What is a good audience retention rate? YouTube, TikTok, Reels

Audience retention is how much of a video people watch: as a curve (the share of viewers still watching at each moment), as an average (average percentage viewed, or average watch time in seconds), or as a single moment (how many got past the first seconds). A good audience retention rate is one that's higher than your other videos of the same format and length, because none of the apps publishes a number to hit:

• YouTube grades a video against your own: its audience retention report draws your typical retention, from your 10 latest videos of similar length, and can compare it with all YouTube videos of similar length.
• TikTok shows each video's retention curve, average watch time and how many watched the full video, with no benchmark.
• Instagram shows each reel's watch time, average watch time, a retention chart and a skip rate, also with no benchmark.

Third-party studies (further down) give rough ranges. The rule that holds everywhere: retention falls as a video gets longer. A 10-minute video that keeps 30% of viewers to the end and a 30-second clip that keeps 60% aren't comparable, so compare each video with your own of a similar length. Retention is how long people watch; whether they react is your engagement rate, a different number.

In YouTube Studio, open Content, pick a video, select Analytics, and look for Audience retention on the Overview or Engagement tab. Retention data typically takes one to two days to process. YouTube gives you two comparisons there, and they're the closest thing it has to a definition of good:

• Typical retention: the retention of your 10 latest videos of similar length, drawn with your video's curve. Above it means this video held people better than you usually do.
• All YouTube videos of similar length: select See more on the report. YouTube's Analytics API describes the same comparison as a score from 0 to 1 at each point of the video, where 0.5 is the middle: half of similar videos keep viewers better, half worse.

The same report highlights key moments on videos at least 60 seconds long with at least 100 views. Intro is the share of viewers still watching after the first 30 seconds; YouTube says a high one can mean the opening matched what the title and thumbnail promised. Top moments are where almost no one left, spikes where people rewatched or shared (or had to rewatch something unclear), and dips where they skipped or stopped. YouTube says a flat line means people watch that part from start to finish, and that videos generally taper off as they play. You can also split retention by new and returning viewers and by subscribers and non-subscribers.

YouTube Analytics shows two averages, and its help pages define both the same way:

• Average view duration is the average minutes watched “among those who stayed to watch”.
• Average percentage viewed is the average share of the video watched among those who stayed to watch.

Both are worked out from engaged views (people who clicked to watch or watched past the first seconds) and their watch time. That matters since August 24, 2026, when YouTube started counting a public view the moment a video starts to play, autoplay previews included: YouTube says CTR, average view duration and retention are still calculated the way they were, so you can compare them with your older videos.

A 10-minute video with a 3:00 average view duration has an average percentage viewed of 30%; a 30-second Short watched for 15 seconds on average is at 50%. Which one matters more? YouTube says its systems use both, and that broadly, relative watch time (the percentage) matters more for short videos and absolute watch time (the minutes) for long ones. So a longer video can be worth making even when its percentage is lower, as long as the minutes go up.

A Short in the Shorts feed has no thumbnail to click: it just starts playing. So the first number to read is whether people stay. In YouTube Analytics, the Content tab's Shorts view has a How viewers engaged report: the percentage of times viewers stayed to watch past the first seconds of a Short. YouTube's Shorts tips page calls it How many chose to view, and its blog described it in March 2024 as viewed vs swiped away, the Shorts version of impressions and click-through.

YouTube's Shorts help page says its ranking uses three of these numbers: the share of viewers who chose to view, average view duration, and average percentage viewed, then likes and surveys to judge whether people enjoyed it. Two things to know before you compare Shorts:

• The public view count isn't the stayed-to-watch count. Since March 31, 2025, a Short's views count every play and replay, with no minimum watch time. Engaged views and the averages above don't count loops.
• Shorts and long videos don't compare. A Short is seconds long, so its percentage runs far higher. Judge Shorts against Shorts.

So what is a good retention rate for a Short? YouTube publishes no number, and we found no independent study we could check. YouTube gives you the same two comparisons as for long videos. On a video's audience retention report, typical retention is drawn from your 10 latest videos of similar length, which for a 30-second Short means your other short videos, and See more compares it with all YouTube videos of similar length. YouTube's help pages describe both for every video, with no separate rule for Shorts. A Short above your typical line, and above the middle of similar videos, held people better than usual.

How the Shorts feed uses these, and what else it weighs, is in how the YouTube Shorts algorithm works.

TikTok's analytics (in TikTok Studio, and under each video's analytics in the app) show three retention numbers for every video. TikTok's Creator Academy calls them the key metrics to watch if you want to improve your videos:

TikTok has said since 2020 that finishing a longer video from beginning to end is a strong sign of interest, weighted more heavily than weak signals. Average watch time in seconds rises with length while the finish rate usually falls, so look at both together, and compare videos of similar length. More on how TikTok uses these is in how the TikTok algorithm works.

Instagram's Help Center lists these insights for every reel: views (each time it starts to play or replay), watch time (all the time it played, replays included), average watch time (watch time divided by initial views, the first play in a session), viewers and follows. In August 2025 Instagram announced two more, as Social Media Today reported: a retention chart showing where people stop watching, and a skip rate replacing view rate. In April 2026, Social Media Today reported, Instagram's redesigned Insights for posts and reels added share rate and skip rate percentages. Meta's developer documentation defines the reels skip rate as the percentage of views skipped in the first 3 seconds, and labels it estimated and in development.

Instagram's own ranking explainer (2023) says the most important predictions it makes for Reels include how likely you are to watch a reel all the way through. So the job on Instagram is the same as everywhere: get past the first 3 seconds, then keep people to the end.

These are the third-party studies with a stated sample that we could read in full. Each defines retention its own way:

For scale, vidIQ published its own channel's numbers (2.39 million subscribers, the 12 months to June 2026): 30.3% average percentage viewed across 16.5 million long-form views, and 73.6% on its Shorts. That's one channel, not a benchmark, but it shows how far apart the two formats sit.

Be wary of round numbers without a study behind them. The often-quoted “50 to 60% is average” for YouTube usually comes with no source; the closest we found is a marketing agency's look at 56,000 views across four small channels. Retention Rabbit measured the videos of people using its retention tool, Socialinsider and Dash Social measured brand accounts, and Metricool doesn't say how it chose its accounts, so none of them is a random sample of creators. Use them to sanity-check your own numbers, not as targets.

The curve tells you where people leave; these are the fixes for the places they usually leave:

• The first seconds (the hook). Most of the loss happens at the start: the skip in the first 3 seconds on Reels, stayed to watch on Shorts, the 30-second intro on long videos. Open on the line that matters, not a greeting, and deliver what the title or caption promised. Examples are in short video hooks, and the hook generator suggests opening lines.
• Dead air and pauses. Every gap is a moment to swipe. Cut ums, false starts and silences; the free remove silence tool cuts the pauses out of a talking video automatically, and lets you hear it without them before you download.
• Length that fits the story. A clip should end when the point lands, not at a target length. More in how long a short video should be.
• Captions on every word. They keep people following with the sound off, and give the eye something to track between cuts.
• Pattern changes. A new shot, a zoom, an on-screen word or a change of speaker resets attention, especially before a spot where your curve dips.
• End on the payoff. Stop on the punchline or the answer. A clip that trails off after its best line loses people right before the end, which is what the finish rate counts. More in how to make viral clips.

By hand, that's finding the moments worth a clip in a long video, then trimming each to start on its hook and end on its payoff. Clyper does that part from a YouTube link or an upload: it reads the transcript to find the best moments, cuts each on whole sentences so nothing starts mid-word, gives every clip a hook score so you can post the strongest first, frames it 9:16 on whoever's talking, and captions every word. You can trim or extend either end in the editor.

780:22

740:39

700:57

Then measure: after a few days, compare each clip's retention with your others of similar length, and its reactions with the engagement rate calculator.