it's the panache, stupid

Retweeting in big numbers

It was two weeks into #OccupyWallStreet, and these two messages were posted on Twitter.

democracynow 9/28/2011 7:23 AM It "boggles the mind" that #OccupyWallStreet protesters are being arrested when no #WallStreet bankers have been arrested, says @MMFlint.

TheNewDeal 10/1/2011 10:28 PM 0 Bankers Were Arrested After Purposely Crashing Our Economy. Nearly 1,000 Have Been Arrested for Speaking Up About it. #OccupyWallStreet

The first was posted by @democracynow on the 28th of September at 7:23 in the morning. The second was posted by @TheNewDeal on the first of October at 10:28 in the evening. It is the same sentiment -- no bankers have been arrested and many have been arrested for protesting what the bankers did. (See also Boynton, 11/2/2011)

The two messages will be used to explore how messages may differ substantially in being retweeted. The sentiment seems to be the same, but the numbers are very different. "Boggles the mind" was retweeted 453 times between 9/23 and 10/01 when it effectively disappeared. "0 bankers" was retweeted 8,794 times between September 1 and September 17 when it effectively disappeared. The difference between 453 and 8,794 is the mystery.

One possible explanation is that @TheNewDeal has a much larger following than does @democracynow. Followers automatically receive messages that they may retweet. If you have more followers then you are likely to get more retweets. But that is not the case here. @TheNewDeal has 19,650 followers and @democracynow has 135,092 followers.

A second possibility is the difference in the time of day. "Boggles the mind" was posted early in the morning, 7:23 am, and "0 bankers" was posted late in the evening, 10:28 pm. We know that retweeting is normally something that happens quickly. Sysmos has shown that 92% of all retweets occur in the first hour. (Symos, 2010) So, if the timing was the reason for the difference in retweeting one would expect that there would be a significant difference in the first minutes after the original post. There would be fewer retweets in the morning and more in the evening. There is a difference in the expected direction. In the first ten minutes after being posted "0 bankers" was retweeted 94 times and in the first ten minutes after being posted "boggles the mind" was retweeted 73 times. The difference is in the right 'direction,' but it hardly seems sufficient to account for the large difference in retweets.

A reason for the very fast response is the organization of Twitter networks. Twitter users have followers. Their followers have followers. And their followers have followers. That produces the potential for a geometric explosion as the message moves from one set of followers to their followers by retweeting, and to their followers by retweeting, etc. A third possible explanation, therefore, is that @TheNewDeal may have fewer followers than does @democracynow but his followers may have more followers. So the move of the message from @TheNewDeal to followers would be followed by a much greater 'explosion' when the message moved to the followers of the followers. Since an 'explosion' [in time] is the logic I checked the followers of the retweeters for the first ten minutes. I checked the number of followers for each of the persons who retweeted the messages during the first ten minutes. The 94 persons who retweeted "0 bankers" had a total of 35,758 followers. The 73 persons who retweeted "boggles the mind" had a total of 42,195 followers. The difference is in the 'wrong' direction.

A fourth possibility is that this was a 'fluke' or was quite different from the usual pattern. Perhaps, except for this one message @democracynow and @TheNewDeal are retweeted approximately the same. To check this I looked at retweets per day for September 28 through October 7.

Twitter messages: @democracynow is dark, @TheNewDeal is light

@TheNewDeal's twitter messages are retweeted much more than are the Twitter messages of @democracynow. With two exceptions retweets of @democracynow are in the low hundreds. October 1 and the 5th are the exceptions. On the first there are almost 2,000 retweets and on the fifth there are something more than 2,000 retweets.

Only on September 28 did the number of retweets of @democracynow exceed the retweets of @TheNewDeal. That was the day "boggles the mind" was first posted to Twitter, and was an unusually low day for @TheNewDeal.

On October 2 @TheNewDeal was retweeted more than 6,000 times. It was the high point for "0 bankers" but it was also a 2,000 retweet day for other messages @TheNewDeal had posted.

"0 bankers" made for a spike in the retweets of @TheNewDeal. If the retweets for "0 bankers" is removed that gives a better indication of 'normal' retweeting of @TheNewDeal. For September 28 through October 7 the mean number of retweets per day, without the "0 bankers," was 1,704. The mean number for the period starting with "0 bankers" was 3,222. One, messages of @TheNewDeal are being retweeted regularly -- 1704 a day discounting the "0 bankers" spike. Two, a single message can have a substantial impact on the total -- with "0 bankers" it was a spike to 3,222 a day.

Messages of @TheNewDeal are being retweeted at a substantially higher rate than is the case for @democracynow. A reasonable ratio would be the number of retweets daily relative to the number of followers. The mean number of retweets for @democracynow, for the ten day period, was 679. The number of followers was 135,092, and the ratio is exceedingly close to zero. The mean number per day for @TheNewDeal was 2481, and the number of followers was 19,650. The ratio was .13 or 13%.

Every way one can compute it @TheNewDeal's Twitter messages are retweeted at a higher rate than are the Twitter messages of @democracynow. And differences in network connections does not seem to explain how the difference is what it is.

One more possibility -- There are two patterns in Twitter messaging that are easily discernible. One is the 'news move'. (Boynton, 10/20/2011) The other is the 'we move'. (Boynton, 3/11/2011) The news move is associated with the use of urls. The general form is "did you see that . . . http://. . .." The we move is about who we are, how we feel and is associated with retweeting. Twitter messages that are about what is happening, the news move, are more likely to contain a url and less likely to contain a retweet. The reverse is true for we move messages. They are less likely to contain urls and more likely to contain retweets. If @democracynow is news oriented and @TheNewDeal is oriented to who we are then we can expect that @democracynow messages will be retweeted less and @TheNewDeal messages will be retweeted more. One way to easily determine their orientations is to look at how they describe themselves in their Twitter profiles.

@democracynow: "An independent, global daily news hour. Follow us for daily news, events, and updates on Amy Goodman's speaking tour."

@TheNewDeal: "I am a Socialist Because I Like the Military, Police, Fire Dept, Library, Post Office, Medicare, Highways/Roads, Polio Vaccine, FBI, CIA, Student Loans, etc."

It would be misleading to assume that it is the statement in the profiles that makes the difference. It is the pattern of posting Twitter messages that makes the difference. @democracynow floods the Twitter stream with news -- this is what is happening now. And @TheNewDeal floods the Twitter stream with messages about who we are and how we are responding to events. The 'natural' outcome is that @TheNewDeal is retweeted as much more as 8,794 versus 453, which was the original mystery.

References

Boynton, G.R. (11/2/2011) How you know you are winning -- When a judge speaks

Boynton, G.R. (10/20/2011) The 'News Move' in Twitter Messaging

Boynton, G. R. (3/11/2011) RT @bobboynton new media and the revolting middle east

Sysmos (2010) Replies and Retweets on Twitter