RJ Metrics provided data analyses for Twitter last fall and again early this year, with the latest data running through the end of 2009 and updating the information for users and engagements. It’s very valuable insight on trends and potential audience on this major social networking venue and while I urge you to read the latest report at your leisure, here’s a summary:
There’s good and bad news in the topline results for the end of 2009. Total Twitter accounts are up to a little over 75 million. While new accounts were still being added at about 2-3 per second, or 6.2 million a month, that’s down about 20% from July 2009, which was the peak for new accounts.
40% have created an account but never tweeted, 25% have no followers, and 80% have tweeted less than 10 times, which means most people who have joined are inactive or barely active. Also, on a percentage basis, December 2009 was the lowest month for activity in Twitter history, as only 17% of accounts sent anything. However, it’s important to remember that this is 17% of the growing base of accounts, and further analysis indicates that those who come back become “more engaged” over time (see below).
The results include a cohort analysis of Unique Tweeters to understand their behavior and loyalty over time. Essentially, it resembles CPG Panel Trial and Repeat analysis and groups them into the week they first Tweeted (Trial) and then looks at when and how often they return (Repeat). About 20% of those who actually send come back and remain active, however, once grouped by month and looked at total content sent (not just whether they sent at all), that 20% becomes more active over time, and, as the report notes, “tweet so much that it makes up for all the people who left”.
The results point out that if 20% of 75 million accounts are active users, that means there are around 15 million people out there who are “highly active tweeters”. To put that in perspective, take a quick look back at the average ratings and viewership for the 2009 NBA Finals on ABC, which, according to the NBA’s press release, “dominat(ed) prime time”. The average rating was 8.4, with 9.6 million Hhds and a little over 14 million viewers. That would put average viewers for the Lakers vs. the Magic below the number of truly active Twitter users.
Data sources: Twitter: Compiled by RJ Metrics, via Twitter’s application programming interface (API) data, in December 2009. Accessed 2 million Tweets from 50,000 users to compile the analysis. NBA ratings: NBA PR release, June 16th 2009
