Elon Musk’s worst enemy may very well be his own website.
On Friday, Twitter owner Elon Musk began pushing back against reports that the company tweaked the platform’s algorithm in order to specifically boost Musk’s own tweets this past week.
« Several major media sources incorrectly reported that my Tweets were boosted above normal levels earlier this week, » tweeted Musk. « A review of my Tweet likes & views over the past 6 months, especially as a ratio of followers, shows this to be false. »
Musk further explained that there was « a bug that briefly caused replies to have the same prominence as primary Tweets, but that has now been fixed. » At the time, Musk did acknowledge an issue with the « algorithm. »
However, Musk’s claims are refuted by Twitter’s own data, which does show a big boost in impressions on Musk’s tweets that line up with the reported timeline for the algorithm changes to Twitter.
Queensland University of Technology researcher Timothy Graham analyzed the data, which was pulled directly from Twitter’s official API, and found that impressions on Musk’s tweets were up 737 percent on Feb. 13, the day after the Super Bowl, shortly after the reported algorithm changes were made. In the days that followed, well after Musk’s tweet acknowledged an algorithm issue, the daily impressions on Musk’s tweets nearly tripled.
Musk supported his version of the story by tweeting about what he called a « review » of his « Tweet likes & views over the past 6 months. » As evidence of this review, he provided a screenshot of the 311 million impressions one of his tweets — the one about putting the cocaine back in Coca-Cola — received in April of last year, and noted that none of his subsequent tweets have « come close » to that number yet. But it should be noted that individual tweets from users with less than even 1,000 followers routinely go viral and rack up millions of views.
As Platformer first reported, Twitter engineers were tasked with making changes to the website shortly after the Super Bowl on Sunday after one of Musk’s tweets failed to perform as well as a similar post from President Joe Biden. The next afternoon, a « fix » to Twitter was pushed out that « artificially boosted Musk’s tweets by a factor of 1,000. »
The change to Twitter’s algorithm was so obvious that users started complaining that their feeds were being filled up with Musk’s tweets.
Regardless, Musk is claiming that the Platformer report was « bogus » and that the outlet’s source is a « disgruntled employee who had been on paid time off for months, had already accepted a job at Google and felt the need to poison the well on the way out. » Musk then claimed that Twitter would take legal action against the individual.
Platformer’s Casey Newton replied that Musk’s claims were inaccurate and that the outlet stood by its story.
This isn’t the first time Twitter’s own data has debunked claims from Musk or his defenders this week.
On Thursday, Musk fans falsely claimed that a Mashable story regarding Tesla unsubscribing from Twitter Blue was false. A Community Note was attached to Mashable’s tweet linking to the story insisting that because Tesla is business verified with the gold checkmark, it could not be a Twitter Blue subscriber to begin with.
Musk signaled his support for this usage of Twitter’s Community Notes in a tweet.
However, Twitter’s own data as pulled from its official API showed that Tesla did indeed unsubscribe from Twitter Blue in the past week. Furthermore, the data provided examples of business verified accounts with gold check marks which are also subscribed to Twitter Blue, disproving the claim that Tesla’s Twitter Blue subscription wasn’t possible in the first place.
The Community Note was later removed from the Mashable tweet linking to the Twitter Blue story.
