YouTube Analytics is easiest to read in the order a viewer meets your video: impressions and click-through rate on the Reach tab show whether people chose it, watch time and average view duration on the Engagement tab show how long they stayed, the Audience tab shows who keeps coming back, and revenue comes last. One definition changed recently, according to YouTube's content performance help page: since August 24, 2026, a view counts the moment a video starts to play, in every format.
This guide is a glossary and a method. It defines the metrics YouTube Studio shows, using YouTube's own help pages, flags the numbers that commonly look wrong, and ends with a monthly review template. How those signals feed recommendations is a separate question, answered in our explainer on what YouTube says about its algorithm.
Where each report lives
YouTube's getting started page lists seven tabs. Overview summarizes the channel or a video. Content, at channel level, shows how each format performs and is filtered by chips for videos, Shorts, live and posts. Reach, at video level, shows how people found a video. Engagement shows how long they watched. Audience shows who they are. Revenue appears for Partner Program members, and Trends shows what viewers are searching for. Open a channel report from Analytics in the left menu, or a video report by pointing at the video on the Content page and choosing Analytics.
For anything beyond the summary cards, use Advanced Mode, which lets you compare periods and export data. Two cautions from the same page: some reports are missing on mobile, and, as the getting started page notes, YouTube began rolling out an updated Studio in July 2026, so your screens may not match older screenshots.
Metric glossary
Definitions are paraphrased from the YouTube Help page in each row, as checked on October 1, 2026.
| Metric | What YouTube says it measures | Where to find it | Decision it informs | Source |
|---|---|---|---|---|
| Impressions | Times your thumbnail appeared on YouTube for more than 1 second with at least 50% of it visible; not counted on external sites, end screens, notifications or the mobile website | Content tab for the channel, Reach tab for a video | Whether YouTube is putting the video in front of people at all | Impressions and watch time |
| Impressions click-through rate | How often a counted impression led to a view | Reach tab, key metrics card | Whether the title and thumbnail persuade the people who see them | Impressions and watch time |
| Views | Legitimate views, counted the moment playback starts in Shorts, long-form videos and live streams | Overview, Content and Reach tabs | Raw reach only; a view no longer implies anyone stayed | Content performance |
| Watch time (hours) | Total time viewers spent watching | Overview and Engagement tabs | How much attention a video or period earned overall | Engagement |
| Average view duration | Average minutes watched among viewers who stayed, calculated from engaged views | Engagement tab | Whether the video delivers what its packaging promised | Engagement |
| Key moments for audience retention | How well each part of a video held attention, with typical retention comparing your 10 latest videos of similar length | Engagement tab for a single video | Where to tighten, cut or restructure the next video | Engagement |
| Unique viewers | Estimated individual people who watched in the period, calculated from engaged views | Reach and Audience tabs | The real size of an audience behind a view count | Impressions and watch time |
| Monthly audience | Unique viewers over the past 28 days, recalculated daily whatever date range you pick | Audience tab | Active audience size, a better guide than subscriber count | Audience |
| New viewers | People watching your channel for the first time in the period, including private-browser viewers and anyone away for over a year | Audience tab | Whether you are reaching beyond your base | Audience |
| Casual viewers | Viewers who watched at least once a month for one to 5 months of the past year | Audience tab | The group you can turn into regulars | Audience |
| Regular viewers | Viewers who watched at least once a month for more than 6 months of the past year | Audience tab | The core audience your schedule serves | Audience |
| Unique reach and views (co-viewed) | Estimates that count everyone watching together, such as several people in front of one TV | Content tab, All chip | How much of your audience watches as a group on TV | Content performance |
| Subscribers by content type | Subscribers gained from videos, Shorts, live streams, posts and other places such as search and your channel page | Content tab, All chip | Which format turns viewers into subscribers | Content performance |
| Traffic sources | Where viewers found the video: browse features, suggested videos, search, the Shorts feed, channel pages, notifications, playlists or external sites | Reach tab, “How viewers find this video” | Which surface to plan packaging for | Video reach |
| Estimated revenue | Net revenue from all Google-sold ads and transactions in the period | Overview and Revenue tabs | The income trend, before later adjustments | Overview tab |
Two definitions deserve a second look. Many older guides talk about “returning viewers”, and the getting started page still says channel impressions are split into new and returning users, but the Audience tab now divides your monthly audience into new, casual and regular viewers. As written, the casual and regular definitions also leave out someone who watched in exactly six months of the year, so treat the split as YouTube's grouping rather than a census. Revenue per thousand views and the two kinds of CPM have their own guide: YouTube RPM vs CPM.
The August 2026 change to view counting
YouTube's content performance page says that from August 24, 2026, views are counted the moment a video starts to play across Shorts, long-form videos and live streams. The same page says the change does not affect Partner Program earnings or eligibility: earnings still rest on engaged Shorts views and engaged watch hours, and eligibility on qualified Shorts views and qualified watch hours. Our guide to YouTube's 2027 monetization changes covers the thresholds those qualified figures feed.
The practical effect is that metrics now sit on different bases. Views count every start, while average view duration and unique viewers are calculated from engaged views. So multiplying views by average view duration will not reproduce your watch time, and any month-on-month comparison that crosses August 24, 2026 mixes two definitions of a view. Mark that date in your notes before you read a trend across it.
A reading order that leads to decisions
- Reach first. Low impressions mean the video is not being shown much; normal impressions with a low click-through rate point at the title and thumbnail. Read click-through rate by traffic source, because the click-through FAQ says it naturally drops when a video reaches broad surfaces such as Home and runs higher where most impressions come from your channel page.
- Engagement second. Average view duration and the retention curve tell you whether the video kept the people it attracted. A high click-through rate with weak retention means the packaging overpromised.
- Audience third. Monthly audience and the new, casual and regular split show whether you are growing a returning audience or renting attention one video at a time.
- Revenue last. Read income once you understand the audience that produced it, so a good month is credited to the right videos and formats.
The click-through FAQ also sets three ground rules: do not judge right after upload, do not react to small swings, and do not swap several thumbnails on one video and compare the results, since traffic sources shift underneath. YouTube's built-in tests avoid that last trap, as explained in our guide to A/B testing titles and thumbnails.
Numbers that look wrong but are not
- More views than impressions: normal when much of your traffic comes from outside YouTube, because not every view starts from a counted thumbnail impression.
- Click-through rate that does not match your sums: it only uses views that came from counted impressions, so dividing all views by impressions gives a different figure.
- A key metrics card that disagrees with a report: YouTube says the card updates more often than the detailed report.
- Dashes, zeros and totals that exceed their rows: YouTube's limited data page says small rows are hidden for privacy while still counted in totals, that its thresholds are unpublished, and that widening the period beyond 30 days or removing filters can reveal more.
- Missing demographics: private and unlisted videos never show them, and, according to YouTube's age estimation page, audience age reports use the birthday viewers entered, not YouTube's age estimates.
- Revenue that moves after the fact: the revenue help page says revenue takes 2 days to appear and is adjusted twice, after 1 week and again in the middle of the following month, while finalized earnings land in AdSense for YouTube between the 7th and 12th.
- High click-through rate and retention but few impressions: the click-through FAQ says this is common for videos watched by a narrow, loyal audience, and that other channels' videos may simply be outperforming yours for the same viewers.
Monthly review template
Work through the rows in order once a month and keep the answers in one running document. The point is a decision at the end, not a report.
| Step | Report to open | Write down | Act if |
|---|---|---|---|
| Headline | Overview, with its comparison to typical performance | Views, watch time and subscribers for the month | A figure sits outside its usual range, in which case find the upload responsible |
| Format mix | Content tab, one chip at a time | Views and subscribers gained per format | One format is growing while another shrinks, and you have not chosen that shift |
| Discovery | Reach tab for your three strongest and three weakest uploads | Impressions and click-through rate by traffic source | Both are below your norm, which points at packaging rather than topic |
| Holding attention | Engagement tab, key moments for one video | The point of the steepest drop and what happens on screen there | The drop repeats across videos at the same stage, such as the intro |
| Loyalty | Audience tab | Monthly audience and the new, casual and regular split | Regular viewers fall two months running |
| Income | Revenue tab, then AdSense for YouTube | Estimated revenue by source and the finalized figure | The two differ by more than you can explain from the adjustment rules |
| Context | Your own notes | Uploads, collaborations, holidays and definition changes that month | Something breaks comparability, so flag it beside the numbers |
| Decision | This document | The one change you will make next month and how you will judge it | Always: one change, reviewed at the next session |
If the Shorts rows keep coming up weak, our guide to Shorts that are not getting views turns the Shorts metrics into a diagnosis. For timing questions, the “When your viewers are on YouTube” report is covered in our guide to the best time to upload on YouTube.
Limitations of this guide
Analytics describes what happened, not why. YouTube does not publish the thresholds behind limited data or the inner workings of its recommendations, and its own pages carry small inconsistencies, so treat every definition here as YouTube's wording on October 1, 2026 rather than a permanent rule. Studio is mid-redesign, which may move or rename reports.
Estimated revenue is not income for tax or accounting purposes until it is finalized and paid. For anything involving tax, withholding or business planning, rely on your AdSense for YouTube records and a qualified accountant.