Which niche should I pick, and how do I write a good script?
A creator asked both halves of this in one breath during our live session on 4 September 2026, and they belong together. The niche decides whether anyone is shown your video. The script decides whether they stay. Here are the three niche checks, and then the script shape lifted, beat by beat, from a video that did 22.9 times its channel's normal.
Short answer
There is no single best niche. A good niche passes three checks: small channels are getting big views in it right now, it earns enough per view for your goal, and you can make 50 videos in it without running out of ideas. A good script is a chain of small promises, each one kept before the next is made.
The three niche checks
- Small channels win there. Search the topic in OutlierKit. If channels with under 20,000 subscribers are getting 5 to 20 times their normal views, the topic is open. Strangers are clicking on videos from channels they have never heard of, which is exactly your situation on day one. If only huge channels appear, the topic is owned.
- It earns enough. Finance, business, and education pay more per view than entertainment. Decide what you need before you choose. A niche that passes the first check but pays a fraction of what you need is a hobby, and that is fine, as long as you chose it knowing. The niches guide has the per-niche earnings picture, and the RPM benchmarks collect every published range.
- You can make 50 videos. Write down 20 video ideas now. If you cannot, the niche is too narrow for you. A channel needs a year of weekly videos to find its shape, and a niche that runs dry at video twelve will not get you there.
Geography is the worked example on this page and passes all three checks. Three geography channels of very different sizes were overperforming on 4 September 2026, from 1,200 to 19,000 subscribers, which is the first check passed; the starting-a-channel post shows all three. Education-style facts content earns at the middle of the range. And "N facts about [place]" can be made by country, by continent, or by theme for years.
The video the script shape comes from
23 Geography Facts That Sound Wrong, Until You Look at a Map
By Price of Travel, 19,000 subscribers. 496,000 views, which is 22.9x the channel's typical video.
Figures pulled from OutlierKit on 4 September 2026; they drift as views accrue. Published on YouTube 2026-07-12.
Price of Travel has 19,000 subscribers. This video got 496,000 views, 22.9 times the channel's normal. It is faceless for most of its length and plainly edited. What it has is a script that makes a small promise every 20 to 30 seconds and keeps every one of them. The timeline below is that script's shape.
Check one, across four channel sizes
Here is the first check passing in geography, with the video above and three more from the index pulled on 4 and 7 September 2026.
12 Geography Facts About the UK That Will Surprise You
By Curios Earth, 1,200 subscribers. 6,800 views, which is 5.7x the channel's typical video.
Figures pulled from OutlierKit on 4 September 2026; they drift as views accrue. Published on YouTube 2026-08-07.
15 Australia Geography Facts You Won't Believe Are True!
By Maply, 5,720 subscribers. 113,000 views, which is 86.9x the channel's typical video.
Figures pulled from OutlierKit on 7 September 2026; they drift as views accrue. Published on YouTube 2025-08-17.
25 Mexico Geography Facts That Sound Fake (But 100% Real)
By Borderline Wonders, 15,400 subscribers. 347,000 views, which is 26.7x the channel's typical video.
Figures pulled from OutlierKit on 7 September 2026; they drift as views accrue. Published on YouTube 2025-10-26.
Curios Earth, with 1,200 subscribers, got 6,800 views, 5.7 times its normal. Maply, with 5,720 subscribers, got 113,000 views, 86.9 times normal. Borderline Wonders, with 15,400 subscribers, got 347,000 views, 26.7 times normal. With Price of Travel at 19,000 subscribers and 22.9 times, that is four channels under 20,000 subscribers all beating their normal by more than five times. The topic is open.
The script shape, taken from the 22.9x video
| Time | What happens |
|---|---|
| 0 to 15 seconds | Hook: one fact that proves the title is true, told as something that surprised you. |
| 15 to 25 seconds | Why it matters: "my mental map was wrong, and yours probably is too." |
| 25 to 30 seconds | What is coming: repeat the number from the title. |
| Main part | Each item is "what you would assume" and then "what is actually true". 20 to 30 seconds each. Put the biggest surprise about two thirds of the way through. |
| Ending | Point to the next video in the series. Do not say goodbye for a minute. |
In words: the first 15 seconds is the hook, one fact that proves the title is true, told as something that surprised the creator. Seconds 15 to 25 say why it matters: my mental map was wrong, and yours probably is too. Seconds 25 to 30 say what is coming by repeating the number from the title. The main part gives every item the same two-beat shape, what you would assume and then what is actually true, 20 to 30 seconds each, with the biggest surprise about two thirds of the way through. The ending points to the next video in the series and does not say goodbye for a minute.
Before you write your own, read the transcripts of the top three videos in your niche in OutlierKit and mark where each one makes and keeps a promise. Your script should follow the same rhythm. The hooks and retention post covers the first 30 seconds in more depth, the script writing guide has templates for other formats, and the free script template gives you the beats as a fill-in document.
Why the niche and the script are one decision
The niche check tells you whether strangers are clicking on videos in that topic. The script shape tells you whether those strangers stay once they click. Pass the first and fail the second and you get impressions with no watch time, which YouTube stops rewarding after a few videos. Pass the second and fail the first and you have a beautiful script nobody is shown. The 22.9x video passed both, which is why a 19,000-subscriber channel reached half a million people.
If you have a niche already and want to know what is moving inside it, the post on finding trends in your niche picks up from here. If you are comparing several candidate niches, the prompt at the end of this page checks three at once and then drafts a script outline from the best one's transcript. There is also a longer discussion of how to find your YouTube niche on the blog.
Words used on this page
- Niche
- The topic and audience a channel serves. Narrow enough that one viewer wants every video, wide enough for 50 of them.
- RPM
- Revenue per thousand views, after YouTube's share. The number that makes finance pay more than entertainment.
- Open loop
- A promise made and not yet kept. A script is a chain of them, each closed before the next opens.
- Outlier multiplier
- A video's views divided by what its channel usually gets. 22.9x means the video did 22.9 times the channel's normal.
Frequently asked questions
How do I choose a niche for a YouTube channel?+
Run three checks. First, small channels win there: search the topic and confirm channels under 20,000 subscribers are getting 5 to 20 times their normal views. Second, it earns enough: finance, business, and education pay more per view than entertainment, so decide what you need before you choose. Third, you can make 50 videos: write down 20 ideas now, and if you cannot, the niche is too narrow for you.
What is the best YouTube niche?+
There is no single best niche. A niche that is best for one creator fails another, because the three checks give different answers for different people. Geography passes all three for a creator who likes maps and does not need a high RPM. Personal finance passes for someone who wants high earnings per view and has 50 money questions to answer. Pick by the checks, not by a list.
How do I write a good YouTube script?+
A good script is a chain of small promises, each kept before the next is made. In the first 15 seconds, prove the title with one fact. In the next 10, say why it matters to the viewer. Then repeat the number from the title as the roadmap. In the main part, give each item the same shape, what you would assume and then what is true, 20 to 30 seconds each, with the biggest surprise two thirds of the way through. End by pointing to the next video.
How long should each section of a YouTube video be?+
In a list-format video, 20 to 30 seconds per item works well, because a new payoff arrives before attention drifts. The hook should prove the title within 15 seconds. The whole opening, including why it matters and the roadmap, should be done by 30 seconds. The ending should be short: point to the next video and stop, rather than saying goodbye for a minute.
Should I study transcripts before writing my script?+
Yes. Read the transcripts of the top three videos in your niche and mark where each one makes and keeps a promise. Note the timestamps. Your script should follow the same rhythm, with your own facts and your own voice. OutlierKit pulls transcripts with timestamps, in the app or through Claude with the MCP connector.
Do this yourself in Claude
Paste this prompt into Claude with the OutlierKit connector switched on. Replace everything in square brackets with your own niche, channel, or topic. Claude calls OutlierKit for you and writes back the answer.
Using OutlierKit, check [topic A], [topic B], and [topic C]. For each, search outliers and tell me how many of the top 20 are on channels under 20,000 subscribers. Pick the most open topic. Then pull the transcript of its best outlier and write me a script outline that follows the same timing.
- Tools Claude calls
- outliers_search (three times), videos_transcript
- What it costs
- 4 credits. Every call is 1 credit, with no daily limits.
Plan requirement: the OutlierKit MCP connector is available on the Pro plan ($49 a month, 500 credits) and the Max plan ($199 a month, 2,000 credits) only. It is not included in Hobby ($29) or the free trial (10 credits). To connect it, open Claude, go to Settings, then Connectors, and add https://outlierkit.com/mcp. The full setup and all ten tools are on the OutlierKit MCP server page, and plans are compared on the pricing page.
The other questions from the same session
This post answers one of sixteen questions creators asked in the OutlierKit "Find Your Next Viral Idea" live session. Each answer uses real videos, verified live on 4 September 2026, and ends with a prompt you can run in Claude for your own niche.
- How do I write hooks that keep people watching, and hold retention mid-video?
- How do I get Shorts viewers to watch my long-form, beyond the related-video link?
- Why do faceless channels get demonetized, and how do I avoid it?
- If my channel answers questions, must I stay in one subject, or can I cover others?
- How do I get multiple viral videos without depending on YouTube recommendations?
- How do I find the trends in my niche?
- What does a channel watchlist reveal about trends in my niche?
- In 2026, is it quality over quantity on YouTube?
- What is the best strategy for starting a YouTube channel?
- How is OutlierKit different from vidIQ and NexLev, and can I use them together?
- What is the one thing every outlier video has in common?
- How do I get more views on YouTube?
- What is the quickest way to find the biggest content gap in my niche?
- How can I use AI to make videos fast without breaking YouTube's rules?
- How do I get deep, relevant data on what works in my niche?
Real channel breakdowns
See these strategies in the wild — full data-backed analyses of channels in this niche, including outlier videos, upload cadence, and growth patterns:
OutlierKit Channel Analysis
Johnny Harris
Documentary storytelling
- Subscribers
- 7.8M
- Avg views
- 2.3M
- Total views
- 1.2B
OutlierKit Channel Analysis
Kurzgesagt
Animated science
- Subscribers
- 25.3M
- Avg views
- 10.1M
- Total views
- 3.7B
OutlierKit Channel Analysis
Cleo Abram
Optimistic tech explainers
- Subscribers
- 8.3M
- Avg views
- 5.0M
- Total views
- 2.9B
Stats are from our most recent snapshot of each channel. For live numbers, outlier videos, and up-to-date revenue estimates, run a fresh analysis on OutlierKit →
Run the three checks on your shortlist
OutlierKit shows which of your candidate niches has small channels winning right now, then hands you the transcript of the best one to script from.
Try OutlierKit Free