What Is the Social Media Algorithm? — Free Tool
The algorithm is the ranking system deciding which posts each person sees. Learn why there is no single algorithm, which signals actually carry weight, and which widely repeated tricks are folklore.
The ranking system that decides which posts each user sees and in what order, based on predicted interest rather than chronology.
What it is
The algorithm is the ranking system that decides which posts each user sees and in what order, based on predicted interest rather than chronology. It is not one system. Every platform runs several, and each surface within a platform ranks separately.
How it is measured
A ranking system scores each candidate post for each viewer, then sorts. The inputs are roughly: what this viewer has engaged with before, the relationship between viewer and author, signals of quality such as watch time and completion, the post's early performance, and recency. A following feed leans on the relationship. A recommendation feed leans on interest matching and can show you an account you have never heard of. Search ranks on relevance to the query. Same platform, different systems, different behaviour.
Commonly misunderstood
Treating the algorithm as one adversary with a fixed rulebook. The single most common consequence is advice that cannot be true everywhere: what raises Reels distribution has little to do with how a LinkedIn feed sorts. A lot of circulating advice is folklore too. Some claims are real and documented, such as platforms preferring content that keeps people on-platform. Others, like edited captions or specific banned words tanking reach, have no evidence behind them beyond repetition. And the golden first hour is overstated, since recommendation feeds regularly surface older posts.
When it matters
It matters when you are diagnosing a drop, and when someone proposes a tactic that only makes sense if all platforms rank the same way.
Related terms
Features
- Plain definition, and why there is no single algorithm to beat
- How ranking differs between a following feed, a recommendation feed and search
- The signals platforms actually document, and how far you can trust them
- Which common ranking tips are folklore rather than documented behaviour
- Links to related terms so you can connect ranking to organic reach and shadowbans
Frequently asked questions
Is there one algorithm per platform?
No. Each surface ranks separately. Instagram alone ranks the following feed, Explore, Reels and search with different systems and different signals, and the company has said so publicly. Advice that talks about the algorithm as one thing is already too vague to act on.
What signals matter most?
Broadly: predicted engagement based on your past behaviour, the relationship between viewer and author, how long people spend with the post, and recency. Platforms publish partial descriptions of these, and the weights change often, so treat any precise ranking recipe with suspicion.
Does the first hour decide everything?
Early response matters because it is the first evidence the system has, but the golden hour is overstated. Recommendation feeds happily surface older posts that keep performing, and short-form video in particular can pick up days later.
Do links, edits or certain words reduce reach?
These are mostly folklore. Some platforms do favour content that keeps people on-platform, which is a real and documented preference rather than a punishment, but there is no evidence for the specific claims about edited captions or banned words that circulate in creator groups.
How should I actually respond to ranking changes?
Watch your own data rather than announcements. Ranking is optimised for the viewer's attention, so the durable strategy is content people choose to watch and reply to, tested by format and measured over weeks.
