How to read the Reels retention graph
The retention graph plots the share of viewers still watching at each second of your reel. A steep drop at a certain second means something in that moment lost people, so line the drop up with your timeline and look at the shot. The first seconds always drop most; what matters is the shape after that.
Skip rate and average watch time give one number each. The graph tells you where. It is the nearest thing Instagram gives you to an edit note, and most creators never open it.
Where do you find the retention graph?
Open the reel’s insights and look for the retention graph. The time axis runs along the bottom and the share of viewers still watching up the side. Some graphs also show a “typical” line from your other reels; use it if it is there.
What do the shapes mean?
| Shape | What it suggests | What to look at |
|---|---|---|
| Cliff in the first 1–3 seconds | The opening didn’t earn attention | The first frame and first line of text |
| Gradual slope to the end | Normal: people leave a little at a time | Compare with your typical line |
| Sharp step at one second | Something at that moment lost people | The shot or cut at that second |
| Flat stretch | Viewers who stayed are hooked | What was on screen: copy it |
| Bump back up | Viewers rewound or replayed that moment | The payoff, usually |
What we see in our own data: on @tomeescapes (data as of 26 October 2026), the 4 Reels retention graphs we have read from screenshots had a median of 40% still watching at 3 seconds and 7% at the end (median length 11 seconds). The 10 YouTube Shorts were different: 122% at 3 seconds and 47% at the end, because YouTube counts replays in the graph. 4 Reels graphs is a small sample; it is not a benchmark. The two platforms measure differently, so don’t compare them number for number.
How do you match a drop to a shot?
- Read the second of the drop off the graph. Note the second before and after.
- Open the exported video or your timeline, and go to that second. Retention graphs are slightly rough, so look a second either side.
- Name what is on screen: a title card, a slow pan, a cut to a different place, a talking head, a loss of motion.
- Check the same moment in another reel that has the same kind of shot. One drop can be chance; the same drop twice is a pattern.
- Change that one thing in your next reel (shorter, move it later, add motion) and compare the graph.
Our own tracing of these graphs from screenshots is accurate to roughly ±3 percentage points, so ignore wiggles smaller than that.
What does a worked example look like?
Say a 12-second reel shows 45% still watching at second 3, a steady slope to 30% at second 8, then a step down to 15% at second 9. The first 3 seconds are close to your usual, so the opening is fine. The step at second 9 is the finding: open the video at second 8 to 9 and see what changed. It might be a cut to a different place, a text card, or the end of the music. Then look at whether your last two reels had the same thing at the same point. If they did, you have a pattern, and the next reel can test the alternative. If they didn’t, it may just be this reel.
What makes a retention graph misleading?
- Few viewers. A graph built on a few hundred plays is noisy. Wait until the reel has reached enough people, and read several reels before you decide.
- Different lengths. A drop at second 6 in a 10-second reel is not the same thing as in a 30-second one. Compare reels of similar length.
- Trials against normal reels. A trial is shown only to non-followers, who leave sooner. Compare like with like.
- Reading one reel. One graph tells a story; three graphs with the same drop tell you what to change.
How is this different from skip rate and watch time?
- Skip rate is the first 3 seconds, as one number.
- Average watch time is the whole reel, as one number.
- The graph shows both, second by second. Use the numbers to find the weak reels and the graph to find the weak moment.
Visteria reads the retention graph from the Insights screenshot you upload and lines the drops up with the shots in your video, so you don’t have to scrub the timeline yourself. See what you get.
Questions people ask
What is a good retention graph on Instagram Reels?
One that stays above your typical line and drops gradually rather than in steps. Compare against your own reels; there is no universal target.
Why does the retention graph drop so much in the first seconds?
Because viewers decide quickly whether to stay. The first seconds always lose the most people, so what matters is how steep the drop is compared with your other reels, and the shape after it.
Why does the graph go back up?
When viewers rewind or replay a moment, that moment counts more than once. A bump usually marks a payoff or a loop that works.