Study finds 46% of company Twitter account followers are bots
We know by now that companies will go to extreme lengths to boost their followers on Twitter or Facebook, and make them appear more popular than they actually are. But given the shady nature of this, it’s been difficult to determine to what extent this is actually being done. A study by an academic in the digital field – Marco Camisani Calzolari – aims to lift the lid on this. He compiled a study into 39 brands on Twitter, both in Italy (where he is based) and internationally, using an algorithm to determine the behaviour of followers. Using key indicators into what type of behaviour was likely that of a bot, he found that up to 46% of followers were likely to be automated bots.
In the study, 10,000 followers were taken at random for each account and analysed. Behavioural patterns such as logging in from different applications, using a real image in your profile or various punctuation in tweets, determined that an account was likely to be human. Dell came out particularly bad in the study, with the highest level of bot followers – at 46%
A flawed ecosystem
While the study in itself may not be overly conclusive – it is a relatively small sample size and the methodology is unlikely to be totally accurate -this is a pretty alarming figure and it raises serious implications for brand activity on Twitter. Much as we might deride the companies that are caught out in a study like this, it points to a much wider problem within social media marketing and its effectiveness for brands. An obsession with numbers might be one thing – doing anything you can to entice people to like your page or follow you on Twitter – but if we’re reaching the kind of level where half of those followers could be fake, we need to start worrying.
What’s perhaps most concerning is that this gaming of numbers is being done across the board. Not just by large corporates, but by celebrities, politicians and general public figures who want to boost their popularity. It’s a kind of poison that’s creeped in, particularly on Twitter where it’s much, much easier to create these automated accounts in the first place. What it means is that we have a completely skewed view of just how useful Twitter is as a platform at all. We already know that the minority of users contribute the majority of tweets, but how much of this content then is pure spam?
Twitter should be able to crack down on this a lot more than they already are. If a professor can singlehandedly come up with an algorithm to detect bot accounts, why can’t Twitter do the same and remove these accounts? Unless, of course, they can and they’re afraid of what might happen if they open Pandora’s Box.