original product's surface area is folded into a much larger codebase.

~ the long tail ~

kurt wilms stayed at google for several years before moving on. ron gorodetzky did the same. the sentiment-on-twitter category as a whole had its peak between 2010 and 2014, then thinned out as twitter restricted its api and as the brands that wanted this capability built it in house or bought the larger players. topsy was the bigger version of the same bet and apple bought it in 2013 for about two hundred million. fflick was the smaller, earlier, cleaner version that disappeared into youtube without leaving a footprint.

@}-->--

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

« previous  ·  random  ·  next »