Twitter's Getting Serious About its Bot Problem – Which is Good News for Social Media Marketers
Twitter analytics tool Followerwonk shared an interesting data point yesterday – one worthy of extra attention.
Twitter is deleting/suspending accounts at an unprecedented rate. In the past 30 days we have detected:— Followerwonk (@followerwonk) March 5, 2018
‣ 2.18 million new accounts
‣ 3.18 million deleted/suspended
We've never seen a net loss over 30 days, before.
This is significant for a couple of reasons – first, Twitter undoubtedly has a bot problem, and has had for years.
Back in 2013, Twitter said that around 5% of the accounts on its platform were fakes - equivalent, at that stage, to 10.75 million users. More recent research has estimated that fake profiles make up around 15% of the platform’s total user count, which, as per their latest earnings results, would suggest there are now around 49.5 million false Twitter users.
The issue of bots became a key focus as part of the investigation into how social platforms were used to spread misinformation, and sow division among voters during the 2016 US Presidential Election campaign, with both Twitter and Facebook playing the most prominent roles in such actions. That’s sparked Twitter into action – they recently announced changes to their API, for example, to limit the use of mass follow and retweeting.
But we’ve heard similar from Twitter before, and they’ve seemingly not taken action – part of the reason for this (or so the logic has been assumed) is that Twitter has struggled to add more users, and removing bots would only hurt their overall figures. The above numbers from Followerwonk would suggest that’s true, but Twitter, evidently, is doing it anyway, which will help improve the actual performance of their ad products, and increase trust in Twitter data.
And the latter is also a key note – another problem with Twitter bots is that it’s become incredibly easy to ‘cheat’ your way to a massive Twitter following, then profit from your knowledge of community and brand building, and credentials as an 'influencer'.
The New York Times published an article on this recently, highlighting several celebrities and social media experts who’d been found to be purchasing followers.
Removing bots enables Twitter to take control of this situation, and improve the credibility of their metrics – as social media marketing becomes a more significant element in modern business, people are also becoming more informed about what those numbers actually mean, and how to identify real influence, as opposed to those who’ve bought their way to prominence.
If people repeatedly see influencer lists filled with people who they know have cheated the system, that reduces their trust in such data, while it would also skew Twitter’s algorithm which favors content from more prominent users, and with higher engagement stats (both of which can be gamed).
Part of the problem thus far has been in identification, but Twitter’s been developing its machine learning tools for some time, they surely have the capacity to identify and weed out bad actors. The question then, as noted, has been whether they have the motivation to do so. If Twitter does continue this trend, and is willing to take a hit in its overall user count in order to clarify its metrics, that would be a big step forward. It may not please investors, but it would improve the real world results of advertisers, and help non-users get a better understanding of what’s actually popular, and actually relevant within their niche.
This, of course, is only one stat, but hopefully this is indicative of a larger trend, and we do start to see Twitter crack down on bots and fakes, helping improve the data flow and uncover more relevant, valuable insights.
Follow Andrew Hutchinson on Twitter