Most stalled channels have one dominant problem, not ten. An audit exists to find that one problem, prove it with the channel’s own numbers, and name the single fix to test first.
A YouTube channel audit checks six things, in this order: whether the channel’s recent videos beat its own median, which videos did and what they share, whether the packaging gets clicks (impressions to click-through rate), whether the opening keeps the people who click (the first thirty seconds of the retention curve), whether the topics have an audience at all (demand in the niche), and whether the channel turns viewers into anything (subscribers, leads or sales).
The order matters because each check rules out a cause. A channel with strong click-through and strong retention that still is not growing has a topic problem, not a thumbnail problem, and no amount of thumbnail work will fix it. Run the checks in sequence, stop at the first one that fails, and fix that.
If it is your own channel, open YouTube Studio → Analytics. You have everything: impressions, click-through rate, average view duration, the retention curve, traffic sources and revenue.
If you are auditing someone else’s channel, you have the public layer only: views, likes, comments, publish dates, durations and subscriber count. Click-through rate, retention and revenue are private to the owner, and any tool showing them for a rival is showing an estimate. An outside audit can still find the dominant problem in most cases, because checks one, two and five use public data. It cannot separate a click problem from a retention problem without Studio access, and an honest audit says so.
Take the view counts of the last 15 to 20 long-form uploads. Sort them and find the median, the middle value. Now look at the most recent five against that median.
Use the median, never the average, and never subscriber count. One viral video distorts an average so badly that normal videos look like failures, and subscriber count is a historical total that says nothing about who still watches. Keep Shorts out of the long-form median; they are distributed separately and get very different view counts. We cover the method in full in how to find outlier videos.
List every video at two times the median or more. For each one write down the topic, the title pattern, the thumbnail’s main element, the length and the opening line. Then do the same for the five weakest videos.
You are looking for a pattern the winners share and the losers lack. Common ones: the winners answer a question the viewer already had, the winners name a specific person, tool or number in the title, the winners have one face and one object in the thumbnail while the losers have text over a screenshot. When the pattern is clear, the audit’s first recommendation writes itself: make more of the thing that already worked for this audience. That is a lower-risk bet than any new idea, because the channel’s own data has already voted for it.
If there is no pattern, or there are no winners, the channel has not yet found a topic its audience wants. Skip to check 5.
In Studio, open a recent video → Reach. You want two numbers: impressions and click-through rate.
Compare click-through against the same channel’s other videos, not against a universal benchmark. What counts as good varies with the traffic source: browse and suggested traffic click at different rates from search traffic, so a channel that moved from search to browse will see its rate fall while growing.
Open the retention curve for the same video. Read the shape, not the average.
Good click-through with early leaving is the most common combination on channels that have worked hard on thumbnails. The packaging is now writing cheques the opening does not cash. The fix is in the first thirty seconds of the video, not in the thumbnail.
Find five to ten channels of a similar size that serve the same viewers. Run check 1 on each of them. If their medians are several times higher than yours on comparable topics, the demand exists and the problem is on your side (checks 2 to 4). If their medians are as low as yours, the niche itself is small at this weight class, and the audit should say that plainly: no packaging fix changes the size of the audience.
Also look for topics that are working on those channels and absent from yours. A format proven for the same viewers, not yet used by you, is the safest new idea available. That is the method in how to find YouTube video ideas.
For a hobby channel this means subscribers per thousand views. For a business channel it means something stricter: does a viewer have a clear next step, is it stated in the video and the description, and can you count how many took it? A channel with healthy views and no measurable leads has a conversion problem that is invisible in YouTube’s own analytics. We cover that side in how to get clients from YouTube.
An audit that lists twenty things to improve is a to-do list, not a diagnosis. Write it as three lines:
YouTube Studio now lets you test up to three thumbnails and titles on a video, so a packaging recommendation can be tested rather than argued about. Record the prediction before you run the test. A recommendation that cannot be scored later was a guess.
VideoCore’s free channel audit runs checks 1, 2 and 5 on any public channel from real YouTube data: the median, the outliers, what they share and the fix to test first, with no account needed to run it. Connect your own channel inside VideoCore and the private Studio layer, click-through and retention, sits alongside it for checks 3 and 4.
For business channels where the audit shows a conversion problem worth solving, the same page can route you to ViewsToClients, the managed service run by Jake Trinder and the VideoCore team, where a strategist implements the plan with you. Whether that is offered depends on what the audit finds; it is not a fit for every channel and the page will say so.
Real data, not estimates: the median, the outliers, the pattern behind them, and the one fix to test first. No account needed to run it.
Run the free channel audit Diagnose one video insteadUsually for one dominant reason rather than many. Check, in order: whether recent uploads fall below the channel's own median views; whether videos that beat the median share a pattern you have stopped repeating; whether impressions are turning into clicks (a packaging problem); whether clicks are turning into watch time (an opening or pacing problem); and whether comparable channels are getting more views on the same topics (a demand problem). Stop at the first check that fails and fix that one thing.
A structured review of a channel's performance data to find the main reason it is or is not growing, and to name the fix to test first. A full audit checks recent performance against the channel's own median, the shared traits of its best videos, click-through rate, audience retention, demand in the niche, and whether viewers take a next step. The result should be one diagnosed problem, one fix and one measurable test, not a long list of suggestions.
Impressions without views means YouTube is showing your thumbnail and people are choosing something else, which is a click-through problem. Compare the video's click-through rate with your channel's other videos in YouTube Studio. If it is below your norm, change the title and thumbnail so they make a clearer, more specific promise, and use Studio's thumbnail test to compare versions rather than guessing.
The packaging is making a promise the opening does not deliver quickly enough. Look at the retention curve: a steep drop in the first thirty seconds followed by a flatter line means the hook, not the thumbnail, needs fixing. State the payoff in the first sentence, remove the introduction, or adjust the title so it describes what the video actually is.
Because it matched something the audience wanted and the others did not. Compare the outlier with your weakest videos on topic, title pattern, thumbnail composition, length and opening line. The traits the winner has and the losers lack are your channel's own evidence of what works, and making more videos with those traits is a lower-risk bet than any new idea.
Partly. Views, likes, comments, dates, durations and subscriber counts are public, which is enough to measure the channel against its own median, find its outliers and compare it with similar channels. Click-through rate, retention and revenue are private to the owner, so an outside audit cannot tell a click problem from a retention problem, and it should say so rather than estimate.
Automated audits from real public data, including VideoCore's, are free. Human audits by consultants and agencies are usually priced as a one-off report or bundled into a monthly managed service. Before paying, check that the audit uses your channel's own measured data rather than estimates, and that it ends in one prioritised recommendation you can test.
A strategist runs the audit, decides which problem to fix first, and then works through the fix with you: the topics to make, the packaging to test, the openings to rewrite and the way viewers are moved towards a next step. The useful ones commit to a measurable prediction for each recommendation and review it against the results.
Whichever the audit points at, not whichever is easiest. If impressions are low, the topic. If impressions are fine but clicks are low, the title and thumbnail together. If clicks are fine but retention drops early, the opening. Testing the thumbnail on a video nobody wanted, or the opening on a video nobody clicked, wastes the test.