Recommendation system
The machine learning pipeline that shortlists candidate posts, scores each for predicted engagement, then ranks and filters them for one person at that moment. Understanding it as prediction rather than punishment explains most of what feels unfair about reach.
Related terms
The problem a system has ranking content from an account it knows little about. It resolves by testing your post with …
Content velocityHow quickly a post gathers engagement in its first hour or so relative to how many people saw it. Fast, strong early s…
Engagement baitExplicitly asking for interactions with no value attached, such as like if you agree or tag three friends. Platforms h…
Engagement podA group that agrees to like and comment on each other's posts to fake early momentum. It breaches most platform terms,…
For You Page (FYP)TikTok's main feed, built almost entirely from predicted interest rather than who you follow. This is why new accounts…
Ineligible for recommendationThe documented state where a post still appears to followers and on your profile but is withheld from Explore, the For…
Knowing the word is the easy part
A five-minute assessment scores you across six skill areas and builds a path through 28modules — skipping whatever you already know. Free, no card.
Take the free assessment