How YouTube Analytics is organised
At the time of writing YouTube Studio groups analytics into an Overview tab, a Content tab, an Audience tab and, for channels in the Partner Programme, a Revenue tab; eligible channels also see a Research tab that reports what viewers search for. YouTube rearranges and renames these from time to time, so treat the help centre listed in the sources as the record of the current layout rather than this page.
Overview is the summary: views, watch timeTotal watch time of a video divided by its number of views., subscribers, a realtime panel and the top videos for the period. Content is where the work happens. It holds the per-video and per-format reports — how viewers found the videos, impressions and how they lead to watch time, key moments for audience retention, and the split between Shorts, long-form videos and live streams. Audience describes the people: returning and new viewers, unique viewers, when they are on YouTube, the other channels and videos they watch, and demographics and geography at the level of aggregation YouTube permits. Revenue reports estimated earnings by source for monetised channels.
Underneath every tab sits Advanced mode, which opens the same data as a table that can be filtered, compared across videos or periods, grouped by dimension and exported. Most of the reading described in this section happens there, because the default charts are summaries and the decisions live in the comparisons.
Metrics by goal
YouTube reports many metrics and a channel needs few. The right few depend on what the channel is for, which is the first thing a measurement plan writes down. The table maps common goals to the metric that should lead the review and the metrics that explain it.
| Channel goal | Lead metric (the KPIA metric tied directly to an objective.) | Explaining metrics | Where to read them |
|---|---|---|---|
| Audience growth | Views from new viewers; subscribers gained | Impressions, impressions click-through rate (CTRShare of impressions that result in a click.), traffic sources | Content, Audience |
| Depth and loyalty | Watch time; returning viewers | Average view duration, average percentage viewed, the retention curve | Content, Audience |
| Discovery through search | Views from YouTube search and from external search | Search terms, impressions by source, titles and chapters | Content (traffic sources), Research |
| Traffic to a site | Clicks on description and end-screen links; sessions in web analytics | Views, end-screen element clicks, card clicks | Content, plus web analytics with tagged links |
| Support for paid campaigns | Views from YouTube advertising; cost per view (CPVCost of each video view.) in Google Ads | Retention of ad-driven views, subscribers gained from ads | Content (traffic sources), Google Ads |
| Revenue | Estimated revenue; revenue per thousand views | Monetised playbacks, ad types, memberships and other sources | Revenue |
Two warnings. Unique viewers is YouTube's approximation of ReachUnique people who saw an ad. and is reported as an estimate over a period rather than a precise count, so use it for direction. And paid metrics such as cost per view are reported in Google Ads, not in YouTube Analytics, where ad-driven views appear only as a traffic source; the YouTube Ads hub covers that side. A KPI tree is the cleanest way to connect the lead metric to the explaining metrics, so that the weekly review has a shape and a reviewer knows which number to open when the lead metric moves.
Retention and traffic sources in brief
Two reports do most of the diagnostic work, and each has its own guide.
The audience retention report draws a curve of how many viewers are still watching at each moment of a video. YouTube shows it in absolute form, as the share of starts still present, and in relative form, which compares the video with others of similar length across the platform. It marks key moments — the intro, spikes where people rewatch or share, dips where they skip or leave — and it can be filtered by traffic source and by new against returning viewers. The curve is the video's own verdict on its script and edit, and the guide to reading retention curves covers the intro drop, dips and plateaus, relative retention, Shorts, and what to change.
The traffic sources report says where views came from: YouTube search, suggested videos, browse features such as the home feed and subscriptions, the Shorts feed, channel pages, playlists, notifications, external sites and apps, YouTube advertising, and a direct-or-unknown bucket. Each source responds to different work — search to titles and descriptions, suggested to thumbnails and session behaviour, browse to the relationship with subscribers — so the mix tells a channel which lever to pull. The traffic sources guide explains each source, how to read the mix per video and per channel, and how to track external traffic properly. The optimisation work those readings call for lives in the video SEO hub.
Further guides
Further guides read the people behind the views and what the channel does for the business:
- Reading YouTube audience reports — returning and new viewers, unique viewers, subscribers, when viewers are on and what else they watch, each read against the channel's own baseline and beside the content reports.
- Measuring YouTube's contribution to the business — showing what an organic channel does for the business through tagged visits and key events, untagged influence, self-reported answers and branded search, reported as a range with its limits.
A weekly review routine
Analytics only changes anything if someone reads it on a schedule and leaves with a decision. A weekly review of under an hour covers most channels.
- Overview, this period against the previous one. Views, watch time, subscribers. Note what moved and whether a single video or the whole catalogue moved it.
- New videos in their first window. For each video published since the last review: impressions, click-through, average view duration and the traffic source mix, compared with the channel's recent videos at the same age. Decide whether the packaging or the content needs attention, and record the decision.
- Retention on the two most important videos. Open the curve, find the intro drop and the largest dip, and write one sentence on what caused each. Those sentences go to whoever scripts and edits the next video.
- Traffic sources at channel level. Has the mix shifted? A fall in suggested or browse usually precedes a fall in views; a rise in search on an older video is a prompt to refresh its metadata.
- Audience. Returning against new viewers, and the other channels the audience watches, which is the cheapest audience research on the platform. The audience reports guide covers how to read each one.
- One action. Each review ends with one change to make and one hypothesis to check the following week.
A weekly review on a personal finance channel
A personal finance channel with 41,000 subscribers reviews on Monday mornings. Views for the week are 212,000, down 8% on the previous week; watch time is flat. The Content tab shows that a video published on Thursday earned 310,000 impressions at 3.1% click-through, below the channel's recent range of 4% to 5%, while its average view duration of 5:40 on a 12-minute runtime is in line with the rest. The traffic source mix for the channel shows browse features down from 38% to 31% of views. The reviewer reads this as a packaging problem on one video rather than a content problem, and logs two actions: run Test and compare on the Thursday thumbnail, and check next Monday whether browse recovers once the new video's click-through improves. Retention on the two biggest videos is unchanged, so nothing goes to the editor this week.
Common analytics problems
Most questions a channel brings to its analytics are one of a few patterns. The table maps each to the guide that reads it properly.
| Problem | What it usually means | Where to read next |
|---|---|---|
| Views fell and nothing changed | Impressions fell on a surface, usually browse or suggested | Traffic sources |
| High click-through, short views | The packaging promises what the video does not deliver | Retention curves, Thumbnails and CTR |
| Good retention, few impressions | The video is not being offered; packaging or topic signals are weak | Thumbnails and CTR, Titles, descriptions and chapters |
| A sharp dip in the middle of every video | A recurring segment viewers skip: a sponsor read, a recap, a tangent | Retention curves |
| Search views without subscribers | The video answers a query and ends; nothing invites a return | Channel strategy |
| Site traffic does not match link clicks | Untagged links, app traffic landing as direct, or counting differences | UTM and naming conventions, Why platforms and analytics disagree |
| Shorts views without long-form growth | The two formats reach different viewers and the bridge between them is missing | Shorts versus long-form |
| Nobody can say whether the channel keeps the people it reaches | Content reports read without the audience reports beside them | Reading YouTube audience reports |
| The business asks what the channel is worth and last click shows little | Most viewers never click a link, and many who do arrive untagged | Measuring YouTube's contribution to the business |
When one of these patterns has already appeared, the YouTube channel troubleshooting pages diagnose it cause by cause: the diagnosis for views that dropped suddenly reads the drop surface by surface, and the diagnosis for retention that drops early reads the opening of the curve by traffic source.
The YouTube marketing pillar sets out how analytics fits with channel strategy and video SEO; this section teaches the reading. For the paid side, where cost per view and conversions are measured in Google Ads rather than here, the YouTube Ads hub covers formats, bidding and counting rules.
