Algorithmic feed

Each time you open the app, the system gathers a pool of candidate posts — from accounts you follow and, on many platforms, from accounts you don't — and gives each one a score. The score is a prediction: how likely are you to react, comment, watch to the end or simply stop scrolling? The highest scores go to the top. That is why two people who follow exactly the same accounts can see very different feeds.

• Engagement — the reactions, comments, shares and saves a post has already collected. See what engagement means.
• Recency — newer posts usually get a boost so the feed does not feel stale.
• Relationship — how often you interact with the author, and whether you follow them.
• Your habits — whether you tend to stop for photos, videos, polls or particular topics.
• Diversity rules — limits that stop one account or one topic from filling the whole screen.

Say you follow 200 people and 40 of them have posted since your last visit. A newest-first feed would show those 40 posts in time order. An algorithmic feed might open with a friend's photo that already has a long comment thread, slot in a post from someone you don't follow that matches your interests, and push a text update from an acquaintance you never interact with further down — sometimes far enough that you never reach it.

Ranking solves a real problem. Once you follow more people than you can keep up with, a strict time order buries the posts you would most want to see under whatever was posted last. The cost is transparency and control: it is hard to tell why a post appears, and a system tuned for reactions can end up rewarding provocative posts over useful ones. The guide to how social media feeds work goes deeper.