Nicowari Survey

» see nicovideo.jp as it lived, on the wayback machine

~ the eulogy ~

Nico Nico Douga had a mechanism called nicowari, short for Niconico interruption, which could cut into the player on every machine watching the site and put something else on the screen. It was built for advertising and for announcements. On 29 July 2008 Niwango pointed it at the audience instead and asked them five questions.

The first Nicowari Survey mixed the trivial with the civic without apologizing for either. One question asked which of the Nico Nico Douga icons you liked best, and two questions later you were being asked what you thought about politicians using Nico Nico Douga to put out their opinions. Five answer buttons per question, a click to advance, and a summary screen before you submitted. 86,225 people answered. The results went up as a video roughly fifty minutes later.

What Japan ended up with was a polling operation grown out of a video site, and no country I have lived in produced anything like it. Ahead of the February 2014 Tokyo gubernatorial race a Niconico survey pulled 906,000 responses, a sample size national newspapers do not reach, taken from a population that had self-selected into an anime and vocaloid website. Both of those descriptions stayed true for six years, and Niconico went on publishing the numbers.

~ the dossier ~

~ ninety seconds ~

The window was ninety seconds. That is the whole instrument. You had a minute and a half to read five questions and answer them, inside a pause in whatever you had chosen to watch, with no way to come back later and no way to have thought about it beforehand.

Asked in that first survey how often they would be willing to do this, a third said they would turn up for a theme they cared about, and slightly fewer said once a month. Niwango took the hint and kept the cadence occasional rather than constant, which protected the thing that made the numbers interesting.

~ no warning on purpose ~

At the Nico Nico Douga summer version launch event on 4 July 2008, three weeks before the first poll, Niwango director Hiroyuki Nishimura explained what the surprise was for: you can get results with almost no organized bloc voting in them. The surveys were not announced beforehand, so the only people who could answer were the ones already watching, and there was no window in which anyone could be rallied.

Every open web poll of that era was a contest in who could mobilize fastest, and the usual defenses were registration walls and after-the-fact filtering. Nicowari Survey solved it by making the poll unfindable until it was already happening. The cost was a permanent sampling floor: you never reached anybody who was not logged in and watching at that second. At least that bias held the same shape on every run, which is more than most open polls of the period could claim.

~ the system that replaced it ~

By 2014 Niconico was no longer a video site with extras attached. Live broadcasting and paid channels had their own audiences who might spend an evening on the site without ever opening a video, and the nicowari mechanism could only reach a person with a player running. On 21 November 2014 the old system ran for the last time and a rebuilt Niconico Survey took over, able to put the same questionnaire in front of users of any of those services at once.

The handover survey ran from noon on 21 November to noon on 24 November and collected 491,052 answers, with a live program on the 28th to argue about what they meant. The successor was better at everything except the one thing the original had been: a poll you could not see coming, that ended before you finished deciding whether to lie.

@}-->--

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

« previous  ·  random  ·  next »