Twitter’s Bot Problem Is Worse Than You Think

Robots on computers.

The last weekend in January was a rather eventful one for Twitter (NYSE: TWTR) investors. On the Friday, shares of the company jumped nearly 10% on rumors that Chinese social-media company Tencent was interested in buying the social media company. While it’s important to note that this is an unconfirmed story from Citron Research, a known short-seller that has a mixed track record of success, some investors saw promise and bid the stock price up aggressively.

Twitter’s positive press abruptly ended the next day. A bombshell report from The New York Times chronicled the mass use of social media fraud on the site. The report centered on a small public relations company, Devumi, that has created more than 3.5 million fake, automated accounts (bots), selling these accounts 200 million times over to 200,000 celebrities, newsmakers, and other influencers to inflate their follower counts.

Even worse, many of those bots used stolen information from real accounts, like pictures, names, personal data, and location data, including some from minors.

Does Twitter know how bad things are?

According to the Times article, Devumi sells followers for a relatively cheap price of a penny per bot. While Devumi also provides bots for Facebook, Twitter’s anonymous nature and lack of verification make the service an easier target. According to the article, some estimates peg up to 48 million of Twitter’s reported users, or 15%, as bots. Twitter countered by noting the number of bots is “far lower” but did not offer an estimate.

Through a spokesperson, the company told The New York Times that it did not typically suspend users who purchased bots because “it would be difficult … to know who is responsible for any given purchase.” The Times reported that Twitter failed to say whether a sampling of fake accounts based on real users violated the company’s policies against impersonation.

Here’s why it’s bad for Twitter’s advertisers

For a company struggling to reverse the narrative that it’s poorly run, this report is problematic on many levels. First, it feeds into that narrative: Let’s not forget that many of these bots are reportedly using identification stolen from real people. While a stolen picture, hometown, and name may not be as damaging to the person who owns them as if a Social Security or credit card number were stolen, it’s still not right.

The biggest risk is that this could continue to push advertisers away from Twitter to other digital outlets like Facebook and Alphabet, which are coming close to having a duopoly in the digital advertising market. Could you blame advertisers for fleeing Twitter? The site has essentially aided and abetted click fraud, helping influencers (and itself) inflate their fees and overcharge brands for marketing.

According to The Times, an influencer with 100,000 followers might earn an average of $2,000 for a promotional tweet, while an influencer with a million followers might earn $20,000. At a penny per follower, a $9,000 initial investment would allow influencers to charge $18,000 more for promotional ads many times over. That’s money essentially stolen from brands and, eventually, customers of their products.

Here’s why it’s bad for Twitter’s investors

Another issue concerns Twitter’s relationship with investors. While admittedly there are key differences, there’s notable overlap between Twitter’s current predicament and the Wells Fargo account-opening scandal. Although equity investments are evaluated on profit and growth, analysts and investors use supporting figures to forecast and evaluate future performance. For Wells Fargo, a key supporting statistic was accounts per customer; for Twitter, a key figure is monthly active users (MAUs).

In fact, in early days, part of Twitter’s investment thesis was MAU growth, as it assumed the company would grow into lofty valuations, which at one time were as high as 58 times sales. That growth never occurred, as MAU growth has slowed to a standstill; now, investors are finding out that many of these monthly active users appear to be bots, which certainly changes the investment thesis.

Here’s why it looks bad for democracy

Finally, there’s overlap with congressional inquiries concerning the use of bots by Russia to attempt to influence the 2016 presidential election. Twitter itself noted that Russian bots retweeted candidate Donald J. Trump 10 times more than they did Hillary Clinton in the final weeks of the presidential campaign.

The Times article dovetails with the federal government’s inquiry by showing how easy it is to get bots to retweet, like, or amplify messages. Making matters worse, Twitter has an outsized effect on news cycles, as journalists and newsmakers use likes and retweets as a proxy for popularity.

If Citron’s report is right and Tencent is a prospective buyer, it could be difficult for the company to sell itself to a company in a foreign country — one sometimes considered hostile, and one where the government often interferes in matters of industry — in the event unanswered questions about bots and election influencing remain.

On the first trading day after The New York Times article, Twitter stock rose nearly 4% as investors seemed to shrug off its implications. But I’m not convinced. While I’ve been generally positive on the company, I’m staying away until Twitter can convince me there’s a fix for these serious problems.

error: Content is protected !!