8 min read
How the Instagram algorithm decides who sees you
Instagram ranks feed, Stories, Reels and Explore separately, and each ranks on different signals. Reach problems are usually a mismatch between the surface you're posting to and the signal you're generating.
Four surfaces, four sets of rules
The feed ranks mostly on relationship: how often you two interact, whether they've messaged you, how recent the post is. Stories lean even harder on that. Reels barely care — they're built to show strangers things, and they rank on whether the video held attention. Explore sits somewhere between, keyed to what the viewer has engaged with lately.
This is why an account can have excellent feed engagement and no Reels reach, or the reverse. They aren't measuring the same thing, and improving one doesn't improve the other.
Watch time is the signal that moves Reels
For video, the dominant question is whether people kept watching. Not whether they liked it — whether they stayed, and whether they looped. A ten-second video watched twice all the way through outperforms a sixty-second one abandoned at four seconds, even when the longer one gets more likes.
The practical consequence is that the first second is doing most of the work. Not a slow build, not a logo sting — the thing the viewer came for, immediately.
Saves and shares outrank likes
A like costs nothing and means little. A save means the viewer expects to want this again. A share means they were willing to spend their own credibility on it. Both are much stronger signals, and both are what "make something useful" actually cashes out to.
This is why reference-style content — the list, the comparison, the thing you'd want to find again — travels further than content that's merely enjoyable, per unit of effort.
What the follower count does and doesn't do here
Follower count is not a ranking input in any direct way. Instagram doesn't show a post to more people because the account is large; it shows it to more people because the people who saw it responded. Large accounts get reach because they have more people in the first test group, not because size is rewarded.
That means a follower number that doesn't engage is neutral-to-negative for reach: it raises the denominator without raising the numerator. If your engagement rate has fallen after a growth push, this is usually the arithmetic, not a penalty.
Consistency is a ranking signal about you, not your posts
Accounts that publish regularly get treated as reliable inventory, and the platform has more recent data to rank them with. Accounts that vanish for six weeks come back cold — not punished, just unknown again.
The lever here is boring and it works: a rhythm you can hold, in one recognisable lane, for long enough that the platform learns who to show you to.