This week, my friend Gilad Lotan of SocialFlow wrote an excellent blog post explaining how trending topics really work, in an attempt it seems, to put rumors of censorship to rest. Twitter has been dealing with these rumors for quite some time, and last December publicly explained that Trending Topics are about velocity, not volume, but their blog post (either because no one read it or no one cared) has done little to silence the calls.
When I was at ONI, we analyzed a number of trending topics that were controversial in some way or another and that people were claiming had been censored, including #Flotilla and#Wikileaks. In both cases, we found that, while the problem was likely algorithmic, it did seem odd that Twitter would choose not to interfere and set a genuinely newsworthy event to trend. Nonetheless, none of us felt it to be ‘censorship.’
Despite my work (or perhaps because of it), I find those calls to be somewhat obnoxious. First off, Twitter is a private company that has been transparent about their algorithms, like it or not. And they’re not denying you your right to speech, they’re simply not trending your speech. That’s like complaining about being left off Twitter’s recommended user list (*cough Scoble cough*).
I think there’s a question here though: How much do Trending Topics actually matter? First off, they have to represent one of two things: Either an intentionally-created hashtag (usually created with the purpose of trending) or a genuinely popular person or thing at that moment (e.g., a celebrity who has just died). In the case of the former, I think it’s a legitimate method of trying to get attention for a cause, but I don’t know that getting a hashtag on the sidebar matters all that much in the long run (I’ll elaborate in a moment). In terms of the latter, I think it’s certainly important – I wouldn’t have known when Farrah Fawcett (of Charlie’s Angels fame) died if it weren’t for spotting her name on the sidebar. Seriously.
But back to hashtags – As far as I can recall, #syria never really trended, but that hasn’t stopped people from using it, searching it, saving it as a search term, and following it. On the other hand, the ability of a well-worded tweet to go viral is proven: I once tweeted a pithy statement, only to find it retweeted 22,000 times.
I think that raises a bigger point: Trending Topics do not imply quality of content. There are certain hashtags full of valuable content all the time (think of some of the smaller, humor-focused ones, or even #ows or #occupywallstreet when they first started). Then there are other hashtags repeated ad nauseum for the sole purpose of getting the hashtag to trend – The #GiladShalit campaign and the #Gaza solidarity campaign have both done this at times, for example, filling up single tweets with only the hashtag. If I spot an interesting hashtag and click on it, only to find a bunch of garbage, chances are I’m not going to check back. On the other hand, if I spot a non-trending hashtag filled with quality tweets, I might save it as a search.
My point is that people complain about censorship, and there are certainly some valid criticisms of Twitter’s algorithm (I don’t like, for instance, that it’s different for different cities), but without much analysis of whether Trending a topic is actually that useful.
I would also add that I’ve been at several conferences where the hashtag of the conference trended locally – which really only implies (given that there were fewer than 500 people in attendance) that the cities in which those conferences took place had a low threshold for trending topics.
Lastly, a comment: Twitter does in fact censor profanity from the Trending list. Whether they did this with Jeff Jarvis’s #fuckyouwashington remains unknown, but I do find it rather absurd that a company that prides itself on free expression is for some reason abiding by unrequired FCC guidelines.