Twitter maintains that spam and fake accounts make up less than 5% of its users
The U.S. Securities and Exchange Commission launched an investigation into Twitter’s spam and fake account methodology in June, according to a new regulatory filing made public Wednesday.
“We note your estimate that the average number of false or spam accounts during fiscal 2021 continues to represent fewer than 5% of mDAU (monetizable daily active Twitter users),” the SEC’s Division of Corporation Finance wrote in a June 15 letter.
“To the extent material, please disclose the methodology used in calculating these figures and the underlying judgments and assumptions used by management.”
In addition, the letter asked how, when and who discovered an error that resulted in an overstatement of Twitter’s mDAU from the first quarter of 2019 through the fourth quarter of 2021.
“Given that the error persisted for three years, please tell us how you concluded there was not a material weakness in your internal control over financial reporting and that your disclosure controls and procedures were effective as of March 31, 2022,” the regulator said.
In a June 22 response from lawyers at Wilson Sonsini Goodrich & Rosati, the social media giant stood by its less than 5% spam and fake account estimate, and its internal review was outlined in May 16 tweets by CEO Parag Agrawal.
“Twitter respectfully informs the Staff on a supplemental basis that the review is conducted manually by humans reviewing (in replicate) thousands of accounts, randomly chosen out of the accounts that Twitter counts as mDAU,” the company explained.
“Each human review assesses the selected accounts against a complex set of Twitter rules that define spam and platform manipulation, and, at a high level, an account is deemed to be “false” or “spam” if it violates one or more of Twitter’s rules regarding spam and platform manipulation.”
To perform its spam and fake account assessment, Twitter uses private and public data, including account activity, characteristics of how and when the account was registered on Twitter, contact information and information shared via the account’s profile. The assessment is also subject to a “multi-step, multi-tier review process where each account is separately and independently reviewed and investigated by multiple trained labelers.”
As for the mDAU error, Twitter said that the overstatement was discovered in March by its data scientists as part of an investigation into potential new product features that involved secondary accounts.
The overstatement represented “less than one percent of mDAU for each of the quarters from the fourth quarter of 2020 through the fourth quarter of 2021,” according to the company.
“The overstatement had no effect on Twitter’s financial statements or the accompanying footnotes, earnings, or other trends and had no impact on Twitter’s other key metrics,” the lawyers added. “Twitter estimates suggest the adjustments are not likely to change the trend in mDAU in any prior quarter in the applicable period.”
In a July 27 letter, the SEC said it had completed its review of Twitter’s filings.
“We remind you that the company and its management are responsible for the accuracy and adequacy of their disclosures, notwithstanding any review, comments, action or absence of action by the staff,” the letter stated.
Representatives for Twitter and the SEC declined to comment further.
Twitter’s correspondence with the SEC comes as it is suing Elon Musk in the Delaware Court of Chancery over the termination of his $44 billion acquisition of the company.
Musk, who has countersued, claims Twitter misrepresented the total number of spam and fake accounts on its platform when it accepted his $54.20 per share acquisition offer in April. In May, Musk urged the SEC to launch an investigation into Twitter’s spam and fake account methodology.
In addition, Twitter’s former security chief Peiter “Mudge” Zatko, has filed a whistleblower complaint accusing the company of “extreme, egregious deficiencies” related to cybersecurity issues that put the company’s users and shareholders and national security at risk.
Zatko also alleges that Twitter executives are “not incentivized to accurately ‘detect’ or report total spam bots on the platform” and that the company does not have the resources to fully understand the actual number of bots on the platform.
Musk and Twitter’s trial is slated to begin Oct. 17. Meanwhile, the Senate Judiciary Committee will hold a hearing with Zatko, and Twitter shareholders will vote on Musk’s deal Sept. 13.