16 Sept 2026
Module 11. Analytics: Understanding CTR, Audience Retention, and Traffic Sources
Don`t want to read? Listen recording!
A practical beginner guide to impressions, CTR, audience retention and traffic sources. The goal is not to memorise the dashboard. It is to know what to check and what to change in the next video.
1. YouTube Analytics for Beginners: What to Check First
At eight on Monday morning, Emma poured herself a coffee and opened YouTube Studio. The coffee nearly went cold.
The Analytics screen was full of cards she did not understand. Graphs, percentages, CTR 7.7%, arrows pointing up and down, 20,596 impressions. Then she opened the audience retention graph for her latest apple pie video and saw the line drop hard in the first thirty seconds. Thirty-five people out of every hundred had already left.
“That’s it. A third of them left in half a minute. My voice is annoying, the pie is boring, and I clearly have no idea what I’m doing. I should just stop,” Emma thought.
Left: Emma sees the drop and decides the video failed. Middle: we work out what the graph is actually saying. Right: instead of checking all day, she gives Analytics fifteen minutes once a week.
Paul went to the other extreme that same week. After Module 9, he published a detailed video on sharpening lawn-mower blades and opened Studio every twenty minutes. At 10 a.m. the CTR was 12%. By lunchtime it was 8%. By evening it had fallen to 6.5%. In between, Paul rewrote the title three times, swapped the thumbnail for a brighter one, and somehow found time to argue with the first sceptical commenter who showed up.
Both reactions are common when you are new, and both are exhausting. One person is scared of the numbers and works blind. The other watches every twitch in the graph and keeps fixing things that may not be broken.
Analytics is not a school report and it is not a verdict on whether you have talent. Think of it more like an oven thermometer. If a cake is burning underneath, you do not sit down and decide you are a terrible cook. You turn the heat down and keep going.
In this lesson we will focus on three signals first: impressions, CTR and retention. Then we will look at where viewers came from, which videos YouTube puts next to yours, and what to change in the next upload. For a beginner, fifteen focused minutes a week is enough.
2. How to Read YouTube Analytics Without Obsessing Over Views
The point of Analytics is simple: find the stage where the video is losing the viewer. Are people seeing it? Are they clicking it? Are they still watching once they arrive? If a number does not help answer one of those questions or lead to a decision, you can ignore it for now.
Impressions put the cake in the window. The title and thumbnail get someone through the door. The video decides whether they finish the slice.
3. YouTube Analytics Metrics Explained: Impressions, CTR, Retention & Traffic Sources
Impressions tell you how many times YouTube showed your thumbnail on eligible surfaces such as Home, Search or recommendations. YouTube counts a registered impression only when at least 50% of the thumbnail is visible for more than one second. External links from websites or messaging apps do not create registered impressions.
Impressions click-through rate (CTR) is the percentage of registered impressions that turned into views: (views from impressions / impressions) × 100. If 8 people click after 100 registered impressions, the CTR is 8%.
View is a view that YouTube has counted for the video.
Audience retention shows what percentage of viewers are still watching at each point from 0:00 to the end. This is where you see people stay, rewatch, skip or leave.
Average view duration (AVD) is the average amount of time viewers spent watching the video.
Traffic sources tell you how viewers reached the video: Browse features, YouTube Search, Suggested videos, external websites and apps, and other surfaces.
Evergreen content is a video on a subject people keep looking for months or years after it was published.
4. YouTube Studio Analytics: How to Read the Overview Tab
Open Analytics in the left menu of YouTube Studio. The first screen is the Overview tab.
The main Analytics screen: the core channel metrics in the centre, the 28-day date range at the top, and realtime views on the right.
Why YouTube Studio uses 28 days instead of a calendar month
The default range in Studio is Last 28 days. That sometimes looks odd when you first see it. Why not simply say “this month”?
Twenty-eight days is exactly four full weeks. You get the same number of Mondays, Fridays and Sundays every time, so weekdays and weekends are balanced. That makes period-to-period comparisons cleaner than comparing calendar months with different lengths.
You can still change the selector to 7 days, 90 days, 365 days, lifetime or a custom range whenever you need to.
Three channel metrics to understand first
Views. Total views across the channel during the selected period. Studio may also show whether the result is within your channel’s usual range. That comparison is more useful than comparing your channel with somebody else’s.
Watch time (hours). The total amount of time viewers spent watching your videos. This matters both as a depth signal and for monetisation progress. YouTube tracks qualified watch hours separately for YPP. As of this lesson, the ads and Premium entry route is 4,000 qualified watch hours in 365 days. For new applicants, that rises to 8,000 hours from 1 February 2027. The earlier fan-funding tier remains at 3,000 qualified watch hours.
Subscribers. The net change for the selected period: subscriptions minus unsubscribes. On a small channel, a swing of a few people is not a crisis. People clean up subscriptions, change accounts, or simply change their minds.
If you forget a term: hover over the question mark or the metric title. Studio usually gives you a short definition, and the help link goes deeper if you need it.
5. YouTube Impressions and CTR: How to Read the Reach Funnel
Open the Content tab at channel level, or open a specific video and go to Reach. This is where the funnel starts: YouTube shows the thumbnail, some viewers click, and then the video has to hold them.
Studio’s impressions card. Registered impressions count eligible thumbnail appearances on YouTube, not every possible place a link to the video exists.
A lot of beginners look only at views. If views are low, the immediate conclusion is: “YouTube is not showing my video.” Check Impressions first. That tells you whether the thumbnail is getting chances to compete at all.
What counts as a YouTube impression
A registered impression is counted when the thumbnail is at least 50% visible for more than one second on an eligible YouTube surface. External websites, messaging apps and some other surfaces are outside this metric.
YouTube CTR: what we use as a practical benchmark
Next to impressions you will see Impressions click-through rate.
For the first meaningful test of a long-form video, we use a simple working rule: once the video has enough test impressions to say anything at all, CTR below 4% means we inspect the packaging first. The most common problem is the title, the thumbnail, or the way the two work together.
This is a working benchmark from running channels, not a magic platform threshold. The point is to decide what to inspect first.
Why CTR drops when reach expands
Emma’s video started around 14% CTR and was down to 7.7% a week later. That does not mean the thumbnail suddenly became worse.
First test. The video is shown to a warmer group: people who know the channel or are very close to its existing audience.
Reach expands. If the video passes that first test, YouTube starts putting it in front of colder viewers.
CTR drops. A colder audience clicks less often. When reach genuinely expands beyond the warm core, CTR comes down.
Views still grow. A smaller percentage of tens of thousands of impressions is still far more viewers than a large percentage of a few hundred.
How to read CTR: 12% on a small warm test and 6.5% on a much wider cold audience are not the same situation. Always read CTR next to the number of impressions.
6. YouTube Video Analytics: Views, Reach and Evergreen Content
The biggest beginner mistake is trying to manage the channel from the channel-wide graph. That graph mixes everything together: strong videos, weak videos, Shorts, long-form, fresh uploads and old ones. It is useful for seeing the overall direction, but not for diagnosing one video.
Most useful decisions happen at the video level.
Go to Content → choose a video → Analytics → Reach.
This example comes from a real channel video, How They Feed You at One of the Best Health Resorts in Belarus, published on 18 January 2022, almost 1,700 days ago.
Look at the funnel:
Impressions: 1.8 million. In 81.3% of those cases, the thumbnail was shown in YouTube recommendations.
CTR: 9.8%. For reach this broad, that is a very strong result in our experience.
Views from impressions: 174.7K.
Watch time from impressions: 8.8K hours.
Evergreen YouTube content: videos that keep working for years
The video is more than four years old. Hundreds of newer uploads have appeared on the channel since then, but this one still receives impressions and views.
The reason is simple: people keep asking the same question. Every year somebody plans a trip, compares resorts, and searches for things like “what is the food like?” or “buffet or set menu?” If a video answers a question that comes back every year, it can keep collecting traffic long after publication week is over.
7. YouTube Traffic Sources Explained: Browse, Search, Suggested & External
Now look at the traffic-source side of the same video’s Reach report.
Four traffic sources, four different questions about how the video is being discovered.
Left: videos that sent viewers through Suggested. Right: the exact search phrases viewers used.
1. Browse features: Home and Subscriptions
This is traffic from places where a viewer is browsing rather than asking a specific question, especially the Home feed and subscriptions surfaces.
What matters here: Home is a crowded shelf. Your thumbnail and title have to win a click against everything else on that screen. A clear promise beats a vague one.
2. YouTube Search: the exact words viewers typed
Search traffic starts with intent. The viewer came looking for something. In this real example, queries included phrases equivalent to “Belarus health resorts”, “Yunost health resort Belarus” and “Alpha Radon resort”.
Practical use: the YouTube Search terms card gives you viewer language, not brainstorming language. If people repeatedly search for “Alpha Radon prices and food”, that is already a topic angle for a follow-up video.
3. Suggested videos: what YouTube puts next to your video
Suggested traffic comes from recommendations alongside or after another video.
What to inspect: open Content suggesting this video. In our example, the neighbouring videos were about factory cafeterias in Belarus, places to eat, and other health resorts.
Good sign: the video about food and travel is sitting next to content for the same kind of viewer.
Bad sign: on an established channel, if a home-baking video keeps appearing next to cartoons or gaming playthroughs, YouTube is matching it to the wrong interest cluster. New channels can look messy while the system is still testing, but once a channel has history, this card is worth watching.
4. External traffic: messaging apps, websites and shared links
The external breakdown in this example includes WhatsApp and Viber.
That is ordinary word of mouth. Somebody watched the video and sent the link to another person or a group chat. For a creator, that is useful evidence: the video was useful or interesting enough to share outside YouTube.
8. YouTube Audience Retention: How to Read the Retention Graph
Open the video’s Engagement tab. This is where you will find one of the most useful graphs in the whole of Studio: Audience retention.
On this successful video, about 66% of viewers are still watching at 0:30.
Four things to look for: the opening drop, a steady middle, replay spikes and sharp dips.
The first 30 seconds: the first real test of the video
Go back to Emma’s panic: “I lost 35% in the first half minute.” On a successful video with roughly 215,000 views, 66% of viewers are still there at 0:30.
Every video loses people at the start. Some clicked by accident. Some expected a different format. Some simply changed their mind. The question is not whether the graph drops. The question is how far it drops.
Our working benchmark for long-form videos:
65–75% at 0:30: strong start;
around 60%: lower working line;
below 60% in a competitive niche: the opening is already losing too much ground. Check the first seconds, the promise made by the title and thumbnail, and how long it takes to get to the point.
The first fixes are usually boring and effective: cut the long greeting, do not ask for a subscription before the viewer has received anything, and show the reason to keep watching earlier.
A steady slope: no obvious reason to leave
After the opening drop, a smooth line without sharp steps is usually a good sign. People are still leaving, because people always leave, but there is no single moment shouting “something went wrong here”.
Spikes: people went back
A spike means some viewers rewound or replayed that section. It might be a measurement, a recipe ratio, a diagram, a surprising result, or simply a moment they wanted to see twice.
Save the timestamp. It may deserve a Short or a full video of its own.
Dips: open that exact timestamp
A sharp step down in the middle is worth opening immediately. Common reasons are simple: a repeated point, a long pause, an overlong intro sting, or a technical passage that adds no value.
You do not need to guess whether the edit felt slow. The graph has already pointed you to the moment worth checking.
9. YouTube Analytics Device Types: Where Viewers Watch
Open the Audience tab and find Device type.
In this example, 47.1% of watch time comes from phones and another 31.0% from TVs.
1. Mobile phones, 47.1%: what this means for thumbnails
Almost half of the watch time in this example comes from phones.
Practical thumbnail rule: a phone preview is small. Tiny words turn into mush fast. If you use text at all, keep it short and large, just as we covered in Module 7.
2. TV, 31.0%: sound becomes more obvious
Almost a third of watch time here comes from television screens. On a TV, people often watch in a room rather than with the phone ten centimetres from their face. Sudden audio changes are much more noticeable.
Practical rule: keep speech, music and intro stings at a sensible relative level. If the host is whisper-quiet and the sting suddenly blasts the room, some viewers will simply switch away. We covered the basic audio workflow in Module 8.
10. YouTube Audience Analytics: Geography, Subtitle Languages & Translations
After devices, check geography and subtitle/CC languages. This is where you can see whether a video is finding viewers outside its original language market.
Step 1. Geography: where YouTube actually found viewers
Open Advanced mode and choose Geography. This report shows where the video was watched, not where you originally imagined the audience would live.
A video does not have to stay inside one country. If viewers in another market respond well, distribution can grow there too. Geography is where you see those markets and whether localisation is doing anything useful.
Step 2. Subtitles and CC: 6.2 million views
Advanced-mode Subtitles and CC report: 6.24 million views, split by the language viewers used for captions.
This real report contains 6,244,621 views. Views with no subtitles/CC selected account for 935,874, roughly 15%. Another 920,040 used Russian subtitles. More than 5 million views came with other added languages selected: German, English, Turkish, French, Polish and many more.
On our channels, those views came from localisation. A translated title gives a viewer a reason to click in the first place. Subtitles let that person keep watching even when the creator speaks another language. Without the localisation layer, most of that foreign traffic would either never start or leave almost immediately.
And the effect does not stop at the foreign-country total. Extra viewers bring extra watch time and extra audience signals. In our experience, that additional momentum can help the video collect more recommendations overall, including in its original market.
You do not need expensive dubbing before you test a market
A common beginner assumption is: “I talk on camera, so foreign viewers are out unless I pay for dubbing.”
Not at the start. Localise the title, description and subtitles first. If another country starts sending views and retention is healthy enough, then test extra audio tracks or auto-dubbing. Do not postpone the basic localisation because the perfect voice track is not ready yet.
The practical difference looks like this:
Without localisation: the viewer cannot understand the promise before the click, or clicks and immediately hears a language they do not know. Most of that audience disappears early.
With localisation: the viewer understands the topic before the click and can follow the video with subtitles. If the content fits, they can stay, and YouTube has a reason to keep testing the video in that market.
What the geography report shows
In this report, Germany accounts for 21.8% of views, the United States 15.0%, Poland 8.1% and Russia 4.2%.
On one of our Russian-language videos, Germany produced 98,895 views, the United States 67,787, Poland 36,691 and Russia 19,047. Germany and the US alone contributed more than 166,000 views, over a third of the total traffic in that report.
That is why we add translations. International viewers bring extra views, watch time and audience signals. On our channels, that can widen the recommendation pool around the video instead of leaving it locked to one language market.
How to avoid translating everything by hand
A few languages are manageable inside Studio. Dozens of languages across a catalogue are a different job. That is why we built batch translation for titles, descriptions and subtitles into Creator Tools, with the translations pushed back into YouTube Studio.
Creator Tools: batch localisation for video metadata and subtitles, with the translated versions sent back to YouTube Studio.
We covered the workflow itself in Module 6. Here the job is simpler: open Analytics and see whether it worked. Check geography, subtitle languages and retention from the countries you are trying to reach. If those markets begin to grow, the localisation is doing its job.
11. How to Analyze YouTube Analytics: 5 Common Scenarios
One rule before the table: do not make big decisions from tiny samples. If a video has 30 viewers, one person can move the percentages a lot. Wait until there is enough data to see a pattern, and check whether the same thing happens on other videos.
One weak video does not mean the channel is broken. The job of Analytics is not to diagnose your entire career. It is to give you the next thing to inspect.
12. How Often Should You Check YouTube Analytics?
Pick one time each week and make that your Analytics slot. Sunday evening after dinner is fine. Tuesday morning is fine. The exact day does not matter. Give it fifteen minutes, get one useful answer, then close Studio.
Step | Where to look | What to look for | Time |
1. Impressions and CTR | Reach on the latest video | Is there enough impression volume to say anything? If the first test is established and CTR is below 4%, check the title and thumbnail first. If impressions have expanded sharply and CTR has fallen, that is normal audience expansion. | 3 min |
2. First 30 seconds | Audience retention on the latest video | 65–75% is a strong start. Around 60% is our lower working line. Below 60%, inspect the opening and the promise-to-content match. | 3 min |
3. Sharp dips | Middle of the retention graph | Find the clearest drop. Open that timestamp and write down what happened there. | 4 min |
4. Search terms | YouTube Search terms | Which phrases did viewers actually type? Put the best two into the Module 9 content-plan list. | 3 min |
5. One change | Your notes | Write down exactly one thing to change in the next video. | 2 min |
That is the whole session. You leave with one note for the next upload, not twelve screenshots and a headache. There is very little a beginner can learn from refreshing the same dashboard all day.
13. YouTube Analytics Mistakes: What Not to Do
Checking every half hour. Early numbers jump around. Give the video enough time and impressions to form a pattern before you react.
Treating any opening drop as a disaster. Every video loses people early. What matters is how deep the drop goes. Around 60% at 0:30 is our lower working line; 65–75% is a strong start. Below 60% in a competitive niche is where we open the video and inspect the first seconds.
Comparing your numbers with somebody else’s channel. Their channel age, niche, geography and upload history are different. Compare similar videos inside your own channel first.
Changing five things at once. New format, new presenter, new thumbnail style, new duration, new upload day. If the result changes, you will have no idea why. Change one meaningful variable when you can.
Turning every edit into a speed run. Fast cuts do not fix a weak idea. Give viewers enough time to understand the recipe, instruction or point. Remove dead air, not meaning.
14. YouTube Analytics Audit: A Simple 5-Step Practice
Open YouTube Studio on desktop or in the app:
Go to Content and choose your latest video, or your most-viewed one if the channel is still very new.
Open Analytics → Reach:
write down Impressions;
write down CTR;
check Traffic source types and note which source is first.
Open Engagement → Audience retention:
hover over 0:30 and note the percentage still watching;
find the clearest drop in the middle, open that timestamp, and write down what is happening on screen.
Check Geography and Subtitles/CC in Advanced mode:
are there views from other countries;
does this video have translated metadata and subtitles available?
Write one rule for the next video. For example: “No greeting. Show the finished result in the first five seconds.”
15. YouTube Analytics Checklist for Beginners
You can find impressions, CTR and audience retention for a specific video without wandering around Studio.
You understand the difference between a view and a registered impression.
A normal opening drop on the retention graph no longer makes you want to delete the video.
You can read the Subtitles/CC report and understand that “No subtitles/CC” means the viewer watched without captions enabled. It is not a language or a country.
You understand why translated titles and subtitles can bring in viewers from other countries and stop language alone from killing retention.
You have one fifteen-minute analytics slot in the week instead of checking Studio every hour.
You can score at least 3 out of 4 on the self-check below.
16. YouTube Analytics: What to Remember
Read the funnel in order: impressions, CTR, retention. It is the quickest way to see whether the problem is reach, packaging or the video itself.
Weak CTR on the first meaningful test means we inspect the packaging first. When distribution expands to a wider cold audience, CTR comes down. That is part of reach growth.
At 0:30, we want roughly 60% or better. 65–75% is a strong opening. Below 60%, inspect the first seconds and whether the video delivers what the title and thumbnail promised.
Suggested tells you what YouTube thinks belongs next to your video. On an established channel, we want to see content for the same type of viewer.
Translations and subtitles open extra audiences. In our real case, more than 5 million views were watched with added subtitle languages selected.
Do not live in Studio. Fifteen focused minutes, one decision, then back to the next upload.
22 Min to read
.jpg)
.png)