Twitter + Nielsons = Bullshit

Last October, Twitter and Nielson’s launched their new Tweet-based rating metric. Scandal, Miley: The Movement, and Saturday Night Live topped the new ratings system. But the new rating system is bullshit. Eejoli Shah from Defamer nails it:

Twitter numbers, in contrast, include people who dislike the show—love or hate, as long as you tweet about it, you become one of the inflated statistic

This is why Twitter is both awesome and terrible. As a public IM service — twitter is amazing. I love watching friends hash out political issues. It works well as a source for breaking news — assuming you are following the right people. And there is some great comedy — also if you follow the right people.

On the other hand, it is a firehose of bots and spam. Spend any time on trending topics, suggested follow list, or the twitter website and it becomes clear that everything is for sale, and the only customers are bots.

It seems lazy that Twitter and Nielsons would not use any semantic filtering or sentiment analysis on a set of tweets. They are already parsing a massive amount of data, why not take the next step and add a quality metric? [1] The percentage of people making jokes about Sharknado far outweighs the number of people who watched the show. Are advertisers really so out of touch that the would by ad spots on a The Real Housewives of Little Rock because snarky 30-year-olds are hate-watching with their internet friends? I dislike the term “Hate-Watch”. But the firehose-of-data “Live Twitting” nature of Twitter seems to prompt people who might not watch a show to watch so they keep up with a friend or celebrity who may be watching. See “Liz and Dick.

Hate-Watching and Live Tweeting become a performance one puts on with and take parts with friends, internet friends and celebrities. The pleasure in comes not just from watching the show, but interacting with other viewers in real time. This can be an very powerful motivation for not time shifting a show. But it does not change the fact that one does not need to watch the show in order to take part in this performance. The volume of Sharknado tweets far out weighed the number of people who watched the show.

Basing their Twitter metric shows how disconnected Nielsons and advertisers are and how desperate the networks are to figure out how many people are watching.

Based on Twitter’s Q1 2014 earnings, its not working. Twitter says they are seeing a growth retweets and favorites. But in a quarter that includes the Olympics, the Super Bowl, The Oscars and the outbreak of political strife in easter Europe, the growth is far below what analysts expected.

Twitter will continue to play for relevance. They can continue to try to prove to investors and critics they are selling a mainstream product. But their long term success is tied to their ability to get advertisers to buy impressions from bots and assholes on the internet. I don’t think that is really a business plan.


  1. And this is not business speak bullshit. I have seen what semantic and sentiment data mining can do. I don’t worry about the NSA flagging me for New World Order jokes.  ↩