Reading Audience Retention in YouTube Studio
For a channel of short explainers, the retention data in YouTube Studio is the closest thing you have to watching over a viewer's shoulder.
Start with whether people stayed at all
For Shorts, the first question is whether viewers watched or swiped past. YouTube Studio reports this with two related metrics. According to YouTube Help's Shorts content tab tips, Shown in feed counts the times your Shorts appeared in the Shorts feed, and viewed vs. swiped away shows the percentage of times viewers watched your Shorts compared with swiping away.
This number reflects the very start of the video: the first frame, the first line and how well they match what the viewer expected. If a large share of people swipe away, the problem is almost always in the opening seconds, not in the middle of the script. Look at the first frame and the first spoken line together and ask whether they give a reason to stay. Our post on YouTube Shorts titles and thumbnails covers how to make the title, cover frame and opening agree.
Then look at how long they watched
Once someone has chosen to watch, the next questions are how long they stayed and how much of the video they saw. YouTube Help on content performance describes average view duration as the average time watched and average percentage viewed as the average share of the video watched.
For Shorts, average percentage viewed is usually the more useful of the two, because your videos differ in length. A forty-second Short and a fifty-five-second Short cannot be compared fairly on time watched, but they can be compared on share watched.
Compare each video against your own channel rather than against an imagined target. What matters is whether a video held viewers better or worse than your typical upload, and why.
Read the shape of the curve
The retention graph for an individual video shows how many viewers were still watching at each point. Look for a few features. Top moments are stretches where almost no one dropped off. Spikes are moments viewers rewatched or shared. Dips are places where viewers left or skipped ahead.
Each feature points to a specific part of your script:
- A steep fall right at the start means the opening did not hold. This usually matches a high swipe-away rate.
- A dip in the middle often lines up with a sentence that breaks the chain of reasons, such as a digression, a dense list of numbers or a screen that sits still too long.
- A spike shows something viewers wanted to see again, which is worth noting as a pattern to repeat.
- A fall just before the end can mean the payoff arrived too early and the remaining seconds felt like padding.
To use this, open the script next to the graph and match the timestamps. The dip at twenty-two seconds is a specific sentence and a specific screen. Once you can name it, you can decide whether to cut it, move it or show it differently.
Look for patterns, not single results
One video's retention can be affected by many things at once, including the topic, the time it was published and who the platform happened to show it to. Changing your whole approach because of one weak curve is usually a mistake.
Instead, review retention across a batch of videos. If dips keep appearing where scripts list several numbers in a row, that is a writing habit worth changing. If videos built around a direct question consistently hold better than those that open with a statement, that tells you something about your audience. The guide on how to make YouTube Shorts with Claude Code shows how to feed lessons like these back into the rules a production pipeline follows.
Turn findings into rules
The point of reading retention is to change what you make next. When you find a pattern, write it down as a rule for future scripts. "Never put more than one number in a sentence" or "show the map within the first two seconds" is something you can follow on the next upload. "Make videos more engaging" is not.
Keep the list short and specific, and remove rules that stop proving useful. Over time this list becomes the most valuable document your channel has, because it is written from your own viewers' behaviour rather than from general advice.
A simple weekly review
- Sort recent Shorts by viewed vs. swiped away and study the openings of the lowest.
- Compare average percentage viewed across the same videos.
- Open the retention graph for the best and worst, and match dips and spikes to the script.
- Write down one pattern and turn it into a rule.