An algorithmic feed — TikTok's For You page being the archetype — assembles each person's stream by prediction rather than subscription: it estimates what will hold this viewer and serves it, follower relationships optional.
The core loop
New content gets tested on a small sample; the system reads completion, rewatches, likes, shares and follows; strong signals earn a bigger sample. Repeat until the signal fades. This is why growth on feed-first platforms arrives in steps and why a nobody can outperform a celebrity on any given day.
What it does to follower counts
Feeds weaken the link between followers and views — a million followers guarantees little if the feed disagrees. Followers still matter as the durable asset (early distribution, credibility, monetization gates), which is why velocity after each post is the number to watch, per the growth tracking guide.