YouTube Analytics is the free dashboard inside YouTube Studio that shows how people find, watch and come back to your videos. For brands, the short answer is this: ignore the vanity numbers and read five things in order. Reach (impressions and click through rate), traffic sources, retention, audience loyalty and the screens people watch on. That sequence tells you whether a problem is your packaging, your content or your distribution.
This guide walks through YouTube Analytics step by step, explains what each metric really measures according to YouTube's own Help Center, flags the big change to how views are counted in 2026, and ends with a monthly report template your team can copy.
The short version
- Views changed meaning in 2026. Since 24 August 2026 a view counts the moment playback starts on every format. Use engaged views and watch time for trend lines.
- There is no single good CTR. YouTube says half of channels sit between 2% and 10%, and CTR drops naturally when Home serves a video widely.
- Retention beats raw views. Average view duration and the retention graph show whether the content earned the click.
- Loyalty is the brand metric. Returning, casual and regular viewers show whether you are building an audience or renting one.
- Report monthly, not daily. Data can lag by days and retention takes 1 to 2 days to process, so weekly panic is mostly noise.
What is YouTube Analytics, and what changed in 2026?
YouTube Analytics lives in YouTube Studio and is split into tabs: Overview for a high level summary, Content, Reach ("how your audience is discovering your channel"), Engagement ("how long your audience is watching your videos"), Audience ("who is watching your videos"), Revenue for monetised channels, and Trends for what people are searching for, according to YouTube Help. A realtime report lets you compare the last 60 minutes and 48 hours with your typical performance (YouTube Help).
At Made on YouTube in September 2026, YouTube announced new Insights and Research destinations in Studio to help creators "understand past video performance and discover outlier videos" (YouTube Official Blog). We covered the wider announcements in YouTube New Features 2026; here we focus on the metrics themselves.
The view count is not the number it used to be
From 31 March 2025, Shorts views began counting every start or replay with no minimum watch time, and the old method was renamed engaged views (YouTube Help). Then, from 24 August 2026, "views are counted the moment a video starts to play across all formats, including Shorts, long form videos (VOD), and live streams." YouTube Partner Program earnings stay "based on engaged views and engaged watch hours" (YouTube Help).
The practical consequence: a year on year view comparison across August 2026 is not like for like. If your board deck shows views jumping in September, check engaged views and watch time before you celebrate.
How to read YouTube Analytics: 6 steps for brands
Work through these in order. Each step answers one question, and the answer tells you where to look next.
Step 1: Set the date range and compare to typical
Pick the last 28 days for routine reviews and use "Since published" for a single video. Overview cards compare your results with "your typical performance", which is a fairer yardstick than your best ever video (YouTube Help). The YouTube Analytics API documentation says each day runs from 12:00am to 11:59pm Pacific time, so UK teams should expect a day to close in the morning UK time (Google for Developers).
Step 2: Check reach with impressions and click through rate
An impression is logged when your thumbnail is "shown for more than 1 second and at least 50% of the thumbnail is visible on the screen." Impressions on third party websites and embeds, mobile web, YouTube Kids, cards and end screens, and email or push notifications are not counted (YouTube Help).
Impressions click through rate is how often viewers watched after seeing one of those impressions. YouTube says "half of all channels and videos on YouTube have an impressions CTR that can range between 2% and 10%," and that "if a video gets a lot of impressions (such as if it appears on the Home Page), it's natural for the CTR to be lower" (YouTube Help). Falling CTR alongside rising impressions is often good news. For packaging fixes, see our thumbnail best practices and title guide.
"You can tell if your thumbnail is clickbait if it's getting high CTR but low average view duration and lower than expected impressions." YouTube Help
Step 3: Read your traffic sources
Traffic sources show which surfaces deliver viewers. YouTube defines browse features as "traffic from the Home, subscriptions, Watch Later, Trending/Explore, and other browsing features"; suggested videos as suggestions "next to or after other videos, and from links in video descriptions"; and direct or unknown as "direct URL entry, bookmarks, signed out viewers, and unidentified apps" (YouTube Help). YouTube says recommendations drive more overall viewership "than channel subscriptions or search" (YouTube Official Blog).
For brands this split is diagnostic. Strong YouTube search traffic means your videos answer real queries (see our YouTube SEO guide). Heavy external traffic with weak browse traffic means you are pushing views in from paid or owned channels but YouTube is not yet recommending the content on its own.
Step 4: Judge the content with retention
Average view duration is "average minutes watched among those who stayed to watch", calculated from engaged views. The retention graph adds key moments: the intro (the share still watching after 30 seconds), top moments, spikes where people "rewatched or shared", and dips where they "skipped or stopped watching". Typical retention compares a video with your 10 latest videos of similar length, and the report needs a video of at least 60 seconds with at least 100 views. Data "typically takes 1 to 2 days to process" (YouTube Help).
Benchmarks vary widely and the public ones come from tool vendors, so treat them as rough guides. Retention Rabbit's study of over 10,000 videos found average retention of 23.7%, with 55% of viewers gone by the 60 second mark (Retention Rabbit, vendor research). vidIQ publishes target bands that fall as videos get longer:
Target average percentage viewed by video length
Step 5: Measure loyalty, not just reach
Unique viewers is "the estimated number of viewers who came to watch your videos over a given time period", counting one person once across devices and repeat viewing, with totals available for periods up to 90 days (YouTube Help). YouTube then splits your audience into new viewers, casual viewers (watching at least monthly for one to five months of the past year) and regular viewers (at least monthly for more than six months) (YouTube Help).
This is the closest thing YouTube offers to a brand health metric. A channel growing regular viewers is building a habit; a channel living on new viewers is paying for attention again every month. It also explains why posting rhythm matters, which we cover in how often to post on YouTube.
Step 6: Check which screens people watch on
The device type report matters more every quarter. Viewers watch over one billion hours of YouTube on TVs every day (YouTube Official Blog), and over half of YouTube's top 100 creators now have TV as their most watched interface (YouTube Official Blog). In the US, Nielsen measured YouTube at a record 14.2% of all TV viewing in July 2026.
YouTube share of total US TV viewing, 2026
If TV is a large share of your watch time, judge creative on a big screen: legible graphics, longer runtimes and fewer phone style captions.
YouTube Analytics metrics cheat sheet
| Metric | What it tells a brand | Where to find it | Watch out for |
|---|---|---|---|
| Impressions | How often YouTube chose to show your thumbnail | Reach tab | Excludes embeds, end screens and notifications |
| Impressions click through rate | How persuasive your title and thumbnail are | Reach tab | Naturally lower when Home sends lots of impressions |
| Views | How often playback started | Overview | Counted from the first frame since 24 August 2026 |
| Engaged views | Views that meet YouTube's engagement bar | Overview and Content | The basis for YPP revenue and eligibility |
| Average view duration | How long viewers actually stayed | Engagement tab | Calculated from engaged views |
| Audience retention | Exactly where people drop off or rewatch | Video analytics | Needs 60 seconds of video and 100 views |
| Unique viewers | How many people you really reached | Audience tab | Totals only for periods up to 90 days |
| New, casual and regular viewers | Whether you are building a habit | Audience tab | Regular means monthly viewing for 6+ months |
| Traffic sources | Which surfaces deliver your audience | Reach tab | Direct or unknown includes signed out viewers |
Definitions summarised from YouTube Help pages listed in Sources.
Template: a monthly YouTube Analytics report for brands
Advanced mode lets you analyse specific videos, compare periods or groups of videos and export data. Downloaded reports are limited to 500 rows, and groups can hold up to 500 videos, playlists or channels (YouTube Help). Larger teams can pull the same data into dashboards with the YouTube Analytics and Reporting APIs (Google for Developers). This is the one page structure we recommend (Advertain guidance):
- Headline: engaged views, watch time and subscribers for the month against the previous three months.
- Reach: impressions and click through rate, split by traffic source.
- Content: top five videos by watch time, each with average view duration and its intro retention.
- Loyalty: unique viewers and the share of regular and casual viewers.
- Screens: watch time by device type, with TV called out.
- Experiments: results from Test and Compare. YouTube picks winners on watch time share, not clicks, and tests run from a few days up to 2 weeks (YouTube Help).
- Decisions: three actions for next month, each tied to one metric above.
If you want a team to build and read this for you, see our work or talk to us.
Frequently asked questions
How often does YouTube Analytics update?
The realtime report covers the last 60 minutes and 48 hours. Impressions data appears within a few hours of publishing, retention typically takes 1 to 2 days to process, and YouTube says some data can be delayed by a few days (YouTube Help).
Is YouTube Analytics accurate?
It is the most accurate source available for your own channel, but numbers move after the fact. YouTube says metrics are "algorithmically confirmed" and it may "temporarily slow down, freeze, or change" counts and discard low quality playbacks (YouTube Help). That is why Studio totals and public counts can differ for a while.
What is impressions in YouTube Analytics?
An impression is a thumbnail shown on YouTube for more than one second with at least half of it visible. Embeds, end screens and notifications are excluded.
What does browse features mean in YouTube Analytics?
Browse features is traffic from the Home feed, subscriptions, Watch Later, Trending or Explore and similar browsing surfaces. Growth here usually means YouTube is recommending your videos to people who did not search for them.
What is a good YouTube Analytics benchmark for brands?
Compare against your own typical performance first. YouTube's CTR range is 2% to 10% for half of channels, and vendor retention studies put averages around 23.7%, but format, length and traffic source change what good looks like.
