Most video analytics measure your audience size back at you
A plain definition, the four numbers that actually separate good Reels from average ones, and the ones to ignore. Built from 2,518 analysed Reels. 2026-10-08.
The definition
Video analytics is the measurement of what happens to a video after you publish it: how many people saw it, how many stayed, how many reacted, and how those compare to something. The last clause is the one most dashboards skip, and it is the only part that turns a number into information.
Why raw numbers mislead
A reel with 131,286 views is exactly median in our sample. Whether that is good depends entirely on the account: on 10,000 followers it is thirteen views per follower and remarkable, on three million it is a failure. Any metric that does not divide by audience size is largely reporting how big you already are.
The four numbers worth reading
Views per follower has no useful single median across accounts of wildly different sizes, which is the point: compute it for your own account.
| Metric | How to read it | Median here |
|---|---|---|
| Views per follower | The only cross-account comparison that works. 3x is an outlier floor. | See note |
| Like rate | Likes divided by views. Low means it reached people who did not care. | 3.11% |
| Comment rate | Comments divided by views. The harder action, so a better signal. | 0.070% |
| Views against your own median | Your baseline. 3x your own median is the goal worth setting. | 131,286 across all accounts |
The numbers to treat carefully
Retention and hook scores produced by AI tools, including ours, are estimates generated by a model reading the video. They are not measured completion data from the platform. They are useful for ranking your own drafts against each other and should not be quoted as a percentage of people who watched to the end.
Our own hook score is a good example of the limit. Grouped into nine families of opening, every family averages between 8.5 and 9.1 out of ten, a spread of six tenths of a point across more than two thousand Reels. It does not discriminate between hook types, even though their actual median views differ by about three times.
What a useful analysis actually tells you
- Which of your Reels beat your own median, and by how much.
- What those Reels had in common in the first three seconds.
- Whether your like and comment rates are above or below the median for comparable work.
- Which length band your Reels fall in, and whether any sit under 15 seconds.
- What a creator in your niche does differently, read from their published Reels rather than guessed.
Figures on this page come from 2,518 Instagram Reels across 528 accounts, analysed by Scrptly. The source table holds 3,554 rows; it is deduplicated to 2,518 unique Reels so no reel is counted twice. Read 2026-10-08.
Run your own account or a competitor's and get every one of these numbers in one pass.
Analyze a creator freePut your own idea through the same checks before you film it.
Create content