He opened his channel analytics to see where the views were coming from. Most videos had reasonable numbers, and one of them caught his eye: a video two years old, about a question people still ask, which the platform was still showing to thousands every month, and barely any of them clicked.
The video itself is good. People who watch it finish it. The problem is that it is being shown and not opened.
And this is the cheapest work available to a creator: a video that exists, already shot, already edited, whose audience is searching for it, and what it needs is an hour's work on its cover rather than a new week of filming.
How to find the videos worth it
Sort your videos in analytics by the last three months rather than all time, and look for three patterns:
1. High impressions, low clicks. The platform is showing the video and people are not opening it. The fault is in the cover: the thumbnail or the title. This is the highest-value opportunity on the list. And before you judge the number, read what it means in Click-Through Rate: What It Means and Why There's No Fixed Good Number.
2. A video that worked and then faded. It used to bring steady views and then slid down. Usually because a fact in it has aged, or because somebody else published something newer.
3. A permanent subject with small numbers from day one. A question people search every year, where your video was not written in the words people use, and that is the subject of How do people find your videos in search?.
What you leave alone
One rule prevents a loss: a video that is working does not get changed.
If a video brings steady views with a reasonable click rate, changing its title or thumbnail is gambling with something that works. The same applies to a video published this week: let it settle for a month before you judge it at all.
And a weak video with no impressions in the first place is not an update case. No impressions means the platform is not showing it, and changing the cover does not fix that. That is a video where you make a different decision: keep it, unlist it, or remake it as a new video.
The order: what to change first
| What you change | When | The effort |
|---|---|---|
| The thumbnail | High impressions, low clicks | An hour |
| The title | The title does not carry what people search for | A quarter of an hour |
| The description and its first two lines | The description is stale or has broken links | Ten minutes |
| The pinned comment | To answer a recurring question or correct a fact | Five minutes |
| Playlists | The video is not in any playlist | Ten minutes |
Start with the thumbnail because it has the largest effect, and its rules are in The Thumbnail: What Actually Makes Someone Click. Then the title, whose formats are in Writing the Title and Description.
The cheapest item in the table is the last one: a video reaching thousands every month and sitting in no playlist is wasting a second view it would otherwise have taken, and building playlists is in Your channel page and playlists.
The one-change rule
Change one thing, then wait.
Anyone who changes the thumbnail, the title and the description in one sitting and then sees the numbers improve does not know which one worked, so they cannot repeat it on another video. The aim is not to fix one video but to learn what works on your channel.
The waiting period is two to four weeks. In the first days after a change the numbers are turbulent, because the platform re-tests the video on a fresh audience, so judging on day three is judging noise rather than a result.
How to compare honestly
Before you change anything, write in a file: the date, the video, exactly what you changed, and three numbers from the last twenty-eight days: impressions, click rate, and views.
After the change, compare twenty-eight days with twenty-eight days. Do not compare a week with a month, and do not compare a period containing a season or an occasion with an ordinary one.
And if the rate improved while total views fell, look at impressions: the platform may have reduced how often it shows the video for some other reason unrelated to your change.
Three cases and their answers
A fact in the video has aged. The audio cannot be edited, so add on-screen text if you can re-export, or put the correction in the pinned comment and in the first line of the description. And do not delete a video because a fact in it has changed.
The title is written in language nobody searches with. A clever or poetic title looks good and never gets typed into a search box. Rewrite it in the phrase people actually use, and save the cleverness for the thumbnail.
The video is good and the subject has outgrown it. An update is not enough here. Film a better new video, leave the old one, and put a link to the new one in its description.
How many videos a month
Do not turn this into a project that eats your production time. Two videos a month is enough: pick the two with the highest impressions and the lowest clicks, change one thing in each, and write down what you did.
After six months you will have updated twelve videos, and you will have a record telling you which kind of thumbnail works on your channel. That record is more useful than any general rule anybody gives you.
And if the problem is broader than scattered videos, with the whole channel stuck at one number, that is a different case, and reading where viewers leave is where it starts, in The Retention Curve: Where Viewers Leave and Why.
The idea that stays
Most creators publish and never go back. Every video you have that is more than a year old and still getting impressions is work you paid for once, still waiting for a cover that deserves it.