Tesla CEO Elon Musk sent Twitter shares down on Friday when he said he would put his $44 billion takeover of the social network “on hold” as he investigates the proportion of fake and spam accounts on the network. platform.
Although Musk later made it clear that he will continue to stick to the deal, he continued to insist on the issue of fake accounts. Writing on Twitter that his team would make their own analysis, he expressed doubts about the accuracy of the numbers Twitter has reported in its most recent financial filings.
In its earnings report for the first quarter of this year, Twitter acknowledged that there are a number of “fake or spam accounts” on its platform, in addition to legitimate revenue-generating daily active users or users (mDAU). The company reported, “We conducted an internal assessment of a sample of accounts and estimated that the mean of fake or spam accounts in Q1 2022 represented less than 5% of our mDAU during the quarter.”
Twitter also admitted to overestimating the number of users by 1.4 million to 1.9 million in the past 3 years. The company wrote, “In March 2019, we launched a feature that allowed people to link multiple individual accounts together to easily switch between accounts,” Twitter revealed. “A mistake was made at the time, so actions via the primary account resulted in all linked accounts being counted as mDAU.”
While Musk is rightly curious, experts on social media, disinformation and statistical analysis say his proposed approach to further analysis falls woefully short.
Here’s what the CEO of SpaceX and Tesla said he would do to determine how many spam, fake and duplicate accounts there are on Twitter:
“To find out, my team will randomly sample 100 @twitter followers. I invite others to repeat the same process and see what they discover.” He clarified his methodology in subsequent tweets, adding: “Pick an account with a lot of followers” and “Ignore the first 1000 followers and then pick every 10th. I’m open to better ideas.”
Musk also said, without providing evidence, that he chose 100 as the sample size for his research, because that’s the number Twitter uses to calculate the numbers in their earnings reports.
“Any sensible random sampling is fine. If many people independently get similar results for %fake/spam/duplicate accounts, that will be telling. I chose 100 as the sample size because that’s what Twitter uses to calculate <5% fake/spam/duplicate."
Twitter declined to comment when asked if its description of the methodology was accurate.
Facebook co-founder Dustin Moskovitz weighed in on the matter via his own Twitter account, noting that Musk’s approach isn’t exactly arbitrary, uses too small a sample and leaves room for huge error.
He wrote, “I also feel that ‘doesn’t trust the Twitter team to help draw the sample’ is its own kind of red flag.”
BotSentinel founder and CEO Christopher Bouzy said in an interview with CNBC that analysis by his company indicates that 10 to 15% of accounts on Twitter are likely “inauthentic”, including fakes, spammers, scammers, nefarious bots, duplicates. and “single-targeted hate accounts” that typically target and harass individuals along with others who spread disinformation on purpose.
Primarily backed by crowdfunding, BotSentinel independently analyzes and identifies inauthentic activity on Twitter using a mix of machine learning software and teams of human reviewers. The company controls more than 2.5 million Twitter accounts today, mostly English-speaking users.
“I don’t think Twitter is realistically classifying ‘fake and spam’ accounts,” Bouzy said.
He also warns that the number of inauthentic accounts may appear higher or lower in different corners of Twitter, depending on the topics being discussed. BotSentinel has discovered that there are more inauthentic accounts tweeting about politics, cryptocurrency, climate change and covid than those about non-controversial topics like kittens and origami.
Carl T. Bergstrom, a professor at the University of Washington who co-authored a book to help people understand data and avoid being picked up by false online claims, told CNBC that taking 100 followers from a single Twitter account should not serve as “due diligence”. for making a $44 billion acquisition.
He said a sample size of 100 orders of magnitude smaller than the norm for social media researchers studying this sort of thing. The biggest problem Musk would face with this approach is known as selection bias.
Bergstrom wrote in a message to CNBC: “There is no reason to believe that followers of the official Twitter account are a representative example of accounts on the platform. Perhaps bots will be less likely to follow this account to avoid detection. “More likely to follow them to seem legit. Who knows? But I just can’t believe Musk is doing anything but trolling us with this silly sampling scheme.”