Last week’s printed edition of Focus had a piece about how Germany’s politicians are using social media. It made the dubious claim that 61% of Green top candidate Katrin Göring-Eckardt’s Twitter followers could have been bought (JPG of that part of the Focus piece here).

Let’s actually instead try to get to grips with what is going on here, and try to draw some conclusions.

Firstly I have drawn up a table of the top 10 German politicians by total number of Twitter followers. I have used only personal accounts, drawing on the data from Pluragraph on follower numbers (tables here, here), and I have also intentionally not included Martin Schulz as his Twitter audience was built in Brussels before his move back to German politics. Hence I do not think it is fair to compare the others to him.

I then ran all the accounts through Status People’s Fake Follower Check – more details about how that tool works can be found here (and there’s a review of the service here). The reason I used that tool is that it tries to distinguish between Fake followers and Inactive followers. The results for the 10 politicians analysed are as follows:

The remarkable headline finding is that the best politician in the check – Sahra Wagenknecht – has only an estimated 25% Good followers. The vast majority of all the followers of German politicians are either Fake or Inactive. My own personal score on the check is Fake 12%, Inactive 30%, Good 58% – that gives me an estimated number of good followers of 13098 – equivalent to that of Peter Tauber! Also once the Fake and Inactive followers have been removed, the estimated number of good followers is actually very low – possibly demonstrating the lack of take-up of Twitter in German politics.

Beyond that in 8 of the 10 cases the percentage of Inactive followers is higher than the number of Fake followers – this is the effect that is documented in an excellent article by Meedia that takes apart the original Vice article that alleges that German politicians have been possibly buying Twitter followers. When you are famous – on Twitter and in real life – Twitter suggests your account as one to follow when a user is new on the network. These people will follow Altmaier or Ströbele, experiment a bit with Twitter, and then never log in again – but their follow still counts towards the follower number for that politician.

Also Twitter is a public network – anyone can choose to follow any politician they wish to follow. And sometimes – as Louise Mensch found to her cost in the UK – sometimes a load of these accounts are bots. The test for accounts like those of Ströbele and Goring-Eckardt that have a higher Fake percentage is then to see if something causes their follower numbers to shoot upwards. Here once more Pluragraph helps – it plots the evolution of the follower numbers. Goring-Eckardt’s numbers grew particularly at the time when she became a top candidate (logical), while Ströbele’s have been gradually increasing. This does not show anything untoward.

A further question is what any politician could actually do with these Fake followers, and if they could indeed control them. Yes, Fake followers inflate a politician’s follower numbers – that much is obvious. But compare the style and approach of German politicians tweeting to the way it is done by ragers on the UK eurosceptic right wing – there the bots come and retweet certain accounts to strategically increase the reach of those tweets, and flood UK Twitter with pro-Brexit bile (look at the sorts of responses to David Vance tweets for example). I cannot see similar patterns among the 10 politicians I have looked at here – these Fake followers are not being controlled by the politicians themselves as far as I can tell.

Then you come to the question: what could any of these politicians realistically do about this? The Status People tool offers a paid option that allows a user to systematically eliminate their Fake followers by blocking them. But is it right for a politician to spend money on that? Conversely renewed efforts to drive up the numbers of real followers through open engagement on Twitter would of course also help.

Largely though I draw three conclusions from all of this. First, we should take the follower numbers of all high level politicians in Germany with a pinch of salt (and the same may apply beyond Germany). Second, the level of real Twitter engagement in German politics remains low. Third, so long as Fake Twitter accounts are not being deployed by the politicians themselves to shape the Twitter debate, and the politicians themselves are not buying followers, then I see no major reason to worry here.