Twitter Chief Executive Officer Parag Agrawal stood by the company’s estimate that less than five percent of its daily active users were “fake or spam”, as Elon Musk expressed doubts about it.
Let’s talk about spam. And let’s do so with the benefit of data, facts, and context…
— Parag Agrawal (@paraga) May 16, 2022
“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),” Agrawal said. “Externally, it’s not even possible to know which accounts are counted as mDAUs (monetisable daily active Twitter users) on any given day.
Agarwal then elaborated on the challenges of fighting spam.
“Spam isn’t just ‘binary’ (human / not human),” he added. “The most advanced spam campaigns use combinations of coordinated humans + automation. They also compromise real accounts, and then use them to advance their campaign. So, they are sophisticated and hard to catch.”
The Twitter CEO said that because the “goals and tactics” of spam accounts are constantly evolving, fighting them becomes an incredibly dynamic process.
“You can’t build a set of rules to detect spam today, and hope they will still work tomorrow. They will not,” he added.
Appearances complicate the detection of fake account, Agrawal added. “Many accounts which look fake superficially -- are actually real people,” he said. “And some of the spam accounts which are actually the most dangerous -- and cause the most harm to our users -- can look totally legitimate on the surface.”
Agrawal said Twitter’s team constantly removes spam from the platform while making sure the accounts of real people are not suspended inadvertently.
“Now, we know we aren’t perfect at catching spam,” he added. “And so, this is why, after all the spam removal I talked about above, we know some still slips through. We measure this internally. And every quarter, we have estimated that <5% of reported mDAU for the quarter are spam accounts.”
The Twitter CEO added that the company’s fake account estimates are based on multiple human reviews of thousands of randomly-sampled accounts“Each human review is based on Twitter rules that define spam and platform manipulation, and uses both public and private data (eg, IP address, phone number, geolocation, client/browser signatures, what the account does when it’s active…) to make a determination on each account,” he added.