Do not invent video ideas; find them. Start from videos that recently beat their channel's average, check that several of them exist on the same subject, and make your version better or different in one clear way. The idea decides more than the editing does: a strong idea survives average production, and no amount of editing rescues a weak one.
Plays from YouTube when you press play. Watch on YouTube
Start from recent outliers
Search your niche with the Videos filter and look for videos that scored at least five times their channel's average, especially on smaller channels. Weigh the upload date: a modest outlier from last month says more than a huge one from three years ago. The best topics have high demand and few existing videos. Chris Barrera 24:59
YouTube Studio offers a second source once your channel has viewers: the Research tab shows what they search for and marks "content gaps", searches with too few good videos. Nate Curtiss 5:57
Train a research account to surface ideas
Nate Curtiss creates a fresh YouTube account and watches at least 30 seconds of six or seven popular recent videos from each competitor. After a few refreshes the homepage shows what YouTube is currently pushing to that audience. An outlier extension puts a score on each recommended video, and he saves the breakout videos from the last six to twelve months as screenshots in his planning document. Nate Curtiss 19:15
The lesson on outlier research in module 3 covers this account in more detail.
Model proven videos, combine two winners, make it slightly better
A proven topic is an opening. Fabio Morena's first video to pass a million views was an improved version of a competitor's popular upload: tighter pacing, less filler, more of the best footage. When it worked he made sequels and spin-offs. Compete with channels you can realistically match, not with studios. Fabio Morena 12:05
Add what the others left out. Watch the top three videos on the topic, note their best points, and read their comments for questions nobody answered. Then cover all of it in one video. You can also combine two proven ideas from the same niche into one new title. Make Money Matt 4:00
A straight copy fails. Steffen Miro shows a direct copy of a viral video that flopped, because the core viewers had already seen the original. His rules: skip a topic that has been made more than three times unless you have a new angle, move quickly on trends, and take winning topics from other niches to your own audience. When something works on your channel, repeat it. Steffen Miro 19:37
Idea formulas: absurdity, familiar names, a format tied to a current topic
- Absurd but believable. Ideas that sound almost impossible earn more clicks, as long as the viewer still believes the video will deliver. List 10 to 20 extreme versions of your subject from the viewer's side, keep the plausible ones, and plan how you will really deliver. The bar keeps rising: what was absurd last year is ordinary now. Nate Curtiss 0:12
- Familiar names. A person or subject your audience already knows lowers the effort of clicking. When everyone covers a trend in the same way, choose the angle nobody took. Nate Curtiss 3:09
- A proven format plus a current topic. The "your life as every rank of…" format moved from the mafia to soldiers and beyond. A small channel applied it to a subject that was in the news and, by Tim Danilov's account, got half a million views in eleven days. Tim Danilov 9:30
- One sentence as a story. For Shorts, Daniel Bitton found ideas in long interviews: one striking sentence can carry a whole short story. This leans on other people's material, so it only holds up when you add your own framing and editing. Daniel Bitton 7:55
Validate every idea with several recent outliers
One successful video can be luck. For every idea, Nate Curtiss looks for at least four similar videos that overperformed in the last six months, and stores them with the idea as evidence. In niches where average views are high, he raises the view filter. Then he describes the video in a voice note and has Claude turn it into a four-sentence concept. Nate Curtiss 26:15
Using AI to widen the list
AI is good at producing many options and bad at knowing which one viewers want. Use it for the first and check the second on YouTube.
- Analyse several channels at once. E'Calm gives an AI the links of five small channels with very high views and asks what they have in common. He gets the pattern (titles that leave a gap between what the viewer knows and wants to know) and fifty topics. E'Calm 7:35
- Generate a series, then verify. A prompt returns lists of "every type of … explained" ideas per niche, and each idea suggests follow-ups. Before producing one, search YouTube for a comparable video with good views. Money Groot 1:45
Checklist
Your ticks are saved in this browser only.
Income and view figures mentioned in these videos (for example "half a million views in 11 days") are the creators' own claims and are not verified.

