Starting this week, some Twitter users may see a drop in their number of followers as the company continues efforts to limit the reach and visibility of suspicious accounts.
Twitter announced today that it will start removing accounts it has locked for “sudden changes in account behavior” from follower counts. The company says that “most people will see a change of four followers or fewer,” but that “others, with large follower counts, will experience a more significant drop.”
In a blog post, the company stressed that the change would not affect Twitter’s monthly active user or daily active user metrics. On Friday, the Washington Post reported that the company was “sweeping out fake accounts like never before,” having suspended more than 70 million accounts in May and June. As a result, Twitter’s stock started to slide on Monday, as Wall Street feared that the purge of fake accounts would negatively affect Twitter’s user counts in the second quarter. Twitter CFO Ned Segal tried to reassure investors in a series of tweets on Monday, saying that “most accounts we remove are not included in our reported metrics, as they have not been active on the platform for 30 days or more.”
Twitter has taken aim at a few different types of troublesome accounts in recent months in the name of promoting a “healthy conversation” on the platform. The first is spam accounts — the company said its systems identified and challenged more than 9.9 million potential spam accounts a week in May, up from 6.4 million in December 2017. The second group is those identified as trolls, whose tweets Twitter started making less visible this Spring.
Today’s announcement deals with accounts that Twitter says “were created by real people, but we cannot confirm that the original person who opened the account still has control and access to it.” The concern here is that the accounts have been hacked or have potential to be hacked. The company said that one of the things it looks for is if email and password combinations are posted elsewhere that “could put the security of an account at risk.”
Vijaya Gadde, Twitter’s legal, policy and trust & safety lead, also wrote that Twitter locks accounts if it notices sudden changes in account behavior, including “tweeting a large volume of unsolicited replies or mentions, tweeting misleading links, or if a large number of accounts block the account after mentioning them.” In May, the company said these were some of the signs it was using to identify trolls. When asked for further clarification, a Twitter spokesperson acknowledged in an email to VentureBeat that “we sometimes use some of the same signals for different efforts,” but said that Twitter’s also looking at other signals that it hasn’t made public.
In short — Twitter is testing out a lot of different factors to try to curb the influence of certain problematic accounts. Of course, if it fails to communicate those procedures clearly to users, it could risk a considerable backlash.
These changes come after Twitter has been criticized for the past year and a half for its failure to quickly remove Russian bot accounts intent on spreading misinformation during the 2016 U.S. presidential election.
Twitter is expected to offer more detail about how its efforts to promote “healthy conversation” are affecting its bottom line when it announces its second quarter earnings on July 27.