• YouTube
  • Growth
  • Creators

YouTube Analytics Explained: Which Metrics Matter and What’s Normal

YouTube’s definitions of impressions, CTR, engaged views, retention and returning viewers, its only published benchmark (2–10% CTR) and how to diagnose a video.

Joanna FinchPublished Last updated 12 min read
Angled close-up of a dashboard on a screen with dark tiles, small line charts and a click-through rate tile reading 14.65%
Short answer

Seven YouTube Analytics metrics explain most of a video’s performance: impressions, impressions click-through rate, engaged views, average view duration, average percentage viewed, the audience retention graph, and new versus returning viewers. YouTube’s only published numeric benchmark is that half of all channels and videos have a click-through rate between 2% and 10%. Since August 24, 2026, a public view counts from the first frame, so compare engaged views, and judge every metric against your own recent videos, by traffic source.

YouTube Studio shows dozens of numbers, and most creators look at the wrong ones first. The metrics that tell you what to change next sit a level down: how often people click, how long they stay, whether they come back, and where they came from.

This guide is for creators reading their own analytics and for the brand managers who get sent screenshots of them. Every definition comes from YouTube’s Help Center and official blog, fetched on October 16, 2026; where YouTube publishes no benchmark, we say so. YouTube has also been rolling out an updated Studio design since July 2026, so a card may have moved.

Which YouTube Analytics metrics matter most?

The short list is seven metrics, in the order a viewer meets your video: impressions, click-through rate, engaged views, average view duration, average percentage viewed, the retention graph, and returning viewers. Traffic sources sit alongside them, because they change what every other number means.

Define the terms once (the table further down says which Studio tab holds each):

  • An impression is a thumbnail shown on YouTube for more than one second with at least half of it visible.
  • Impressions click-through rate (CTR) is how often viewers watched a video after seeing one of those impressions.
  • A view now counts the moment a video starts to play; an engaged view counts when someone clicks to watch or keeps watching past the first seconds.
  • Average view duration (AVD) is the average minutes watched among those who stayed to watch.
  • Average percentage viewed (APV) is the average share of a video watched among those who stayed.
  • Audience retention is the graph of how many viewers are still watching at each moment of a video.
  • Returning viewers are people who had watched your channel before and came back in the period you selected.

What changed about YouTube views on August 24, 2026?

Since August 24, 2026, YouTube counts a view the moment a video starts to play, across Shorts, long-form videos and live streams, and the old definition is now called an engaged view. YouTube’s help page on how engagement metrics are counted gives the date and says Partner Program earnings still rest on engaged views and engaged watch hours, and eligibility on qualified views.

The edge cases are in YouTube’s blog post of August 19, 2026, What are ‘Engaged’ Views on YouTube?. A thumbnail you see but never touch stays an impression. One you hover over until it autoplays is a view from the first frame, and an engaged view only if it keeps playing “for some amount of seconds”, a number YouTube doesn’t publish.

The post adds that “the vast majority of analytics remain anchored on engaged views, so the way metrics like CTR, AVD, and retention are calculated hasn’t changed.” The Content tab defines engaged views as “how many times viewers stayed to watch past the initial seconds, not including any loops.” So compare engaged views, not public counts, across the August boundary; our guide to how much YouTube pays per 1,000 views explains why RPM now runs on engaged views. Eligibility still uses qualified views and watch hours, covered in the YouTube monetization requirements.

What is a good click-through rate on YouTube?

YouTube’s own figure is that “half of all channels and videos on YouTube have an impressions CTR that can range between 2% and 10%.” That sentence is still on the Impressions and click-through-rate FAQ as of October 16, 2026, and it’s the only CTR benchmark YouTube publishes. It isn’t a target: half of channels sit outside it, and CTR “will vary based on the type of content, audience, and where on YouTube the impression was shown.”

Three things move CTR that have nothing to do with your thumbnail:

  1. Traffic source. A video shown widely, for example on Home, naturally has a lower CTR; one whose impressions come mostly from your channel page may run higher. A falling CTR on a video that’s gaining impressions can be good news.
  2. Age and size. New videos or channels, “like those less than a week old”, or videos with fewer than 100 views, show an even wider range. Don’t check CTR immediately after uploading, and act only on a statistically significant change.
  3. Uncounted impressions. External sites and apps, the YouTube Kids and YouTube Music apps, cards, end screens, notifications and thumbnails under 50% visible don’t register impressions, so with heavy external traffic you may see more views than impressions.

YouTube also warns against swapping several thumbnails on one video, since each may reach a different traffic mix, and it defines clickbait precisely: “high CTR but low average view duration and lower than expected Impressions.”

What is a good average view duration or average percentage viewed?

YouTube publishes no benchmark for either, but it does say which one to weigh: “relative watch time is more important for short videos and absolute watch time is more important for longer videos.” That line is from the YouTube performance FAQ, which says the discovery system uses both absolute and relative watch time as signals.

So a 30-minute video with 7 minutes of average view duration (23% viewed) can be healthy, and a 90-second video with 35 seconds (39%) can be weak. Those examples are ours, to show the arithmetic.

Both metrics ignore early leavers. The Content tab defines average view duration as “average minutes watched among those who stayed to watch” and average percentage viewed as the “average percentage of a video watched among those who stayed to watch”, both “calculated from engaged views and their corresponding watch time.” People who left in the first seconds show up instead as the gap between views and engaged views.

Use the comparisons YouTube builds in instead: SEE MORE on the retention report compares a video with all YouTube videos of similar length, and the “typical retention” line compares it with your own 10 latest videos of similar length.

How do you read the audience retention graph?

The audience retention report, available at video level only, marks four kinds of moment, and YouTube’s key moments page says what each one usually means. Data takes one to two days to process, and highlights appear only on videos at least 60 seconds long with at least 100 views.

  • Intro is the percentage still watching after the first 30 seconds. A high figure can mean the opening matched the thumbnail and title; if it’s low, YouTube suggests changing them to reflect the content, or reworking the first 30 seconds.
  • Top moments are stretches where almost nobody left. If they come late, bring that material forward, because “audience sizes typically decrease over the length of the video.”
  • Spikes are segments rewatched or shared: the best part, or one that wasn’t clear the first time.
  • Dips are segments viewers skipped or abandoned.

A gradual decline is normal; videos “generally taper off during the playback period.” The segment drop-down compares new with returning viewers, subscribers with non-subscribers and organic with paid traffic.

Black monitor on a desk showing a white line chart rising across a dark grid, with a blurred frosted-glass office behind it
The shape of the line matters more than any single number: YouTube’s report marks the intro, the top moments, the spikes and the dips for you.

What are new, casual, regular and returning viewers?

YouTube splits your monthly audience, the unique viewers of the last 28 days, into new, casual and regular viewers, and separately reports returning viewers for any date range. The definitions are on YouTube’s Audience tab page and its page on new, casual and regular viewers:

  • New viewers watched your channel for the first time in the period; private-browser viewers, people who deleted their watch history and anyone away for over a year count as new.
  • Casual viewers watched at least once a month for one to five months in the past year.
  • Regular viewers watched at least once a month for more than six months in the past year. YouTube calls them your most loyal audience members and says a regular share lower than you expect, “as low as” under 1%, is common, especially for newer channels, trending videos and Shorts-heavy channels.
  • Returning viewers had watched before and came back in the selected period.

Two YouTube statements keep these in proportion. “Audience segments do not affect reach or monetization”, so a low regular share doesn’t throttle you. But “if viewers are returning to your channel regularly to watch more, they’re more likely to be recommended more of your videos in the future.” And on why views run below subscriber count, YouTube says viewers “are subscribed to dozens of channels and may not return for every new upload”, and recommends monthly audience, not subscribers, as the measure of your active audience.

Which traffic sources should you watch?

Traffic sources explain the other metrics, because CTR, retention and subscriber conversion all differ by where a viewer came from. The Content tab’s “How viewers find your content” report, defined on YouTube’s content performance page, splits traffic inside YouTube from external traffic: browse features (Home, Subscriptions, Watch Later), suggested videos, YouTube search, channel pages, the Shorts feed, playlists, notifications, end screens, external sites and apps, and direct or unknown. Cards on the Reach tab show the search terms, the suggesting videos and the external sites behind each.

Read sources before rates: search traffic brings motivated viewers, while Home reaches people who weren’t looking, which lowers CTR by design. When Home or Suggested traffic falls, YouTube’s performance FAQ lists the usual causes: your audience watching more of other channels or less YouTube, a viral video whose viewers didn’t return, uploading less often, or a topic losing popularity. None is fixed by a new thumbnail.

How are Shorts analytics different?

Shorts replace click-through rate with a single question: did the viewer stay or swipe? YouTube’s blog post says a Short appearing in the feed counts as a view from its first frame, and the Shorts view of the Content tab has its own key metrics card (views, likes and subscribers) and a report called “How viewers engaged”.

Its metric is Stayed to watch: “the percentage of times viewers stayed to watch past the initial seconds of a Short.” Older guides call this “viewed vs swiped away”; Stayed to watch is the current label. YouTube’s Shorts help page says the former Shorts view metric is now reported as engaged views, and that Shorts ad revenue sharing is based on engaged Shorts views. Percentage viewed matters more than minutes here, in line with YouTube’s rule for short videos, and a Shorts-heavy channel should expect a very low regular-viewer share.

How should you read the first 24 to 48 hours?

Look, but don’t act. YouTube gives you early data on purpose and then tells you not to over-read it.

The Overview tab’s Realtime report compares 60-minute and 48-hour performance for recent uploads. Impressions, views, CTR and average view duration are “available within a few hours of publishing your content”, according to YouTube’s impressions page; retention takes one to two days; and counts may be slowed, frozen or adjusted while YouTube discards low-quality playbacks.

Against that, the CTR FAQ says not to check click-through rate immediately after uploading. The performance FAQ says publish time “is not known to impact a video’s long-term performance”, and that a new title or thumbnail re-ranks a video only because viewers react differently: “don’t change what’s working.” Our guide to how the YouTube algorithm works covers what the system weighs in that window.

How do you diagnose a video with pairs of metrics?

No single metric explains a video; pairs do. Where a row rests on a YouTube statement, the source column says so; the rest is our reasoning from the definitions.

What you seeWhat it usually meansWhat to trySource
High CTR, low AVD, fewer impressions than expectedThumbnail or title promised something the video didn’t deliverMake packaging match the content; fix the first 30 secondsYouTube’s CTR FAQ calls this the sign of clickbait
Low CTR, high AVD and retentionPackaging undersells a video people like once they startTest a clearer title or thumbnail; keep the videoOur reasoning
Falling CTR, rising impressionsThe video is being shown to broader audiences, such as HomeNothing; judge AVD insteadYouTube’s CTR FAQ says Home impressions lower CTR
Low intro retention, decent CTRThe opening doesn’t match the promise, or takes too longRe-cut the first 30 seconds; match thumbnail to contentYouTube’s key moments page
More views than impressionsA lot of traffic came from outside YouTubeCheck the External card; ignore CTR for this videoYouTube’s CTR FAQ
Views far below subscriber countNormal: subscribers follow dozens of channelsWatch monthly audience and returning viewers insteadYouTube’s performance FAQ
High views, low engaged viewsMany autoplays or hover plays that stopped earlyCompare engaged views across videos; for Shorts, check stayed to watchOur reasoning from YouTube’s definitions
Strong retention for returning viewers, weak for newThe video assumes context newcomers lackAdd a short setup; test packaging aimed at new viewersOur reasoning

Where do you find each metric, and what does a problem look like?

MetricYouTube’s definitionWhere to find itWhat a problem looks like
ImpressionsThumbnail shown on YouTube for more than 1 second, at least 50% visibleContent tab; Reach tab at video levelFalling across many videos at once: the channel is being offered less
Impressions CTRHow often viewers watched after seeing an impressionReach tab; Content tab key metricsWell below your own recent videos with a similar traffic mix
ViewsCounted the moment a video starts to play (since August 24, 2026)EverywhereComparing pre- and post-August counts as if they were the same
Engaged viewsTimes viewers stayed past the initial seconds, excluding loopsEngagement tab; Advanced ModeA wide, widening gap between views and engaged views
Average view durationAverage minutes watched among those who stayed to watchEngagement tab; Content tabFalling while video length stays the same
Average percentage viewedAverage share of a video watched among those who stayedAdvanced Mode; Engagement tabLow on short videos, where YouTube says relative watch time matters most
Intro retentionPercentage still watching after 30 secondsVideo-level audience retention reportLower than your typical retention line
New, casual, regular viewers28-day monthly audience split by watch behaviourAudience tabGrowing new viewers with flat casual and regular viewers
Returning viewersViewers who had watched before and came back in the periodAudience tab; Advanced ModeShrinking while total views hold up
Traffic sourcesWhere views came from, inside and outside YouTubeContent tab: How viewers find your contentA sudden fall in one source with no change in your uploads
Stayed to watch (Shorts)Percentage of times viewers stayed past the initial seconds of a ShortContent tab, Shorts view: How viewers engagedLower than your other Shorts

What does a 15-minute weekly analytics routine look like?

Once a week is enough for most channels. Pick the same day, use the 28-day range, and write the numbers down so you see trends instead of one video.

  1. Minutes 1 to 3: Overview. Views, watch time and subscribers against the previous 28 days, plus any typical-performance note, which YouTube says explains why views are higher or lower than usual.
  2. Minutes 4 to 6: the latest upload. Impressions and CTR on Reach, engaged views and average view duration on Engagement, against your last five videos of the same format.
  3. Minutes 7 to 9: retention. Intro retention and the biggest dip, with its timestamp.
  4. Minutes 10 to 12: audience. Monthly audience and the new, casual and regular split; the Content tab’s Subscribers report, which splits new subscribers by format (videos, Shorts, live streams, posts); and the “Channels that your audience watches” card.
  5. Minutes 13 to 15: traffic. The top three sources, the search terms behind YouTube search, and which suggested videos are feeding you.

Once a month, add 10 minutes in Advanced Mode: group videos by format or series (a group holds up to 500 videos), compare groups on engaged views and average percentage viewed, and export the table. That is the cadence at which a decision about format or length is safe.

Open spiral planner on a light wooden desk, showing a blank monthly calendar grid with a black pen and a white pen resting on it
Five numbers written down on the same day each week beat a daily refresh of the dashboard.

Which analytics should you share with sponsors?

Share the numbers a brand can’t see from outside and would otherwise guess: engaged views, average view duration, audience geography and age, and the share of returning viewers. Name the metric and date range on every figure, since public view counts changed on August 24, 2026.

For long-form, 28-day and 90-day engaged views and average view duration, plus the age, gender and geography reports, answer most questions; for Shorts, add the stayed-to-watch rate. Brands that follow our guide to vetting YouTube creators will ask for these and for your traffic sources. Put the figures on a dated page of your media kit rather than in screenshots, update them monthly, and quote subscriber count only with context.

What doesn’t YouTube publish about analytics?

Plenty. There is no YouTube benchmark for average view duration, average percentage viewed, intro retention, stayed to watch or subscriber conversion. The 2% to 10% CTR statement describes half of all channels, not a target, and the FAQ adds that its suggestions “are based on aggregate success rates across YouTube” and “do not guarantee any particular outcome.” YouTube doesn’t say how many seconds make an engaged view, and notes that geography, traffic source and gender data may be limited. Your own 10 latest videos of similar length are the comparison YouTube built for you.

What to do next

Before the next upload, open the retention report on your last three videos and note where each lost its first big group of viewers; that habit does more than any benchmark. If you post on other platforms too, our guide to calculating engagement rate gives one formula for all of them. When your numbers are steady, our guide to getting YouTube sponsorships explains how to turn them into deals, and our list of YouTube growth educators collects people who teach this well.

Brands read these same numbers from the other side; our guide to measuring influencer marketing ROI shows what they do with the figures you share.

Frequently asked questions

What is a good CTR on YouTube?
YouTube’s only published benchmark, still on its Impressions and click-through-rate FAQ in October 2026, is that half of all channels and videos have an impressions CTR between 2% and 10%. It isn’t a target: CTR falls naturally when a video reaches Home and runs higher on channel-page traffic, and videos under a week old or under 100 views swing more widely. Compare CTR between your own videos over time, by traffic source.
What is a good average view duration on YouTube?
YouTube publishes no benchmark for average view duration or average percentage viewed. It does say that relative watch time matters more for short videos and absolute watch time for longer ones, and that both metrics are calculated from engaged views, so early leavers don’t drag them down. Judge a video against YouTube’s typical retention line, which compares it with your own 10 latest videos of similar length.
What is the difference between views and engaged views on YouTube?
Since August 24, 2026, YouTube counts a view the moment a video starts to play, including a thumbnail that autoplays on hover. An engaged view, the old definition, is a click to watch or watching past the initial seconds; YouTube doesn’t publish the number of seconds. CTR, average view duration and retention are still calculated on engaged views, and Partner Program earnings still rest on engaged views and engaged watch hours.
What are returning viewers and regular viewers in YouTube Analytics?
Returning viewers had watched your channel before and came back in the date range you selected. YouTube also splits your monthly audience, the unique viewers of the last 28 days, into new, casual (watched at least once a month for one to five months in the past year) and regular viewers (at least once a month for more than six months). YouTube says a regular share as low as under 1% is common, especially for newer and Shorts-heavy channels, and that these segments don’t affect reach or monetization.
What is “Stayed to watch” in YouTube Shorts analytics?
Stayed to watch is the Shorts metric in the Content tab’s “How viewers engaged” report: the percentage of times viewers stayed to watch past the initial seconds of a Short rather than swiping away. Older guides call it “viewed vs swiped away”. A Short in the feed counts as a view from its first frame, so this stands in for click-through rate. Engaged views for Shorts exclude loops, and YouTube says Shorts ad revenue sharing is based on engaged Shorts views.
Why are my YouTube views lower than my subscriber count?
YouTube’s performance FAQ says viewers are subscribed to dozens of channels on average and don’t return for every upload, and many stay subscribed to channels they no longer watch. Subscriber count isn’t a measure of active audience; YouTube recommends monthly audience, the unique viewers of the last 28 days, instead. Watch returning viewers and the new, casual and regular split rather than subscribers.

From the lists

Keep reading

Written by Joanna Finch, editor of Topfluencers. Spot a mistake? Email corrections@topfluencers.co and we’ll fix it within 48 hours.