December 5 Update | When I wrote this piece a week ago I was certain that Twitter wasn’t blocking Wikileaks from trending. Now I’m not so sure.
So the Wikileaks organization released the first batch of a promised quarter million US State Department cables today, and it seems like everyone on the planet is talking about it. The story is front-page news at media sites across the globe, it’s all over the television, and for a while this afternoon more than two percent of all Twitter traffic was about the leaks.
You read that right — one in every fifty tweets was about Wikileaks this afternoon. To put that in context, it’s about three times as many as mentioned Justin Bieber.
And yet “Wikileaks” hasn’t hit Twitter’s trending topics list all day. Right now #becauseofjustin, a Bieber-related hashtag, holds the top spot on Twitter’s global trending topics list, even though it’s running just one tenth the volume of #wikileaks itself.
Twitter being Twitter, there’s no shortage of speculation about why this is. Some suggest it’s a conspiracy of some kind. Others claim that because Wikileaks is a username on Twitter, it’s excluded from appearing on the trending topics list. (This is false, as a quick investigation will demonstrate.)
So what is it? Basically, it’s the algorithm.
It turns out it’s tougher than you’d think to put together a trending topics list that really means anything. If you just go by the raw frequency with which words appear, you’re going to wind up with stuff like “the,” “and,” and “RT” at the top of the charts forever. And even if you exclude words like those, you’re still going to wind up with “lunch” trending every lunchtime and Glee trending every Tuesday.
Which is fine if that’s what you’re interested in, but the folks at Twitter have decided that they’re interested in something else. What they’re interested in is finding out what’s breaking — what people are interested in today that they weren’t interested in yesterday. And to find that out you need to look beyond the raw numbers. You need an algorithm, and it needs to be a sophisticated one.
Because any time you create a system like this, there will be people who try to game it. Twitter doesn’t want its trending topics list to be a list of the day’s ten most successful “RT if you love puppies!!!!!” campaigns, and it doesn’t want the list — or its feed — to be clogged up with “Tweet about Velveeta to win a free iPad” garbage, either.
All of which means that Twitter doesn’t really care whether “wikileaks” was the most-tweeted word of the day. What they’re trying to measure is something more subtle and more complex. And they won’t tell us what exactly that is. They can’t tell us, because if they told us, we’d have a leg up in gaming the system again.
So why didn’t #wikileaks trend? It’s impossible to say for sure, but it’s most likely a combination of things. First, #wikileaks is a hashtag that sees considerable traffic on an ongoing basis, so it has to spike higher to make a splash than a less common word would. Second, a large portion of the traffic #wikileaks has seen today has been in the form of retweets, and Twitter gives retweets much less weight than original tweets in calculating trends. Third — and I’m truly guessing on this one — #wikileaks is a hashtag, and as a hashtag it necessarily reflects a co-ordinated, organized push to boost discussion of a topic, rather than an organic outpouring of interest. I wouldn’t be at all surprised to learn that putting a # in front of a word depresses that word’s weight in the trending topics algorithm, if only slightly.
Wikileaks rolled out a new hashtag today — #cablegate — and it quickly rose to the top of the trending topics list, despite producing far less traffic than #wikileaks. (That fact alone should quiet the conspiracy theorists, though it probably won’t.)
Some will take this last piece of info as a reason to spawn new hashtags every week, so that the tags’ novelty works in their favor. There may be something to be said for such an approach, but my own sense is that it’s counterproductive, for a few reasons.
For starters, trending topics are wildly overrated as an organizing tool. You constantly see people urging others to “get [whatever] trending,” or fretting about the fact that it’s not, but the reality is that having your hashtag show up briefly on a list in a sidebar isn’t going to do a hell of a lot for your movement. I’ll say more about why that is in a future post — this one is already ridiculously long.
Beyond that, there’s the fact of what you lose by switching hashtags — continuity and predictability.
Consider #wikileaks vs #cablegate. #Wikileaks drew huge traffic all day because people associated the tag with the topic, so using it came naturally to them. And they kept using it in the face of encouragement to switch to #cablegate. Even after #cablegate trended, in fact, #wikileaks stayed much more popular.
The point of hashtags — the point of Twitter itself — isn’t to get your tweets in front of random people. It’s to build a community of discussion. It’s to connect with people who are interested in what you’re interested in, and get them more interested. It’s to turn weak ties into stronger ties. And trending topics don’t have a whole hell of a lot to do with any of those projects.