A multiple is a post's performance divided by the account's baseline. It is a single number and it gets read as a single scale, as though 30× were simply more of whatever 3× was. It usually is not. Different bands tend to have different causes, and the cause is the part you can learn from.
Under about 2×: probably noise
Posts vary. Time of day, day of week, what else was competing, how the first few hundred viewers happened to behave — all of it moves a post within a band around normal without anything meaningful having happened.
Treating small multiples as findings is how creators end up rewriting a strategy every week. If your baseline is properly trimmed, most of your posts should land here, and that is the system working.
Roughly 2× to 5×: a better post
This is the band where craft shows up. The hook was sharper, the topic was closer to what the audience wanted, the edit was tighter. It is the least dramatic band and the most useful one to study, because the cause is usually something you did on purpose and can do again.
It is also the band that gets ignored, because it does not feel like a win. That is a mistake: a repeatable 3× is worth more than an unrepeatable 30×.
The band you can reproduce is worth more than the band you can screenshot.
Roughly 5× to 15×: the post left your audience
At this scale the post is no longer being carried by people who follow you. It has been pushed to people who do not, which is a different mechanism — Instagram has been explicit that Reels is built to surface content from accounts you do not follow, and sends per reach is reported as the signal that matters most for that.
So the lesson from this band is rarely “this topic is good”. It is “this post was legible to people with no context”. Those are different properties, and the second one is the one to isolate.
Over about 15×: check for a cause outside the post
Very large multiples are worth investigating before they are worth imitating. A meaningful share of them have an external explanation: the post was shared by a much larger account, it attached to a moment that was already moving, it was boosted, or it caught a platform test.
None of that makes the post bad. It makes it a poor teacher, because the thing that produced the number is not a thing you can put in the next video. The question to answer before drawing any conclusion is simply what else happened.
The comparison that is never valid
Multiples are comparable within an account and across accounts. They are not comparable to raw view counts, and the most common analytical error in short-form is sliding between the two — studying a post because its view count was large rather than because its multiple was.
That error has its own piece: breakout Reels vs popular Reels. It is the single most expensive habit in content research, because it sends you to study the posts with the least to teach.