Audience: the reach of political messaging on Twitter

Television audience is measured by the number of TV sets turned on. Reach on Twitter is measured rather differently. Anyone may sign up as a follower of anyone with a Twitter account. Becoming a follower means all of the messages of the person being followed go directly to your Twitter account. The number of followers of the person writing a Twitter message is the first audience for that specific message. As with the TV set that is turned on there is no guarantee that someone is watching the TV nor that each follower reads the message. There is a secondary audience for Twitter messages, however. Search is frequently used to find messages about a subject of interest. In general, those 'hits' do not leave a trail. Given the prevalence of search in Twitter communication it seems the estimate of individuals reading a message might be about as many as the number of followers; for every follower who does not read the message someone finds it and reads it without leaving a trail. At least that could be considered a first approximation of reach.

The 'first audience' can be quite impressive. Mr. Obama is the leader for politicans with 5.7 million followers as of October 15, 2010; the number of his followers increases daily. How far ahead of other politicians he has become is clear when you learn that Sarah Palin has only 277K followers.

Instead of approaching audience via individual followers this report looks at the results for 11 specific searches. One search was for "obama AND mosque." It started on August 13 and closed on the 27th. It was the period immediately after President Obama made his statement about the right to build a Muslim community center in lower Manhattan. The search was run continuously during the 14 days, and 31,985 messages containing those two words were found. The finding reported here is about the number of followers of the persons producing each of those 32K messages.

The table gives the number of tweets, the total number of followers for the tweets, the average number of followers, and the number of retweets. All of the searches were conducted in 2010.

There is no plausible way to take a sample of political messages circulating through Twitter. About the best one can do is variety, and this set of searches has variety. #iamthemob and #rino are streams of messages of alienated persons; they are alienated from the way both Democrats and 'standard' Republicans do politics. For example, their favorite rino, Republican in name only, is John McCain. Two collections were about large controversies: the nomination of Kagan for the supreme court and Obama's statement about the community center. Financial regulation and jobless benefits were two pieces of legislation passed by congress during 2010. Two were about courts. The stream about Arizona and suit was about the suit filed against the Arizona law dealing with illegal aliens. The stream about Miranda concerned a supreme court decision requiring that suspects explicitly tell the police that they wanted to remain silent. The FDA approved a morning after pill that would work for up to five days, and that precipitated a small stream. And the smallest stream was about the report on economic growth for a quarter.

They are equally varied in number of tweets ranging from 44K to 353.

The reach is impressively large in all cases. Take the smallest. There were only 353 tweets about the economic growth for the quarter, and there were 798,530 followers of the 353 individuals. Of course, 798K pales in comparison to 129.9 million and 84 million. Thirty-five thousand tweets had a reach of 129.9 million, and 31.9K reached 84 million. All of the numbers are impressive: 894 has a reach of 1.4 million, 13,780 reaches 62.4 million.

In addition to the very large audiences these numbers capture there are two additional ways the numbers are quite different from ordinary Twitter practice.

RJ Metrics published an analysis New Data on Twitter's Users and Engagement on January 26, 2010. The data was collected in the months toward the end of 2009. They reported the number of followers for their sample of users. The figure shows their results.

Only approximately 1% of Twitter users had over 500 followers. If you compare that to the averages for the 11 searches of political streams there is a dramatic difference. The average number of followers ranges from 1414 to 4531. Following is more standard practice in political messaging than for all users of Twitter.

Sysmos, Inc. published Replies and Retweets on Twitter in September 2010. It was based on a collection of 1.2 billion tweets. They found that only 6% of the Twitter messages in their sample were retweets. There are many more retweets in political streams. The smallest percentage is 15% and most are 20% or 30% retweets.

Tweeting about politics reaches very large audiences. They rival the reach of TV news; Beck and O'Reilly, for example, average about 3 million viewers apiece. The communications are not comparable, of course, but an audience of 3 million is not particularly impressive in political tweeting. As important as the total numbers are the relationships in this communication domain are equally important. This is a much more connected domain of communication than is true for standard practice in Twitter. There are many more leader-follower relations, there are many more retweets than standard practice giving a clear indication that politics on Twitter is an active and inter-connected communication domain.

© G. R. Boynton
October 17, 2010