Meta conducted an experiment where thousands of users were shown chronological feeds on Facebook and Instagram for three months. Users of the chronological feeds engaged less with the platforms and were more likely to use competitors like YouTube and TikTok. This suggests that users prefer algorithmically ranked feeds that show them more relevant content, even though some argue chronological feeds provide more transparency. While the experiment found that chronological feeds exposed users to more political and untrustworthy content, it did not significantly impact their political views or behaviors. The researchers note that a permanent switch to chronological feeds could produce different results, but this study provides only a glimpse into the issue.
I think this is bullshit. I exclusively scroll Lemmy in new mode. I scroll I see a post I already have seen. Then I leave. That doesn’t mean I hate it, I’m just done!
In the 80’s, Pepsi was gaining quickly on Coca Cola with the Pepsi challenge: having tasters blindly tasting Pepsi versus Coke and choosing which one they liked better. Pepsi won a majority of these. But over the decades, it turns out that consumer preference for a sip of each didn’t necessarily translate over an entire can, or an entire case of cans. When asked to drink 12-20 ounces (350 to 600 ml) of the soft drink, regularly, people behaved differently than what they did for a 2 ounce (60 ml) taste.
Asking consumers to rate things in the moment still suffers from their less reliable momentary ratings of things they experience all day, day after day. Especially of things that tend to be associated with unhealthy addictions.
Yeah, you’re right that even having users rate content is still limited.
I’d argue it almost definitely has to be better than engagement, though. It also has the potential to be less punitive to people who actually are thoughtful with what they like by using the likes as more of a classification problem and less shoving the same trash in everyone’s face.
It’s a hard problem, but sites aren’t even attempting to actually attempt to do anything but tie you to a shitty dopamine loop.
Totally agree. I think those who design the algorithms and measure engagement need to remember that there is a difference between immediate dopamine rush versus long term user satisfaction. User votes can sometimes be poor predictors of long term satisfaction, but I imagine engagement metrics are even less reliable.
They don’t want satisfaction.
They want addiction.
That’s not a sustainable model, either. Zynga had a decent run but ended up flaming out, eventually purchased by a large gaming company.
That’s to say nothing of the business models around gambling, alcohol, tobacco, and addictive pharmaceuticals. Low level background addiction is the most profitable, while intense and debilitating addictions tend to lead to unstable revenue (and heavy regulation).