TV by the Numbers

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

~ the eulogy ~

Robert Seidman and Bill Gorman started publishing in September 2007 out of Calabasas, California, on a plain premise. Nielsen overnight ratings were public and so were the network press releases, and almost nobody was reading either of them carefully in front of an audience. The site posted the numbers every morning and then did the part the numbers do not do on their own, which is say what they mean for a show you like.

Gorman's Renew/Cancel Index was the thing that made it matter. It compared a scripted broadcast series against its own network's average rather than against the whole schedule, because a show is renewed relative to what else that network has. Seidman ran a looser companion called Bubble Watch for the cases where the formula would be too confident. The Cancellation Bear was the mascot attached to the verdict, and it got quoted in enough places that people who had never visited the site knew what it meant when it showed up.

Nielsen leaned on them in 2009 over how much of the archive they were keeping. Zap2it partnered with them in November 2010 and Tribune Digital Ventures bought them outright in April 2014, which is how the site ended up inside Nexstar's media group. On 31 January 2020 the staff posted a goodbye and the site went inactive the following day. The URL now redirects to Zap2it's listings.

None of this measurement was ever about me. Nielsen counts American households through a sample of American set-top boxes, so watching an American show from Ankara contributed exactly nothing to the number that decided whether it continued. I was not in the panel and there was no version of me that could be. What reached this side of the world was only the outcome, a series that had a next season or did not.

~ the dossier ~

~ what the arithmetic was for ~

The reason a ratings blog developed a fanbase is that it answered the only question viewers actually had. Nobody watching a show on the bubble in March wants a demographic breakdown. They want to know whether to get attached. TV by the Numbers gave a straight answer with the reasoning shown, months before the network would say anything, and it was right often enough that people planned their television around it.

The 18 to 49 demographic was the center of it, and that is where the fights started. Kathy Bates went after the site over Harry's Law, which was pulling large total audiences with an older skew, and the show's executive producer said publicly that they knew nothing. He was arguing with a mirror. The site did not invent the rule that advertisers pay for the younger half of the audience, it just published what the rule was doing to his show.

~ the arithmetic stops working ~

Every part of that model assumes a published number. Broadcast overnights existed because advertisers were buying against them and everyone in the business needed a common measurement, so the figure came out every morning whether the network liked it or not. Streaming had no such requirement. Netflix canceled things on the basis of internal numbers it declined to show anyone, and there was no formula to build against a secret.

So by 2020 the site was analyzing the shrinking half of television with great precision while the half that mattered stayed dark. The closing note said as much. The thing they were good at had not stopped being true, it had stopped being where the shows were.

~ what it proved ~

Two people with a spreadsheet and public data built the most trusted renewal predictor in American television, and they beat trade publications with sources inside the networks. That is worth stating plainly, because it is the argument for open measurement in one sentence. When the number is published, anyone can check the work.

What replaced it is the current arrangement, where a platform announces a top ten list it computed itself and journalists write it up. Nobody now can tell you why a series was canceled, and there is no bear.

@}-->--

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

« previous  ·  random  ·  next »