← blog

What a Good Retention Curve Looks Like

2026-09-09 · Nik · updated 2026-08-29

A healthy retention curve drops sharply in the first fifteen seconds, then flattens into a slow decline. The initial drop is normal — it is viewers deciding. What matters is where the curve settles: holding above 40% at the halfway mark is strong for long-form, and any upward bump marks a moment worth repeating.

Where to find it

YouTube Studio → pick a video → Engagement → Audience retention. Two numbers sit above the graph: average view duration in minutes and seconds, and average percentage viewed. The graph itself is more useful than either.

The four shapes

1. The cliff

Steep fall in the first 15–30 seconds, then a low flat line.

Means: the opening is not delivering what the thumbnail and title promised.
Fix: re-cut the first fifteen seconds. Lead with the payoff. This is the single highest-value edit in video.

2. The slide

Steady, unbroken decline from start to finish.

Means: nothing is wrong, nothing is gripping. Common on tutorials that answer the question early.
Fix: structure. Put a reason to stay at the two-minute mark — a second question, a demonstration, a result.

3. The plateau

Sharp initial drop, then a nearly flat line to the end.

Means: the best shape there is. The people who stayed are the right audience.
Fix: nothing. Promote this video. Make more like it.

4. The spike

Curve rises at a specific point.

Means: viewers are rewinding or arriving late to that moment. It is the most valuable second in your video.
Fix: find out what it is, and open your next video with that.

Benchmarks worth using

Video length Weak Decent Strong
Under 3 min <40% 50–60% >70%
5–10 min <25% 35–45% >50%
10–20 min <20% 30–40% >45%
Over 20 min <15% 25–35% >40%

Percentages fall as length rises — that is expected, and it is why absolute watch time matters more than the percentage for monetisation.

The first 30 seconds decide the rest

Most of the variance between a video that spreads and one that stalls happens before the half-minute mark. Retention there is a function of three things: whether the opening matches the promise, whether something is actually happening on screen, and whether the viewer can tell where this is going.

If your curve drops below 60% at 30 seconds, that is the place to work. Nothing further down the timeline will save it.

What promotion does to the curve

Advertising does not change the shape of the curve — it changes how many people are on it, and it tests it against a wider audience than your subscribers.

That is useful in a specific way. Your existing audience is forgiving; strangers are not. A campaign shows you the retention your video earns from people who have never seen you before, which is exactly the audience the recommendation system would send it to.

Curve Is promotion worth it?
Plateau Yes — this is the ideal candidate
Spike Yes — and open the next video with the spike moment
Slide Maybe — try a small test, watch whether a new audience behaves differently
Cliff Not yet — fix the opening first

Promoting a cliff-shaped video is the most common way to waste a budget: it delivers the same verdict as before, just to more people.

Reading a campaign afterwards

After a week of delivery, compare the campaign period against the video's lifetime average:

  • Average view duration up — the targeting found a better-matched audience than your organic reach
  • Roughly flat — the audience was right, the video is what it is
  • Down — targeting was too broad; narrow the country or interest and retest

Frequently asked questions

What is a good average view duration?

For a ten-minute video, three to four minutes is solid, five is strong. Absolute duration matters more than percentage once you are chasing watch hours.

Should I make shorter videos to improve retention percentage?

Only if the content is genuinely shorter. Padding a five-minute idea into twelve minutes hurts retention; cutting a twelve-minute idea to five loses the value.

Does the curve include ad views?

Yes. Everything that counts as a view feeds the retention graph.

How many views before the curve is trustworthy?

A few hundred gives you a shape. A thousand makes it reliable enough to act on.


Once you have a video that plateaus, see what a week of promotion delivers on the pricing page.

Put this into practice

Pick a budget, see the reach it buys, launch in minutes.

Start a campaign