officials on Thursday doubled down on defending the accuracy of their calculation of spam on the platform, addressing an issue that has become a potential stumbling block in
$44 billion deal to buy the company.
The officials, who asked not to be named, reiterated on a call with reporters Thursday that spam accounts make up less than 5% of the company’s daily monetizable users, which Twitter defines as daily users who are logged in and authenticated by Twitter. They added that it would be difficult for outsider auditors to accurately measure the figure that is based on private user information, such as internet addresses, geolocation data and contact information, that Twitter doesn’t share.
The issue of bots and spam has emerged as a complicating factor in Mr. Musk’s deal to buy Twitter. In May, Mr. Musk said his deal was “on hold,” as he questioned the way Twitter calculates spam and fake accounts. Mr. Musk has suggested he thinks the number could be closer to 20%, without providing evidence that the company’s disclosures were unreliable.
Twitter’s board has insisted it is moving forward on the deal as agreed. Executives have told employees that there is no such thing as a deal being put on hold, The Wall Street Journal previously reported. After some legal back-and-forth between the two sides, Twitter in recent weeks provided Mr. Musk with historical tweet data and access to its so-called fire hose of tweets, people familiar with the matter have said
Twitter’s stock was up around 2% in Thursday trading.
In the briefing with reporters, the Twitter officials declined to say whether the company shared the private user information with Mr. Musk. They would only say that Twitter has fully complied with the obligations of the purchase agreement and that the privacy of user information was important to the company.
Twitter and rival platforms have spent years trying to stamp out bots and fake accounts through content-moderation teams and artificial intelligence tools, but many describe spam and fake accounts as a whack-a-mole problem since bad actors are constantly updating their methods.
The Twitter executives said their estimate of spam is based on sampling a portion of its accounts and human reviews of account activity. Those reviewers, Twitter executives said, analyze public information like what that account tweets and interactions with other accounts. They also factor in information that is available only to Twitter, including IP addresses, client and browser signatures and contact information.
Twitter said it typically removes more than one million spam accounts daily while they are being set up or soon after. Additionally, the company said, it locks millions of suspected spam accounts a week if humans can’t verify they aren’t spam.
Executives said that some legitimate users might appear to be fake because they kept automatically generated usernames and never took the additional step of making their username their own. There, private information, not available to the public, can help Twitter determine whether it is real.
Twitter officials added that because outsiders don’t have access to its internal data, those spam estimates can’t be accurate. They said the company hasn’t found a way to share this information with outside experts in a privacy-safe way.
Copyright ©2022 Dow Jones & Company, Inc. All Rights Reserved. 87990cbe856818d5eddac44c7b1cdeb8