Elon’s ExTwitter Engagement Stat Exaggeration: Outside Stats Paint A Bleaker Picture
from the new-record-high-vibes dept
Does anyone actually trust Elon to be honest about, well, anything? Last week he claimed that ExTwitter hit a new “all-time high” on engagement, with “417 billion user-seconds globally” and that in the US it was 93 billion “user-seconds.”
First off, what the fuck are “user-seconds”? This is not a typical measure in the internet world. It’s also a potentially misleading one. The grand sum total of “user seconds” can be pretty misleading. Does it count people just seeing tweets in the wild? What counts as a “user second”? If someone just leaves a tab open does that continue to count? Are they actively engaging on the site? There are so many questions as to what that stat even means.
But, more importantly, as Media Post points out, Elon had announced numbers back in March that suggested even higher engagement data than what he claimed was this new “record” high. Of course, there appears to be some gamesmanship with the numbers, as the March numbers Media Post is discussing are per month, and Elon seems to be talking about a single (very newsworthy, right after the assassination attempt) day on ExTwitter:
Musk posted that X saw a cumulative “417 billion user-seconds globally” in one day — equating to 27.8 minutes per users — at 250 million daily active users, which does not align with X’s user reportage in March, when the company said users were spending 30 minutes per day with the app on average.
The company also claimed 8 billion total minutes in March, but 417 billion seconds only equates to 6.95 billion minutes, which either negates the “record high engagement” now or invalidates the previous numbers.
On July 15, Musk also posted that in the U.S., user seconds reached 93 billion — “23% higher than the previous record of 76B.” This equates to 15.5 minutes per user, on average, based on X’s previous reportage of 100 million U.S. users — a figure that is lower than expected.
The lack of standard and standardized reporting allows playing with the numbers to misrepresent how popular the site is.
Meanwhile, that same article also highlights how outside observers see little to no evidence of higher engagement on the site, and plenty of evidence of decline:
… a new report by data intelligence platform Tracer shows “significant drops” in user engagement and “drastic drops” in advertising unlike competitors like YouTube, Instagram and Pinterest.
In June, X advertising saw drops month-over-month and year-over-year, the report shows, with click-through-rates (CTRs) declining 78% month-over-month, which the report suggests reflects a sharp downturn in user activity. In addition, cost-per-thousand (CPMs) decreased 17% from May to June, suggesting that advertisers are also leaving X.
[….]
Comparatively, Instagram, Pinterest and YouTube have seen dramatic user engagement increases recently. Instagram’s CTRs surged by 89% over the past year, while YouTube and Pinterest saw increases of 77% and 385%, respectively. The success these platforms are seeing is likely a direct result of the introduction of new video-first launches, such as Instagram Reels, YouTube Shorts, and a host of new features on Pinterest.
Is it any wonder that the site is struggling, with Elon telling advertisers to go fuck themselves and then threatening to sue those advocating pulling ads from the site?
But really, when it comes down to details, does anyone believe Elon’s random “best day ever!” tweets to be trustworthy?
Filed Under: elon musk, metrics, stats, user seconds
Companies: twitter, x
Comments on “Elon’s ExTwitter Engagement Stat Exaggeration: Outside Stats Paint A Bleaker Picture”
Elon’s ExTwit metrics are about as reliable as Trump’s crowd size metrics.
Re:
The boy who cried best day ever… no, I meant wolf.
How many seconds attributed to AI bots?
Not only is it a meaningless metric...
…but even if we qualify it in some way to make it something worth measuring — for example, just to pick one of many available options “time logged into the site”, it still means nothing.
“Leaving the site open in a browser tab” or “leaving the app running on a device” isn’t engagement – it has little to do with people actually doing things actively.
And beyond that: Twitter is increasingly overrun by bots – porn bots, crypto bots, spam bots, Nazi bots, incel bots, whatever bots. Of course it is: Elon WANTS it to be, because bots inflate the user count and make Twitter look much more popular than it really is.
Here’s a data point for you: last night, I watched one of my colleagues log into their (small, little used) Twitter account. They’d picked up 282 new followers, which is impressive given that they only had about 30 previously. And all of those 282, ALL OF THEM, were porn bots, and all of them had followed 3,000 to 5,000 accounts. This is significant because Twitter rate-limits follows: you can only follow so many other accounts per unit time. Which means that all 282 of those porn bot accounts had to be cumulatively logged in long enough to follow thousands of accounts – each.
The point being that if “time logged in” is used to calculate user-seconds, then an awful lot of that is being racked up by bots, not by people. I’m sure that Elon knows this, but like the insecure miserable failure that he is, he’s lying about it in order to mislead investors, analysts, regulators, and everyone else.
Re:
From what I understand, when someone posts a video to “X”, it counts instantly as a view even if you just scroll past it in your feed. Which is why so many of his Nazi friends have high view counts – it’s pushed to anyone not really interested and if it plays for so much as a second while you scroll past it’s a view. That explains why people not popular in the mainstream have such high engagement numbers, and why advertisers aren’t returning (since serious advertisers need verifiable numbers and don’t like being on Nazi feeds).
I’m not sure who still uses it, but I have no doubt the number of real users is way, way lower than he pretneds.
It's underwear gnomes' logic!
1) invent a new metric
2) declare “by this new metric, we’re doing better than ever!”
3) ?????
4) profit!
'Can I check your work!' 'No, it's secret! Just trust me and give me money.'
The man that is both desperate to get advertisers to return to his platform and can’t keep his mouth shut, ever… yeah, I wouldn’t trust him to accurately tell me what day it is, I certainly wouldn’t trust him to accurately list how many active human users are on the site at any given moment.
I think the term ‘user seconds’ is a result of someone being of Toke Mind High Russ
Re:
I can assure you this is not Reefer Madness.
Re: Re: Lying High Musk
So this would be “reefer optimism”
or Reefer Does Math?
Whatever, Musk lies. He really cannot tell the truth.
Re: Re: Re:
Did you say Reefer Mathness?
Re: Re: Re:
The only time Elon isn’t lying is when he shills for totalitarian regimes.
I man, he DID write a puff piece for Xi and has blocked SOME anti-Russian nonsense for Putin, as well as complying with Modi’s censorship nonsense…
What?! Musk lying about one of his businesses?? Say it ain’t so!
Re:
He’s not lying, it’s just that when everybody use stupid rows and columns spreadsheets to make some basic calculus, Musk masters up to 12D spreadsheets.
In his upper reality world, that we cannot hope to understand, it seems to make enough sense for him to be happy.
So please don’t try to drag him into our good old reality where he’s not needed.
That’s one of the benefits of renaming to X — he doesn’t have to compare apples to apples against old Twitter stats — he can just measure stats while it was called X and pretend Twitter never existed.
Why is TechDirt still using Xitter?
When is TechDirt gpoing to drop Xitter?
Re:
Where are we using Twitter? Our account has stopped posting (over a year ago).
Re: Re:
Told you it was a bad idea to implement RFC 6214.
Re: Re: Re:
Who ever said that cessation of engagement through ExTwitter wasn’t a choice?
I feel that anyone who calls themselves a genius and literally just tossed money at something, gets placed in the “very stable genius” category with Trump. So that ALSO means, having been to the same “very stable genius” school as Trump, he most likely lies as much.
Re:
in re: “very stable genius”
The short-and-sweet of it is, if you have to brag about it, you ain’t it, period. You are something else, but we try not to talk about repositories of sewage here.
You only trust Elon when he’s shilling for despotic totalitarian regimes.
Otherwise, assume he’s lying.
This comment has been flagged by the community. Click here to show it.
Oh surprise. Another article attacking ex-Twitter and Elon Musk, despite the site’s delivering a better experience for users than ever.
Re:
…hallucinated no mentally competent human being, ever.
Re:
I think you’ve got more useful things to do with your time than post in a Techdirt comments section, Ms. Yaccarino.
Re:
Hi Elon!
Elon is right-wing garbage who thinks Trump is a good idea.
Why is anyone surprisedfuckingpikachu.jpg that he’s full of shit most of the time his lips are moving?!?!
Re:
I mean, he is a good idea… for Elon. A disaster for most people, but for a manbaby who lucked into much of his wealth and is throwing a tantrum because one of his kids dared to be trans, Trump’s policies will be ideal. The brownshirts he’s gathered might be disappointed with the outcome if he gets elected, but Elon will be happy.
Numbers are Hard
Tell me it’s not true that my 28.7 billion pico-seconds using X is a world record.
The only all-time high is Musk himself.
Seriously, kids, to smoke marijuana, you don’t want to end up like him.
Geez, Elon is having a $camatology moment: They continually bang on about growth statistics of 47X, 25X, 10X and so on and so forth. Totally spurious and not backed up with anything even resembling reality. Utter bullshit!