[panick21_]

I mean if its true and he has 100 million followers and he has only 10000ish impression then there is something seriously wrong. Because that flat out makes no sense.

Even if you assume 60% bods, and 70% of users not reading their timeline.

[benjaminwootton]

The timeline of people you actively follow isn’t the default any more. They push the machine learning driven feed as the default option and switch back to it when you restart the app.

With that in mind, Elons main option is to ask the algorithm for more impressions in the black box feed. That will get him into the feeds of non followers and show his tweets to followers when they eventually log in.

I experienced something similar on LinkedIn. I used to have a lot of followers and high engagement. At some point it changed such that the algorithm could bury you or promote you as it sees fit. At that point the only option is to write content that the algorithm promotes rather than content which your followers find interesting. Everyone worked this out and started writing their vulnerable virtue signalling stories for engagement, and the platform went downhill.

[joshspankit]

It makes sense from the perspective of the “social media bait and switch”.

Intuitively you’d think that following someone indicates that you want to see their posts immediately after they post it, but the “algorithms” distort that entirely as a way of making money for the platform.

Whether it’s requiring people to pay $$ to reach more followers, or promoting “posts” (ads) from other paying accounts that you’re not following, or even promoting sticky content designed to keep users on the app a little bit longer (and thus expose them to more ads + boost their DAU count). The whole “timeline” paradigm is a lie. I mean, it’s rarely even sorted by time.

[pavlov]

We don’t know how long the tweet in question had been up. 5 minutes? An hour?

Without this information it’s impossible to guess about the reasons because impressions accumulate over time.

[panick21_]

Agree. I not taking him at his word that he is correct, and of course we would need more data. But it does sound strange.

My problem with all these twitter reporting reminds me of Tesla a few years ago, people just wildly extrapolating and infering from tiny amount of information and then deriving prove that Musk is a piece of shit and the company is going down in flames.

The first can be argued, but the second doesn't seem to be happening nearly as much as people claim.

[z3c0]

Is it though? Just because somebody follows Musk does not mean they engage with his content enough for it to be floated up past everything else they may follow. It's very likely that a great deal of those followers don't find his content interesting, but don't find it objectionable enough to unfollow him.

[panick21_]

I mean, an impression counts views normally right? It would seem strange that so few people saw the tweet.

Its also the case that he usually gets quite a few retweets and often lots of responses as well. So the number just seem low to me.

[z3c0]

I missed this prior to now, but (in my experience with web analytics - I'm not a Twitter dev) impressions are typically when it appears in a users view. A "scroll past" is usually enough, because advertisers just want to know that it made it in front of somebody. Their marketers will handle "grabbing your attention." However, whether it ever makes it to a users view to begin with is determined by "The Algorithm", which would be seeded by a users past interactions. Ignore Musk tweets enough, and presumably, they'll stop showing up by default.

[skywhopper]

Sure, something was wrong, but it’s important whether it was a anti-Elon conspiracy or just a buggy complex system. When you’ve gotten rid of most engineering staff, I’d bet on the latter.