Vanity Fair published an interesting piece on Donald Trump’s bot problem: he claims 22 million followers, but how many of those are fake?
As it turns out, a lot: By the estimation of Status People, a site that can look at the followers of a twitter account and evaluate the demographics of the followers, 41% of Trump’s followers are fake (lest you think I’m bashing Trump here, I’m not: Clinton’s numbers are similar. These fake followers are easily and cheaply bought and standard practice with many large accounts to boost their perceived follower base).
But no, this piece isn’t about Trump, or politics. It did get me curious about the scope of fake followers, though, so I hooked my account up to the service and had them evaluate my account. The results are not pretty:
According to Status People, 11% of my followers are fake. For the record, I’ve never bought or solicited fake followers, so these are all bots that have followed me on their own initiative. This doesn’t surprise me, because I see these fake accounts follow me when I get the notifications, typically right after I post something — so clearly they’re randomly surfing the public streams and randomly follow people to be less obvious about being bots. I’v also noticed they tend to randomly retweet postings at times, all, I think, in the name of trying to look real.
But look at that other number
That’s not why I’m writing this, though, it was the other number that caught my eye: 42% of my followers are inactive, which means that they haven’t posted a tweet in at least 100 days.
Now, some of those inactive accounts might still be quietly watching, but I think it’s safe to say the vast majority have just given up on twitter. Basically, half my follower list isn’t there.
Tap. Tap. Tap. Is this thing on?
The bigger picture: how badly is Twitter in trouble?
Here’s what interests me about this:
Last quarter, Twitter again missed Wall Street’s numbers, and its stock got nailed. Pretty much the only bright spot in their quarterly results were new user growth at 310 million, and they claim a growth in active users of about 3% year over year.
How much of that 310 million new users are fake and/or bots? I haven’t really seen any discussion about how Twitter’s fake account problem is impacting the numbers being used by the market to set Twitter’s stock price.
But if 10 million of Trumps followers are fake, and ten million of Clinton’s followers are fake, the numbers of fake accounts starts becoming statistically significant pretty quickly.
Or think about this: 11% of my followers are fake: that’s roughly 300 of my 2700ish followers. My account is, I think, pretty typical of a generic user’s account, so can we imply that 10%(ish) of followers on most typical accounts are fake?
How many is that? And how many of those, since they’re following and retweeting to make themselves look real, are showing up in Twitter’s monthly active user numbers? And how many of these fake account creations are showing up in that 310 new users every quarter?
I’ve felt for a long time that Twitter’s done an astoundingly bad job of tracking down and closing down these fake accounts, and I couldn’t understand why this wasn’t more of a priority for them. Now I’m starting to understand.
Take a look at the numbers above, and ask yourself how much worse Twitter’s numbers would be if they did take on and kill off the botnets and fake account networks. The reality seems to be that Twitter, and Twitter’s stock price, is to a degree dependent on their existence, and Twitter (pretty literally) can’t afford to be too aggressive and successful shutting them down. It might tank their stock, which might make them a more attractive and cheaper acquisition target.
But wait. It gets worse
This problem is has nastier implications: a subset of those fake/bot accounts are the accounts being used by groups like GamerGate to attack and harass other users. Twitter has, frankly, done an abysmal job of trying to protect the users being attacked here on Twitter and the tools they’ve built to allow users to protect themselves are terrible. There’s really not excuse for this, and it’s one of the biggest problems on their service today, and they’ve shown very poor leadership and initiative at trying to grapple with and resolve the problem.
But take a step back: is this because financially they’re dependent on fake accounts to make their financials, and if they attack the abuse problem they also kick their stock price in the kneecap? Could Twitter be dragging their feet on fixing the abuse problems because it risks hurting the numbers that keep their stock price somewhat above water?
I would like to hope not, but I have to admit, this has me thinking. (and unfortunately, the alternatives to this are they are incompetent and can’t fix it and they simply don’t care and I don’t like those options, either).
Is Twitter the next USENET?
I used this comparison with someone the other day, and the more I think about it, the more I think it’s spot on: Twitter is reminding me more and more of USENET, and not USENET in the good days.
The signal to noise ratio is dropping. Spam and porn problems are growing. The user community is fragmenting into groups that only interact with outsiders on the service by fighting with them.
You’re starting to see more and more people drop off Twitter either because of abuse or because, well, just not worth it any more. Many more are either being much less active or going mute and just watching. Remember, half my followers aren’t really there.
And the abuse problem is off the charts, and Twitter has shown no strong interest, much less urgency, in getting it under controlled. Instead, they do as little as possible until public pressure forces them to react, and then they do some minimal act (like banning Milo) that they can point to and say we’re trying. Twitter, your not trying, you’re failing at fixing this problem.
And that looks a lot like how USENET was during it’s decline: It got big, it got noisy, it got angry, it got ugly, and people slowly abandoned it, and then quickly abandoned it, and then USENET turned into a few alcoves of sanity and community surrounded by vast waves of hate speech and porn.
Sound familiar? Well, you know how this turned out for USENET — especially all of you younger people reading this and madly looking it up in Wikipedia to see what I’m talking about…
The good news for Twitter is that it’s not far down that curve of decline, and there’s still time for them to get serious and fix this. The bad news is that they’ve waited long enough and let the bad aspects of the service get so ingrained in the service that the challenge to fix this is a tough one; worse, it’s one they’ve shown no real understanding of, much less interest in resolving.
And my gut says management at Twitter isn’t the group to solve this, and doesn’t really understand the implications of the toxic environment they’ve allowed to grow on their service. Which saddens me, but doesn’t surprise me.
The more I study USENET, the more I become convinced it’s a failed service and has started it’s path down the long slow decline into irrelevance (with porn). It’s not too late to fix this, but time is running out.
So I think the days of Twitter as an independent company are numbered. The way to fix it, I think, are to bring it new owners and disconnect it from the stock price, allowing them the financial freedom to crack down on bots and fakes and abusers and porn and refocus the service on conversation. It needs new management that has to be willing to focus on understanding and fixing these problems, not waving their hands or banning someone like Milo and feel like they accomplished something.
Or leave it alone, and, well, you see how strong and vibrant USENET is today…