← Back to Blog

YouTube Analytics: The Studio Metrics That Actually Matter

Four things drive growth on YouTube: whether people click when your video is offered (impressions and click-through rate), whether they keep watching (retention), whether they come back (returning viewers), and where YouTube chooses to show you (traffic sources). Almost everything else in YouTube Studio is either a breakdown of those four or a distraction from them. This guide explains how to read each one in context and what to change when one is weak.

Key takeaways

  • CTR falls as impressions widen. A lower CTR on a video with far more impressions is usually a success, not a problem.
  • Use average view duration for minutes and average percentage viewed to compare videos of different lengths.
  • The retention graph has three places to look: the first 30 seconds, the dips and the spikes.
  • Returning viewers are the best indicator of whether the channel, not just one video, is working.
  • Traffic sources tell you which job a video is doing: answering searches, riding suggestions or winning the home page.
  • For income, watch RPM. CPM is the advertisers’ number.

Impressions and click-through rate, in context

An impression is counted when your thumbnail is shown to a viewer on YouTube, for example on the home page, in search results or beside another video. Impressions click-through rate is the percentage of those impressions that became a view.

Two details change how you read these numbers. First, not every view comes from a counted impression. Views from external sites, direct links and end screens arrive without one, so views will not equal impressions multiplied by CTR. Second, CTR depends heavily on who is being shown the thumbnail.

Why CTR drops when a video does well

YouTube shows a new video first to the people most likely to want it: subscribers and recent viewers. They click at a high rate. If the video performs, it is offered to wider audiences who know you less, and they click at a lower rate. So CTR declines as impressions grow. That is the mechanism working as designed.

Here is a hypothetical example. In its first two days a video gets 40,000 impressions at a 6.5% CTR, which is 2,600 views. Over the next four weeks YouTube widens it to 400,000 impressions and CTR settles at 3.4%, which is 13,600 views. The CTR “fell by almost half” and the video earned more than five times the views. If you had swapped the thumbnail in a panic on day five, you would have been fixing a healthy video.

So judge CTR this way:

  • Compare videos at a similar impression volume, or at the same age, such as the first 48 hours.
  • Compare within a traffic source. Search CTR and Browse CTR behave differently.
  • Read CTR together with impressions. High CTR with tiny impressions means only your core audience has seen it.

When a video has plenty of impressions and a CTR clearly under your norm for that volume, the packaging is the problem. Test a new thumbnail or title, change one thing at a time, and use the A/B thumbnail test calculator to check that the difference is larger than chance before declaring a winner.

Average view duration vs average percentage viewed

Average view duration (AVD) is total watch time divided by views, in minutes and seconds. Average percentage viewed (APV) is how much of the video the average viewer watched.

They answer different questions. Suppose, hypothetically, a 10-minute video has an AVD of 4:12. That is 42% viewed. A 22-minute video has an AVD of 6:36, which is 30% viewed. The longer video holds a smaller share of its length but contributes about 57% more watch time per view.

Neither is “better” in general. Use APV to judge how tightly a video was made and to compare videos of different lengths. Use AVD to understand how much viewing time each click delivers.

Watch time, the total of all minutes viewed, is the product of the two halves of your funnel: views multiplied by AVD. It is also the figure that counts toward monetization, which you can plan with the watch time calculator.

Reading the audience retention graph

The retention graph on each video shows the percentage of viewers still watching at every moment. It is the most useful single chart in Studio because it points at exact seconds.

The intro drop

Every video loses viewers in the first 30 seconds. The question is how many. A steep early fall means the opening did not confirm what the title and thumbnail promised, or took too long to start. Cut channel intros, long greetings and “before we begin” housekeeping. Show or state the payoff first, then build to it.

Dips

A dip is a point where the line falls faster than the slope around it. People are leaving or skipping ahead. Open the video at that timestamp. You will usually find a tangent, a sponsor read with no transition, a repeated explanation or a slow stretch with nothing new on screen.

Spikes

A spike is a point where the line rises. People are rewatching that moment or jumping to it. That is your most valuable section. It tells you what the audience came for, and it often deserves its own video or its own Short.

Returning viewers

The Audience tab separates new viewers from returning viewers, and depending on your Studio version may break the audience into finer groups by how regularly they watch. However it is labeled, the question is the same: are people coming back?

A channel with strong view counts and a flat returning-viewer line is collecting one-time visitors. This is typical for channels driven by search or by a single breakout video. It is fragile, because each video has to find its audience from scratch.

A rising returning-viewer line means each upload starts with a warm audience, which improves early CTR and retention, which earns wider distribution. To build it, make the next video an obvious follow-on from the last: series, recurring formats, playlists and end screens that point to one specific video rather than “more videos”.

Subscribers are a promise to watch. Returning viewers are the proof. If the two disagree, believe returning viewers.

Traffic sources and what each means

Source What it means What to do with it
Browse features Home page and subscription feed. YouTube is recommending you to people based on their history. Depends on packaging and broad appeal. Protect CTR and early retention.
Suggested videos Shown beside or after other videos, yours or other channels’. Check which videos send the traffic, then make content that naturally follows them.
YouTube search Viewers typed a query and chose your video. Steady and long-lived. Review the search terms list and cover neighboring queries.
Shorts feed Viewers swiping through Shorts. Judge by Shorts metrics, not by long-form standards.
External Websites, apps and social platforms linking to the video. Shows where your audience gathers off YouTube. Views arrive without impressions.

Knowing your mix tells you which lever to pull. For search-led videos, titles and descriptions carry more weight, and the YouTube video SEO tool helps you check them before publishing.

Shorts vs long-form metrics

Shorts and long-form videos share a channel but not a scoreboard. Do not average them together.

Since 31 March 2025, a Shorts view is counted every time a Short starts to play or replay, with no minimum watch time. That made view counts jump for many channels without any change in real audience. The previous-style count is now labeled engaged views, and that is the figure used for YouTube Partner Program eligibility and for earnings.

For Shorts, the key numbers are:

  • Viewed vs swiped away: the share of people who chose to watch instead of swiping on. This is the Shorts equivalent of CTR.
  • Average percentage viewed: on a Short this can exceed 100% when viewers loop it.
  • Engaged views: the number to compare over time and against thresholds.

For reference, the full Partner Program tier requires 1,000 subscribers plus either 4,000 public watch hours in the past 12 months or 10 million public Shorts views in the past 90 days. The earlier tier requires 500 subscribers, 3 public uploads in the past 90 days, and either 3,000 public watch hours in 12 months or 3 million public Shorts views in 90 days. The YPP eligibility checker shows how far you are from each.

RPM vs CPM for monetized channels

CPM is the cost advertisers pay per 1,000 ad impressions, before YouTube’s share. RPM is your revenue per 1,000 views, after YouTube’s share, counting all views including those that showed no ad, and including non-ad income such as Premium and memberships.

A hypothetical month: your videos get 120,000 views. Ads are shown 80,000 times at a CPM of $9, so advertisers pay $720. Your 55% share of long-form ad revenue is $396. Add $24 from Premium viewers and memberships and your total is $420. RPM is $420 ÷ 120,000 × 1,000 = $3.50.

The CPM says $9. Your bank account says $3.50 per 1,000 views. Plan with RPM. It moves with topic, viewer country, season and video length, so compare it against your own previous months, not against other creators. You can model scenarios with the YouTube RPM calculator.

Diagnosis table: symptom, likely cause, fix

Symptom Likely cause Fix
Low impressions, normal CTR and retention Topic has a small audience, or the video has not been tested widely yet Give it time. If it stays low, choose topics with broader demand.
High impressions, low CTR for that volume Title and thumbnail do not communicate the value Test one new thumbnail or title. Make the subject readable at small size.
High CTR, steep intro drop The opening does not match the promise of the packaging Rework the first 30 seconds to confirm the promise immediately. Tone down packaging that overpromises.
Good intro, steady decline Pacing: too much setup, repetition, no reason to stay Tighten the edit. Tell viewers early what is coming later.
Strong views, few returning viewers One-off topics with no follow-on Build series and point end screens to a specific next video.
Views fine, RPM falling Seasonal ad demand, audience country mix, or a larger share of Shorts views Compare with the same month last year and check the geography report before changing content.
Search traffic only, flat growth Videos answer queries but do not earn recommendations Improve retention and packaging so videos can win Browse and Suggested.

Metrics you can safely ignore

  • Daily subscriber changes. They lag behind views and include routine unsubscribes. Review the monthly trend.
  • The real-time counter. Hour-by-hour views tell you nothing you can act on.
  • Dislikes on a video with healthy retention. Retention shows what viewers did. Read comments for specifics instead.
  • CTR compared across videos with very different reach. As shown above, it is not a like-for-like comparison.

Put it into practice

Once a week, with videos that have been live at least seven days:

  1. Note impressions and CTR for each new video at the 48-hour and 7-day marks, next to your usual figures at those ages.
  2. Record AVD and APV, and compare APV with videos of similar length.
  3. Open the retention graph. Write down the intro drop, the timestamp of the biggest dip and the timestamp of any spike.
  4. Check the top three traffic sources and whether the mix changed.
  5. Look at the returning viewers trend for the last 28 days.
  6. For Shorts, log engaged views and viewed vs swiped away separately.
  7. If monetized, note RPM for the last 28 days against the previous 28.
  8. Pick one packaging change and one editing change for the next video.

You can get a quick outside view of your channel with the free YouTube channel audit. For ongoing tracking, YouTube analytics in CreatorsMetrics syncs watch time, retention and traffic sources through Google’s official read-only sign-in. Impressions CTR is not available through YouTube’s API, so keep reading that one in Studio. If you also publish elsewhere, the same weekly approach is applied to short-form in TikTok analytics explained.

Turn your watch time, retention and traffic source data into a scored audit and a 30-day plan. CreatorsMetrics connects to your channel with official read-only sign-in and never posts or asks for your password.

Create a free account

Frequently asked questions

What is a good click-through rate on YouTube?

There is no fixed target, because CTR falls as a video is shown to wider, colder audiences. Compare a video’s CTR with your own videos at a similar impression volume and from the same traffic source.

Is average view duration or average percentage viewed more important?

You need both. Average view duration shows the minutes each view contributes, and average percentage viewed shows how well the video held attention for its length, which makes videos of different lengths comparable.

What is the difference between RPM and CPM?

CPM is what advertisers pay per 1,000 ad impressions before YouTube takes its share. RPM is what you earn per 1,000 total views after the share, including views that showed no ads, so RPM is always the lower and more useful number.

Why did my Shorts views jump in 2025?

Since 31 March 2025, Shorts views count every play and replay with no minimum watch time. The previous-style count is now called engaged views, and that is the figure used for Partner Program eligibility and earnings.

Which YouTube metrics can I ignore?

Subscriber count swings, dislikes on an otherwise healthy video, the real-time counter, and CTR comparisons between videos with very different reach. None of them tells you what to change in the next video.