Skip to main content

Are Zombies taking over Twitter?

14 May 2015

Earlier this month BBC correspondent Rory Cellan-Jones published an interesting article about Twitter's recent fortunes. At the time Twitter's share price had fallen around 25% over the course of several days. This share price drop was triggered by the publication of Twitter's Q1 2015 results which included a cut in Twitter's 2015 revenue forecasts. The silver lining to this cloud was that Twitter's user growth had picked up. According to Twitter's results it gained an extra 14 million 'monthly active users' in the first quarter of 2015 bringing the total year on year gain to a staggering 47 million users. That's a lot of new users. It's not far off the entire population of England joining Twitter in one year.

Investors understandably like to see an ever upward trend of new users, proving that Twitter is still the darling of social media platforms, and more importantly, likely to generate a profit sometime soon. Something it has never done to date. If Twitter's user growth had dropped at the same time as revenue falling short of analyst expectations then we could have seen a stampede for the exit from Twitter's shareholders. But how can we verify these new users exist and are run by real people? In short, we can't, we have to take Twitter's word for it. There's no reason to doubt the user accounts exist but were they all really created by real people?

A couple of years ago we discovered that there is a murky world of services where you can buy Twitter followers to boost your apparent popularity. Several of these services offer you the option of buying up to 500,000 followers for a fee in the region $2000. The sheer scale of the numbers means that the new followers you buy definitely aren't real people, they're fake accounts created programmatically. The companies behind these followers are essentially running 'botnets' of zombie Twitter accounts.

Back in October 2013, to verify the effectiveness of buying Twitter followers, we setup a dummy Twitter account called @Not_ZodiacMedia and bought 2000 new followers. The company we paid delivered on their promise and our account gained 2357 followers within a week. Interestingly now over 18 months later that account still has 1860 followers, all of which are fake. This indicates that Twitter hasn't taken any steps to flush out these zombie user accounts. The number of companies selling followers by the 100,000 indicates that there are probably millions of zombie user accounts on Twitter.

Cellan-Jones's article goes on to detail his own experience of contributing to Twitter's revenue stream, by paying to promote a tweet for the @BBCTechTent Twitter account. This promotion apparently ran for around 36 hours at a cost of £20 or roughly $31.50. The key metric for success for this campaign was the number of followers gained over the time it was running for. The result was a gain of 643 followers. That equates to a cost of approximately 4.9 cents per follower gained.

The going rate for buying 1000 followers seems to be around the $12 mark equating to a cost of 1.2 cents per follower gained.

From our point of view, the fact that it's so cheap and easy to buy followers means that Twitter's advertising proposition is severely undermined. Unscrupulous social media marketing companies are presented with a clear opportunity to profit by artificially improving the number of Twitter followers for clients whilst simultaneously appearing to reduce client spend on Twitter advertising.

Until Twitter takes steps to identify and remove zombie accounts it's very difficult to discern which of your followers are fake and which are real. With their stock price being strongly linked to user base growth there's no incentive for Twitter to do this.

There seems to be little awareness in the mainstream media and general public about this situation, although recently Hilary Clinton came under fire for the percentage of her Twitter followers who appear to be fake. To verify that buying Twitter followers is still as effective as it was back in October 2013, and to draw attention to this phenomenon, we paid $42 for an additional 5000 followers for the @BBCTechTent twitter account. The before and after stats are telling: On 3rd May 2015 the account had 6.3k followers and at the time of writing the account has 11.6k followers.