How a Video Idea Earns Its Place on a YouTube Channel

It starts with a pet topic. Something you've held strong views on for years, or a question a client asked last week, or an idea that turned up in the shower.

So you make it. Time on camera, an edit, a thumbnail someone agonised over. It gets 200 views, and you blame the algorithm.

The algorithm isn't the problem. The audience is the algorithm. It put your video in front of some people, they chose not to watch it, and it stopped. Nobody chose that video except you.

This is the process I use on client channels, set out so you can apply it yourself. It comes from the work I do every month inside Channel Intelligence.

Before any of it: write down who the channel is for

None of the questions below can be answered without this, so it comes first. A page or two, written once, saying who you want watching: what they do, what they're trying to fix, what they've already tried, what they call the problem in their own words, and what would make them pick up the phone.

Most owners have this in their head and nowhere else. That works until a second person touches the channel. Then the researcher picks topics that sound right, the editor cuts for the wrong viewer, and whoever writes the title aims it at whoever happens to be reading. Everyone is doing their best and no two people are aiming at the same person.

It also decides two of the four questions. Question one asks whether buyers use your words or their own, which you can't check until you've written down what their words are. Question three asks whether the people a video attracts could ever pay you, which is unanswerable if you've never defined who pays you.

The Takeout: If you can't hand someone a document that tells them who the channel is for, you can't delegate a single decision about it, and every one of them will come back to you.

Where the ideas come from

The list doesn't start in my head, and it shouldn't start in yours. I look in four places. Each one shows what people already want, rather than what you'd enjoy explaining.

First, I look for white space on YouTube: questions people are clearly asking when the existing answers are thin, dated or beside the point. Real white space is rarer than it sounds, as I'll explain in question two.

I also look at rising searches in the niche. These are terms growing right now, before the market has noticed. They've got a shelf life, so I keep them separate from the evergreen list.

Then I read the communities where buyers spend time. In forums, groups and threads, people describe the problem in their own words, unprompted and without a salesperson in the room. That's where I learn their phrasing, which matters more than the topic.

I also read comments under the videos already ranking. Viewers say what the current answer left out. A comment with 400 likes saying "but nobody explains what happens if you're self-employed" is a video brief someone has written for you.

The Takeout: Spend an hour reading the comments on the five videos that rank for your main subject before you plan anything. It's the cheapest research available, and almost nobody does it.

1. Is anyone looking for this?

Every industry has its own vocabulary, but buyers rarely use it. They describe the symptom rather than the category.

I ran a demand check for a therapy practice. The practice's own term for its work had almost no YouTube demand at all. On the face of it, the subject was dead. But the demand was real. People were searching with the words they use to describe how they feel, which was a completely different set of phrases. If you check only the industry phrasing, you'll reject a good lane.

It happens in my own category too. I'd call what I do YouTube strategy. The people who hire me describe it as having a channel that gets views and brings in no enquiries. Same work, two vocabularies, and only one of them gets typed into a search bar.

I run the check twice: once using your phrasing and once using the words a customer would say out loud. Then I name the video using their words.

The Takeout: Say your video title out loud as though you're a customer explaining the problem to a friend. If it doesn't sound like something they'd say, you're using your vocabulary rather than theirs.

2. Can you beat an existing video

Demand on its own can mislead you. Most ideas die at this point.

Open the results page and read the top five properly. Check how many views they've got, how recent, how strong the packaging looks and how well they answer the question. Often, the page is full and the answers are good. The honest conclusion is that you should make something else.

When I pulled Holly Willoughby's channel, her own show name had 590 UK searches a month, and she owned it. One of the dishes she'd actually cooked on camera had 60,500.

The video exists, but it wasn't built to answer the search. It doesn't appear, so the hundred-fold bigger audience goes to somebody else.

The Takeout: Before you commit, open the results page in a logged-out browser and read the top five. If you can't explain in one sentence why yours would be the better answer, you shouldn't make it.

3. Would the people searching it ever buy from you

Plenty of topics are popular but commercially useless. They bring in people who've got no reason to become customers, and they teach YouTube to send you more people like them.

The subject you're best known for often isn't the one that works. I pulled all 425 videos on Martin Lewis's channel. Ask anyone in Britain what he talks about, and a good share will say energy bills.

Across 33 energy videos spanning eighteen years, the typical video lands exactly on his own average. The videos that travel are the tax explainers and budget reactions. The budget and tax groups are small samples with a wide spread, so treat the direction as real and the exact multiple as unstable. Energy, with 33 videos, is the solid one.

Asking around can tell you which subject you're famous for. It can't tell you which subject your videos win on.

The Takeout: For each idea, write down who'll arrive if it works and what you'd want them to do next. If you can't finish the second half of that sentence, the idea is entertainment.

4. What will it teach you or earn you

An idea that gets through the first three questions still needs to do a job beyond getting views.

Learning means finding out something you can act on next time. Perhaps this title format beat the last one, this topic brought viewers who'll never buy, or this opening lost half of them inside twenty seconds. Earning means getting a customer you can trace back to the video that sent them.

You should be able to answer both questions after every video. "It earned nothing" is a perfectly good answer. If you don't know, the video can't tell you anything, and you'll make the same decisions next month with the same lack of evidence.

The Takeout: Before you film, decide which of the two this video is for and put in place the one thing that'll answer it. Usually, that's a tracked link and a note of what you changed.

The three verdicts

Every idea that goes through this process gets one of three verdicts. Giving it a verdict is the point of the exercise.

1. Make it. There's demand, you can beat the existing answers, the audience could buy from you, and you know what you'll learn or earn.
2. Make it with an edge. The demand is real and the page is competitive, so the video will only work if you can explain in one sentence what yours does that the top five don't. That could be access, data, a client story or a stronger demonstration. If you can't name the edge, use the third verdict.
3. Don't make it. Most ideas end here, which is why I run the process.

You need to be willing to give the third verdict to an idea you liked.

Put simply

A video costs you a day on camera, an edit, a thumbnail and one of the limited slots you've got. An idea needs to justify that cost before the camera comes out. Most ideas can't.

Build your list from the places where your buyers already spend time, rather than from your own head. Check demand using their words. Read the page you'd compete on. Ask whether the people the video attracts could ever pay you. Then decide what it needs to teach you or earn you, and set up the method that'll give you the answer.

Your pet topic might make it through those checks. Some of mine do. I still put it through the same process as everything else.

Method and limitsWhere the examples come from and what this article can't tell you

The Martin Lewis figures come from all 425 videos on his channel, pulled from the YouTube API on 22 September 2026. I compared each long video with the median of the long videos published in the same calendar half-year. I counted only videos that were at least 90 days old and only half-years containing five or more. Lifetime medians are confounded by age, so I never use them to claim a rise or fall.

The Holly Willoughby figures come from a pull of all 40 videos on that channel in August 2026, with UK search volumes from Google Ads data for the same month. Search volumes are monthly averages, and they move with the seasons.

The therapy practice is a real 2026 case, described without naming the business.

This is the shape of the process, rather than the whole method. Reading a results page properly, separating real demand from a seasonal spike and distinguishing organic performance from bought views each require their own method. I haven't covered any of them here.

Nothing here proves cause. When I suggest a reason, I'm giving my reading of what happened rather than a proven explanation.

Analysis and write-up: Stewart Read.