Open YouTube Studio and you face dozens of numbers: impressions, click-through rate, views, unique viewers, watch time, average view duration, subscribers, traffic sources, returning viewers, and a dozen more. Most creators either ignore all of them or stare at the wrong one — usually views — and wonder why nothing improves. The truth is simpler: only seven metrics decide whether a channel grows, and each one points at a different fix. Learn what they mean, what good looks like, and the weekly routine that turns them into decisions.
The 7 metrics that matter
1. Click-through rate (CTR). Of every 100 people shown your thumbnail, how many clicked. Most small channels see roughly 2–10%; below 2% the packaging is the problem, not the video. Fix the title and thumbnail, not the topic.
2. Average view duration. How many minutes the typical viewer actually watches. This is the raw retention number, and it is the one that most directly feeds the 4,000 watch hours. Longer is not automatically better — a 4-minute average on a 6-minute video beats a 6-minute average on a 20-minute one.
3. Average percentage viewed. The same retention expressed as a share of the video. Roughly 40–60% is strong for most formats; under 30% means viewers leave early, and the fix is almost always the first 30 seconds, not the middle.
4. Watch time. Total hours watched. This is the number the recommendation system rewards most, because it is the closest proxy for "viewers were glad they clicked". Two videos with the same views can have wildly different watch time — the one with more is the one YouTube pushes.
5. Unique viewers. How many different people watched, as opposed to how many times videos were played. When unique viewers grow faster than views, the channel is reaching new people; when views grow faster, the same audience is watching more — both are good, but they mean different things.
6. Subscribers gained per video. Not the total count — the per-video figure. It tells you which videos actually convert a viewer into a subscriber, which is the only honest measure of whether a topic builds the channel. A video with 50,000 views and 20 new subscribers was entertainment; a video with 5,000 views and 200 was recruitment.
7. Traffic sources. Where viewers came from: search, suggested, browse, external, or the Shorts feed. A channel that lives on search grows slowly and steadily; one that lives on browse and suggested rides the algorithm's mood. Knowing the mix tells you which lever to pull — search means better titles and descriptions, suggested means better retention and related topics.
Read them in this order: the 3-minute weekly check
Once a week, open Analytics and read in this order, not the order Studio shows them. First, traffic sources for the last 28 days — it tells you where the channel actually lives. Second, CTR and average percentage viewed for your last five videos — the two numbers that diagnose packaging versus content. Third, watch time and subscribers gained per video — the two numbers that say whether the channel is compounding. Ignore everything else that week. The point of the routine is that it is short enough to do every week: a dashboard you never open is worse than a small one you actually read. Track the numbers over days rather than trusting memory with our Channel Stats Tracker, and run the full diagnosis monthly with the Channel Audit.
A worked example: two videos, one winner
Take two videos from the same small channel. Video A: 12% CTR, 32% average percentage viewed, 8,000 views. Video B: 5% CTR, 68% average percentage viewed, 8,000 views. Same views, opposite problems. Video A wins the click and loses the viewer — the thumbnail and title overpromised, or the opening 30 seconds failed, and the fix is in the hook, not the packaging. Video B loses the click and keeps the viewer — the video is good and almost nobody finds it, so the fix is a better title and thumbnail, not a re-edit. Compare the pair side by side with our Video Performance Comparison tool and the diagnosis writes itself: fix the weak number, not the video. The most common error is fixing both at once and learning nothing — change the thumbnail on Video B, wait a week, and read the CTR again before touching anything else.
Mistakes that quietly hurt
Four habits waste the data. First, checking daily — daily numbers are noise, and reacting to them means changing strategy on a coin flip; 28-day windows are the honest unit. Second, judging one video — a single video's numbers say nothing about the channel; judge the last five. Third, chasing views instead of retention — a million views at 20% retention teaches YouTube your channel wastes clicks, while 10,000 views at 60% retention teaches it the opposite. Fourth, ignoring traffic sources — a channel whose views all come from one Shorts spike looks healthy for a week and then collapses, because Shorts viewers almost never become long-form subscribers. Read the seven numbers, in order, weekly, and let them tell you which single thing to fix next.
Frequently asked questions
How often should I check Analytics?
Weekly, in the 3-minute routine above, plus a deeper monthly look. Daily checks change nothing except your mood — the numbers need a week to mean anything.
What is a good CTR?
For most small channels, roughly 2–10% depending on niche and format; above 10% is excellent. Below 2% consistently, the title and thumbnail are the problem — but judge the average across videos, not one outlier.
Why did my views drop after a spike?
Because a spike is usually one traffic source — a Short, a browse push, a search trend — returning to normal. Compare the 28 days after the spike with the 28 days before it, not with the spike week itself; that is the real trend.