A new study estimates that upwards of 10% of Twitter active accounts post spam content — double the company’s own claims.
The report from U.K.-based data analytics and consulting firm GlobalData comes as Elon Musk, the billionaire Twitter power-user, has threatened to nix his $44 billion deal for the social network over the question of the prevalence of spam and fake accounts on Twitter. On Monday, Musk’s lawyers sent a letter to Twitter alleging the company was in “clear material breach” of the acquisition agreement because Twitter has refused to furnish information backing up its claim that fake/spam accounts represent less than 5% of daily active users.
A Twitter rep declined to comment.
According to GlobalData, its 10% spam estimate is conservative but the firm acknowledged that “there is no conclusive way of knowing if a certain account is a bot or spam.” According to Sidharth Kumar, GlobalData senior data scientist, the discrepancy between Twitter’s internal sub-5% estimate and the GlobalData model’s 10% estimate is likely due to a difference in criteria as to what counts as “spam.”
“The precise proportion of spam accounts is difficult to compute, as it is almost impossible to confirm the identity of the entity behind a tweet handle,” Kumar said, adding that “the definition of a spam account may differ for everyone. Incessant tweeting of non-original content can be considered spam, but some may choose to see it as a very active user sharing articles/opinions.”
For the study, GlobalData analyzed about 4 million recent tweets from a sample of 20,976 Twitter accounts to discern patterns — and concluded that 10.9% of those represented spam accounts. The firm’s model used multiple factors to determine whether a specific account was “spam,” including whether its tweets originated from third-party applications; whether it is Twitter Verified; number of tweets per day; proportion of retweets; the median time between any two tweets; the length of an account’s bio description; and proportion of links shared.
Wall Street has viewed Musk’s sudden interest in conducting due diligence about the spam/bot metric as an attempt to either back out of the acquisition or to drive the deal price down. Twitter has disclosed its estimate that spam and fake accounts represent less than 5% of its active users for years, dating back to its IPO filing in 2013.
Last month, Musk tweeted that Twitter’s active user base could represent “20% fake/spam accounts” and asserted, without citing any evidence, that it “could be *much* higher.” To be sure, the question of how many of Twitter’s 229 million daily active users (as of Q1) are actual monetizable users is an important factor in valuing the company, but analysts have wondered why Musk only zeroed in on the issue weeks after clinching the original buyout agreement.
In a series of tweets, Twitter CEO Parag Agrawal responded to Musk’s questions about fake/spambot accounts, tweeting in part that “Unfortunately, we don’t believe that this specific estimation can be performed externally, given the critical need to use both public and private information (which we can’t share).”