So I’m taking a day off from Twitter, in solidarity with #womenboycottTwitter. It was an easy decision to make, because Twitter is objectively awful on several levels. It’s a performative rage-machine with a distressingly bad signal-to-noise ratio that enables racists, fascists, harassers and other sociopaths and punishes the targets of same. In a sane world we wouldn’t put up with it, but we do—and Twitter’s management is counting on that—because we think Twitter is somehow necessary, in terms of getting our message out, in terms of sending traffic to our websites …

Yeah, about that. It turns out that Twitter is terrible at sending traffic to websites. Despite all the noise and the rage, we tend not to click on the links attached to tweets.

To see what I mean, here are some traffic stats from The Map Room, my map blog. New posts on The Map Room are automatically posted to Facebook, Google+, Twitter and Tumblr (and someone’s even imported the RSS feed into LiveJournal), but there are share links at the bottom of each post as well. So where does most of The Map Room’s traffic come from?

Figure 1. The Map Room’s inbound traffic referrers

Google Search and Facebook, mainly. Social media traffic is down in 2017, but that’s because I’ve posted a lot less this year than I did in 2016, and inbound traffic is more or less correlated with how much I post. Google Search traffic is not just up proportionally as a result; so far this year it’s already higher in raw numbers (20,318) than it was for all of last year (15,159). That’s interesting in and of itself, but what I want to talk about today is the ratio between Facebook and Twitter: both this year and last, I’ve been getting roughly twice as much traffic from Facebook as I do from Twitter.

That’s all the more interesting when you take into account that The Map Room has twice as many followers on Twitter than it does on Facebook. As of this moment The Map Room has 6,111 Twitter followers, 3,015 Facebook Page likes, and 2,949 Page followers.

Figure 2. The Map Room’s Facebook and Twitter followers vs. visitors

Which is to say that Twitter needs twice as many followers to send me half as much traffic as Facebook does.

That’s even more astonishing when you consider that Facebook’s algorithms tend to throttle page views when they pick up on words like “buy” in the text of the post—so that they can make you pay to boost the post to be seen by more viewers. Even with those hijinks, I get twice as much traffic from Facebook than I do from Twitter, even with half the followers.

So what’s going on here?

I can only make some guesses, but this probably has to do with the relative quality and size of their respective user bases. Facebook has five times as many users as Twitter does, and they’re much more likely to be real human beings. So when one of The Map Room’s Facebook followers likes or shares something I’ve posted, my post benefits from Facebook’s stronger network effect, even if it’s starting from a smaller follower base.

And that doesn’t even take into account whether Facebook users are more likely to click on links than Twitter users are. I suspect they are, but that’s just a hunch. (Are Facebook users more likely to click because they’re bored at work and killing time? Are Twitter users getting short fast hits of adrenaline and outrage on the go?)

It’s not just link-clicking. Anecdotes are not data, but I can’t help but notice that when I cross-post something to my social networks from my personal accounts, I get fewer responses from Twitter than I do from any other network except Tumblr (where I have a handful of subscribers in either case). When I share a cat picture from Instagram, I get more likes on Flickr than I do on Twitter. A new blog post sometimes gets more feedback on Google+.

Both Facebook and Twitter intermediate themselves between publisher and reader, but Twitter’s algorithms have seemingly curated out all of the signal and left us with the noise, making it extremely difficult to keep up with your feed and forcing people to tweet and retweet the same thing over and over again (“ICYMI: ”) so it’ll be seen. And Twitter’s focus on brands means that there are a lot of institutional presences there that are interested in disseminating information but not in conversation. The end result: we’re shouting into the void, and no one is listening.

All of which is to say that Twitter, at least as a source of traffic, is a lot less important than Twitter itself would like for us to believe. Based on the information I have, I’d be much more willing to spend money to promote a Facebook post than I would a tweet: I’d get more value for my money. And simply walking away from Twitter would hurt a lot less, traffic-wise, than I originally thought.

If my data are in any way typical, that can’t be good in the long run for Twitter. Because if their objectively terrible experience is not in some way mitigated by usefulness, there’s a real risk that the people who walk away will start to include the ones who would otherwise be paying to keep Twitter’s lights on.