X open-sourced its ranking algorithm, and the engagement weights are sitting right there, and they explain almost everything about why a new account gets zero views. The weights X scores posts for your timeline using a set of engagement weights.
X open-sourced its ranking algorithm, and the engagement weights are sitting right there, and they explain almost everything about why a new account gets zero views.
Model each action as a session-length signal: a like is a ~500ms interaction, then you keep scrolling.
a reply the author answers is a live two-person conversation that can run ten minutes and pulls both users back repeatedly.
The page is ready to read now. The fuller skim-friendly version will appear here automatically.
I ship products solo, which means I'm also the marketing department, which I'm bad at. So instead of guessing, I did the thing a developer does when the docs are bad: I read the source. X open-sourced its ranking algorithm, and the engagement weights are sitting right there, and they explain almost everything about why a new account gets zero views.
Why the numbers look like that The weights are a proxy for one metric: time on app. Once you read the ranking as "maximize time-on-app," every piece of folk wisdom about the algorithm either falls out of it cleanly or turns out to be wrong. What this means for a small account The standard advice is "post consistently." For a new account that's backwards.
A reply the original poster answers is worth 150x. It's the entire design philosophy compressed into three numbers. this is the jackpot, and the 75 weight says so out loud.
Open the app view to save this story, compare related coverage, and continue from the same source.