On #OccupyBoston, Occupy Boston, and Twitter’s Trending Topics Algorithm
October 11, 2011 in Media, Occupy Wall Street, Organizing
The ongoing saga of Twitter’s seeming censorship of activist trending topics got new datapoint in the wake of the arrest of a hundred Occupy Boston demonstrators late last night — the phrase “Occupy Boston” trended globally for a while this morning, while #OccupyBoston, with three times the total traffic, did not.
What’s curious about this is the fact that unlike the failed TTs I’ve discussed in the past, the trendlines for Occupy Boston and #OccupyBoston were quite similar before last night’s surge. In fact, the two terms rose almost exactly in sync yesterday until #OccupyBoston shot up as the arrests began after midnight. Now, #OccupyBoston had seen more traffic in previous days, but those raw numbers were tiny, and I don’t think they explain the failure of the tag to trend. Rather, I think the after-midnight spike holds the answer.
Why did #OccupyBoston spike so heavily as the arrests took place? Because of intensive, concerted efforts by OWS supporters to get the word out. And as I’ve noted before, Twitter (and it’s algorithm genies) hate intensive, concerted efforts — they don’t want anyone gaming their system but them.
But while #OccupyBoston was getting tweeted and retweeted by that (admittedly large and growing) core group, Occupy Boston was being tweeted and retweeted by a much broader and more diverse crowd — the kind of people who aren’t up on the latest hashtags. Those folks — journalists, local Bostonians, curious onlookers from all over the world — gave Occupy Boston a breadth of traffic that #OccupyBoston lacked, and the oomph to put it onto the trending topics list on a global basis.
I should make it clear that all this is an educated guess on my part, rather than established fact. Twitter doesn’t release much info about their algorithms. But I shook some Twitter folks out of the woodwork last year when this came up around #Wikileaks, and everything I’ve said here jibes with what I learned from them (and my own investigations) back then.
2 comments
October 19, 2011 at 6:06 am
Can an algorithm be wrong? Twitter Trends, the specter of censorship, and our faith in the algorithms around us « Social Media Collective
[...] it has been asked. Gilad Litan at SocialFlow (and erstwhile Microsoft researcher), spurred by questions raised by participants and supporters of the Occupy Wall Street protests, asks the question: is [...]
October 19, 2011 at 6:14 am
Scrutiny » Can an algorithm be wrong? Twitter Trends, the specter of censorship, and our faith in the algorithms around us
[...] it has been asked. Gilad Litan at SocialFlow (and erstwhile Microsoft researcher), spurred by questions raised by participants and supporters of the Occupy Wall Street protests, asks the question: is [...]