Every time you go to write anything about the platform formerly known as Twitter these days, whether you like it or not, whether you intend to or not, you are also going to be wading into political waters.

That’s because Elon Musk has made X into a political story in itself, while any criticism is viewed through the lens of personal hate for Elon, and an effort to stop Musk’s inherently political goals.

“You hate free speech” is the most common counter, but in reality, nobody hates free speech, and the vast majority of critics aren’t trying to make a comment on Musk or his stated goals. In fact, those same critics have been just as harsh about Meta, TikTok, Snapchat, etc.

But Elon has intentionally tried to make criticism of X a bigger issue than just the flaws at X itself. Yet, it’s difficult not to highlight flaws in Musk and Co.’s logic when they keep amplifying misleading stats, which run counter to every other analysis of the X project.

For comparison, here’s a look at some of the claims made by Musk and Co., versus third party analysis and external data, and you can judge for yourself which is more likely correct.

“X is seeing record high usage”

One of the most common claims amplified by Musk is that X’s traffic and engagement stats are rising, and are regularly surpassing new record highs.

And while graphs like this look compelling, they don’t tell the full story, or really, even a significant part of it.

This data display is sourced from SEMRush, an SEO analytics provider. SEMRush only measures referral traffic from Google, not app usage, and for both X and Instagram, upwards of 80% of their respective usage comes via their apps (in the case of IG, it’s around 97%). So this is only a fraction of the usage that both apps are seeing.

But that’s not all. That big spike at the end of March/beginning of April? Yeah, that’s not reflective of usage, but a change in SEMRush’s reporting methodology, in which SEMRush began counting instances of traffic being driven by Google’s “SERP Features”, those breakout boxes that are included within Search results.

Which, as SEMRush has told me, disproportionately benefits platforms like X and YouTube, because they have more of these breakout displays, while the vast majority of IG posts are not indexed by Google at all.

So in the end, this chart represents less than 15% of the total traffic to X, and significantly less of the traffic sent to IG, and even then, the numbers are not representative of any major shift in X usage.

But even so, Elon’s still happy to blast this out to his 176 million followers as a misleading display of X’s actual performance.

It’s examples like this which make it difficult to trust the numbers being shared by X, and even more so when virtually every third party analysis report indicates the opposite.

• In October last year, Apptopia reported that X has seen a 13% decline in daily active users since Musk took over at the app in late 2022. Apptopia’s data suggests that the re-brand to X accelerated this decline.
• SimilarWeb has reported that X has seen a 14% decline in mobile and web visits since September 2022.
• AppAnnie’s 2023 “State of Mobile” report found that X’s monthly active user count declined significantly between 2022 and 2023.
• Edison Research recently published a report which showed that X has seen a 30% decline in usage over the past year.

On balance, you would have to assume that X’s usage numbers are actually declining, not rising, but you can also refer to X’s own data on this.

Back in December, X CEO Linda Yaccarino claimed that X was seeing 1.5 million new sign-ups per day.

Which would mean that X’s total user count should be increasing by around 45 million new users every month. But it isn’t.

In March 2023, Elon claimed that X had reached a new milestone of 500 million monthly users. In October 2023, when X provided an update on its usage stats, it reported the exact same figure for MAU.  

X is still seemingly stuck on 500 million actives, which suggests that if it is seeing 45 million new sign-ups every day, those new sign-ups are not sticking around.

Then again, that could also relate to the next issue with X’s stats.

“Elon has eliminated bots”

One of the biggest elements of focus for Elon Musk when taking over the app was bots, and eradicating bot profiles from the app.

Which he has repeatedly claimed to have done.

But he hasn’t.

At the numbers of sign-ups that X claims to be seeing, it seems very likely that many of these new accounts would be bot profiles, while X continues to test out new experiments, like charging all new sign-ups $1 to interact in the app, as part of a program called “Not a Bot”.

Which suggests that bots are still a significant problem in the app.

Indeed, recent analysis conducted by ad measurement provider CHEQ, found that of all of the referral traffic driven via X ads during the weekend of the Super Bowl, almost 76% of it came from bot accounts.

Bots are definitely still a significant factor in the app, which would also be impacting its overall usage numbers.

Back when Elon was trying to get out of buying the app, after his $44 billion bid was accepted by the Twitter board in mid 2022, Musk and his team conducted their own review of the Twitter bot problem, and found that the number of bot profiles was significantly higher than Twitter had claimed.

Twitter had long held that the number of fake profiles in the app didn’t exceed 5% of its total mDAU count, based on its own sampling, but Musk claimed that it was actually much, much more, with his own team’s analysis finding that around 33% of Twitter’s active profiles were likely fake.

Musk eventually settled on it being a more modest 20%, while noting that the true bot count was likely much higher.

Part of Twitter’s challenge in combating bots was that there was limited incentive for it to do so, because removing bot accounts would reduce Twitter’s overall usage stats, and thus negatively impact its stock price.

Back in 2022, amid discussion around Elon’s potential takeover of the app, former Twitter CEO Jack Dorsey said that the platform “needs cover for a while”, in the form of going private, in order to reduce market scrutiny, thereby enabling it to effectively combat its bot problem.

Essentially, when you get rid of bots, Twitter’s user count will inevitably drop. But that hasn’t happened under Musk.

At the time of the takeover, Twitter had 237.8 million monetizable daily active users (mDAU). Yet, just three months later, Musk claimed that Twitter was up to 250 million mDAU.  

So not only did the platform not lose any active users as a result of Musk’s claimed anti-bot measures, it actually gained 12 million more. That seems to belie the claim that the platform could have gotten rid of the bots, because it would not only have to have deleted millions of bot accounts, but added tens of millions more to make up for them, and more on top of that again, leading to this higher usage figure.

Considering that Twitter had never added more than 20 million new users in full year, that seems like an unlikely claim.

So is X stretching the truth? Only Elon and Co. know for sure, but it does seem questionable, at best.

Lack of Context

Finally, X also keeps sharing data cards like this, as a means to highlight its significance.

Which looks pretty good, right? Data presented in graphic form always looks more impressive, while millions of impressions is significant, no matter how you look at it.

But then again:

• In 2014, the State of the Union Address generated 2 million Tweets, the exact same amount, a decade ago, at a time when Twitter had half the users that X does now.
• In 2017, the State of the Union Address generated 3 million tweets.
• In 2018, the State of the Union Address generated 4.5 million tweets.

So while X is keen to present its usage data in a flattering light, the truth is that X engagement is clearly down on previous years.

The same also relates to questions around how X is now measuring video views.

In August last year, X claimed that an interview between Tucker Carlson and Donald Trump generated 236 million views in the app, though the video itself had actually only been watched 14.8 million times.

The discrepancy here is that X had moved to using its more general ‘views’ figure, i.e. the number of times that a post has appeared in user feeds, as opposed to its slightly more accurate video view count, which measures how many times users have actually tapped on the play button and watched more than two seconds of the clip.

Shortly after this, X removed the video views count from public view, which now means that general users have no way of knowing how many times a video has actually been viewed in the app.

As another example, in January this year, YouTube star MrBeast uploaded a clip directly to X, which ended up getting over 156 million ‘views’, though only 5 million X users actually watched the clip.

That’s a 96% variance in the ‘view’ count displayed on X, and the actual number of people who intentionally watched the video.

Pretty misleading, which again raises questions as to the data that X shares as a representation of what it actually means, in relation to the platform’s performance.

That’s not to say that X is doing horrendously bad. 250 million users is still significant, and X still plays a key role in general discourse, especially around real time events.

But it does suggest that the data that X shares should not be trusted as 100% indicative of the app’s performance.

Really, its revenue data is the most significant either way, and reports suggest that X’s overall ad revenue is still down by around 60% versus when Elon took over at the app. That’s not a sustainable intake, and unless X can convince a heap more people to pay for X Premium, it’s hard to see how it can continue to remain in operation, unless it can reverse that trend at some stage.

But the numbers also suggest that X’s misleading performance stats are failing to convince its major ad partners, whom it really needs to get its business back on track.

The bottom line is that while X seems to think that it’s highlighting its growth with these examples and stats, it actually seems to be hurting its overall market standing.

And whether you view that as a political stance or not, the data is what it is in this respect.