When a client tells me their views are lower than expected, the first thing I check is their retention data. Not their view count. Not their subscriber growth. Their retention curves. Because retention is the metric that explains almost everything else: why the algorithm is or isn't recommending the content, why views are growing or stalling, and whether the content itself is actually delivering value to the people who click on it.

YouTube's algorithm in 2026 is fundamentally a satisfaction prediction engine. It tries to predict which videos a viewer will enjoy, and the strongest signal it has for that prediction is how long previous viewers watched. A video that holds 60 per cent of its audience through to the end is a video the algorithm has confidence in. A video that loses 70 per cent of its viewers in the first 30 seconds is a video the algorithm stops recommending.

This guide covers the specific retention benchmarks that matter in 2026, segmented by video length, content type, and format (long-form vs Shorts). If you want the broader picture of what "normal" looks like for a new channel across all metrics, our parent guide on average views for new YouTube channels covers the full landscape.

Average View Duration: What the Metric Actually Measures

Average view duration (AVD) measures the average amount of time a viewer spends watching your video before leaving. If your 10-minute video has an AVD of 4 minutes, viewers are watching 40 per cent of it on average. YouTube Studio reports this both as an absolute number (minutes and seconds) and as a percentage of total video length.

The percentage figure is more useful for benchmarking because it normalises across video lengths. Four minutes of watch time on a 10-minute video is 40 per cent. Four minutes on a 30-minute video is 13 per cent. The absolute number is the same, but the percentage tells you something very different about how well each video held its audience.

A common point of confusion: average view duration and audience retention rate measure related but distinct things. AVD gives you a single number representing the average. The retention curve shows you where viewers drop off, which is diagnostically much more valuable. Both are available in YouTube Studio under the Engagement tab.

Retention Benchmarks by Video Length

Retention expectations shift significantly with video length because viewer commitment varies. Someone who clicks on a 3-minute video has a different tolerance for pacing and density than someone who clicks on a 45-minute deep dive.

For videos under 5 minutes, a healthy average percentage viewed is 50 to 70 per cent. These are short enough that most viewers who click will watch a meaningful portion, but the opening seconds are disproportionately important because there's no runway to recover from a weak hook. If your sub-5-minute videos are below 50 per cent, the hook is likely the problem.

For videos between 5 and 15 minutes, the most common range for business and educational content, a healthy average percentage viewed is 40 to 55 per cent. This is the length where pacing becomes critical. A 10-minute video with 40 per cent retention means the average viewer watched 4 minutes. That's enough time to deliver substantial value, but it also means 60 per cent of the content goes unseen by the average viewer. The question is whether the most important content is front-loaded or buried in the second half.

For videos between 15 and 30 minutes, 30 to 45 per cent is healthy. Longer content naturally has lower percentage retention because the time commitment is higher and viewers are more likely to find their answer partway through and leave. Research on YouTube watch time benchmarks confirms that 50 per cent retention on content over 15 minutes is exceptional and typically indicates a highly engaged, niche-specific audience.

For videos over 30 minutes, 25 to 35 per cent is normal. Podcast-format content, long-form interviews, and comprehensive tutorials fall into this range. The absolute watch time is still high (25 per cent of a 40-minute video is 10 minutes), which provides strong signals to the algorithm even though the percentage looks modest.

The critical insight: do not compare percentage retention across different video lengths. A 35 per cent retention on a 30-minute video represents 10.5 minutes of watch time, which is a much stronger algorithmic signal than 60 per cent retention on a 3-minute video (1.8 minutes). YouTube's algorithm weights total watch time heavily, which is why longer videos that hold a decent percentage often outperform shorter videos with higher percentage retention in terms of algorithmic distribution.

Retention Benchmarks by Content Type

Different types of content retain audiences differently because the viewer's relationship with the material varies.

Tutorial and how-to content typically retains well in the first half and drops off once the viewer finds the specific answer they were looking for. A 45 to 55 per cent retention is healthy. With the 2026 algorithm's "good abandonment" framework, a viewer who leaves after finding their answer is no longer penalised. If someone watches 3 minutes of your 10-minute SEO tutorial, finds the setting they needed to change, and leaves YouTube entirely without searching for the same topic again, that's a positive signal.

Thought leadership and commentary retains differently: it holds a loyal audience well (the people who care about your perspective) but has a steeper early drop from viewers who clicked out of curiosity but don't connect with the format. A 35 to 50 per cent retention is healthy, with the shape of the curve mattering more than the number.

Product demonstrations and case studies tend to have the most variable retention. Viewers who are actively evaluating a product will watch most of the video. Viewers who are casually browsing will leave quickly. A bimodal retention curve, where some viewers watch nearly all of it and others drop off early, is typical and healthy for this content type.

Interviews and podcasts have the lowest percentage retention but often the highest absolute watch time. A 25 to 35 per cent retention on a 60-minute interview represents 15 to 21 minutes of viewing, which is a very strong signal. The audience self-selects: people who click on a long-form interview are generally prepared to commit significant time.

What Our Own Data Shows

I'll share some numbers from the Humble&Brag channel because I think transparency about our own performance makes these benchmarks more credible than citing generic studies.

Across our video library over the past 28 days, our channel-wide average percentage viewed is 19.43 per cent, which sounds low until you factor in that our average video duration is over 10 minutes. Our highest-retaining long-form videos sit in the 33 to 43 per cent range: "5 Lesser Known YouTube Growth Tactics" holds 42.6 per cent retention across its full length, "How to Use Google Ads to Grow Your YouTube Channel" holds 36.8 per cent, and "These 7 Formats Will Grow Your YouTube Channel" holds 32.7 per cent. What these three have in common is dense, actionable content with a clear structure that gives viewers a reason to keep watching through each section.

At the other end, our videos that underperform on retention (below 15 per cent) tend to be longer pieces where the specific answer the viewer searched for appears in the first few minutes. That's not a content quality failure; it's the "good abandonment" pattern working as intended. The viewer found what they needed and left satisfied.

The most useful comparison isn't between our best and worst videos. It's between videos of similar length and format. Our talking-head videos between 8 and 15 minutes average roughly 20 to 25 per cent retention. Our more structured, format-driven pieces in the same length range average 30 to 37 per cent. The format and pacing make the difference, not the topic.

Reading Your Retention Curves: Four Diagnostic Patterns

The retention curve in YouTube Studio is the single most valuable diagnostic tool available to you. I review these for every client video at Humble&Brag, and the shape of the curve tells a specific story.

The Cliff: A steep drop in the first 15 to 30 seconds followed by a relatively flat line. This means your hook is failing. The viewers who survive the first 30 seconds are engaged, but you're losing the majority before they reach the substance of the video. The fix is in the opening: deliver on the title's promise immediately, introduce the specific value the viewer will get, and cut any preamble that delays the payoff. At CareerFoundry, we found that restructuring our video openings to lead with the key insight rather than a channel introduction improved retention past the 30-second mark by 15 to 20 percentage points.

The Gradual Decline: A steady, even slope downward throughout the video. This is the most common pattern and it's generally healthy if the retention at the midpoint is above 40 per cent. The decline represents the natural attrition of viewers who find their answer or lose interest as the video progresses. If the decline is steeper than expected, pacing is usually the issue: the content is too slow, too repetitive, or not structured with enough "pattern interrupts" (visual changes, new segments, questions) to maintain engagement.

The Bump: A visible spike at a specific point in the video, often indicating a moment that viewers rewatch or skip forward to. If the bump aligns with a particularly valuable segment, consider creating a chapter marker there or even extracting that section as a standalone Short. Bumps are gold in your retention data because they tell you exactly what your audience values most.

The Flat Line: Retention that holds steady through most of the video with minimal decline. This is the ideal pattern and it's rare. It indicates that the content is consistently engaging from start to finish and that the audience is well-targeted. If you see this on a video, study everything about it: the topic, the title, the hook structure, the pacing. It's your template for future content.

Shorts Retention Benchmarks

YouTube Shorts retention works differently because the format and viewing context are different. Viewers don't choose Shorts deliberately; they scroll through a feed and decide within the first one to two seconds whether to keep watching or swipe past.

For Shorts, the primary retention metric is average percentage viewed, where 70 per cent or above is healthy. Below 50 per cent indicates a structural problem with either the hook or the pacing. Above 100 per cent means viewers are replaying the Short, which is one of the strongest positive signals available.

The first three seconds are the "swipe or stay" window. If your Shorts retention curve shows a sharp drop in the first three seconds, the opening frame and first spoken words aren't earning the viewer's attention. For a deeper dive into all Shorts-specific metrics, see our dedicated YouTube Shorts benchmarks guide.

Like-to-View Ratio Benchmarks

Like-to-view ratio is an engagement metric that correlates with retention but measures a different signal: emotional response. A viewer can watch your entire video without liking it (they found it useful but not remarkable), or they can like it within the first minute because the opening resonated strongly.

For long-form content in 2026, a healthy like-to-view ratio is 4 to 8 per cent. Above 8 per cent indicates content that's resonating emotionally, not just informationally. Below 3 per cent suggests the content is useful but not memorable enough to prompt an active response.

For Shorts, like-to-view ratios are typically lower (3 to 6 per cent) because the viewing experience is more passive. Shorts viewers are scrolling through a feed, and the threshold for an active engagement action is higher relative to the effort of watching.

For business channels specifically, like-to-view ratio is a secondary metric. It's useful as a health check but shouldn't drive content decisions. The metrics that predict business outcomes, retention, return viewer rate, and traffic source distribution, are more directly connected to the viewer journey that leads to conversion.

The Metric That Matters More Than Retention: Return Viewer Rate

Retention tells you whether a single video held the audience. Return viewer rate tells you whether the audience comes back for the next one. For business channels, the second metric is more important because it's what drives compounding growth.

A return viewer rate above 10 per cent, meaning more than 10 per cent of the people who watched a previous video come back for the next one, indicates a channel that's building a real audience rather than just attracting one-time viewers. We covered this in depth in our guide on what "normal" views look like for new channels and why return viewer rate is a better predictor of long-term success than any single-video metric.

When Retention Is a Problem vs When It's Normal

Low retention is a problem when it's combined with other weak signals: low CTR, declining impressions, flat subscriber growth. In that case, the content isn't holding the audience the algorithm is sending, and the algorithm responds by sending fewer viewers.

Low retention is normal when the content is intentionally serving a quick-answer function (tutorials where the viewer finds their answer and leaves), when the content is very long (percentage retention on 30+ minute videos will always look modest), or when the channel is new and the algorithm is still testing different audience segments.

Before assuming your retention is a problem, check it against the benchmarks for your video length and content type. If it's within the healthy ranges listed above, the retention itself isn't the issue. The issue is more likely in your packaging, your CTR, or your topic selection.

If you want a professional assessment of your channel's retention health alongside CTR, traffic sources, and content strategy, a YouTube channel audit covers all of these dimensions and produces a prioritised action plan.