They pick topics they're "passionate about." Or they copy the biggest channels. Both are reliable ways to build a channel that never monetizes.
Passion is irrelevant when you're making AI-generated content. You're not the talent — you're the operator. What matters is the math: does this niche have an audience that advertisers pay a premium to reach? Can you produce content without showing your face? Is the competition workable for a new channel?
This lesson gives you a repeatable scoring framework so the decision is mechanical, not emotional. Pick the niche with the best score. Move on.
Rate each of these four variables from 1 to 10. Average them. That's your Niche Score out of 10. Anything above 7.5 is worth building. Below 6, walk away.
YouTube search volume + trend direction. Use VidIQ or TubeBuddy to check monthly search estimates on 5 seed keywords. Growing = 9–10. Flat = 6–7. Shrinking = 1–4. A niche nobody is searching is a hobby project.
Revenue Per Mille — what you earn per 1,000 views from YouTube AdSense. Finance pays $20–45. Lifestyle pays $2–8. Use the RPM ranges in the niche table below to score this. High RPM = ads from banks, software companies, insurance firms targeting your viewers.
How naturally does this niche work without showing a face? History, finance explainers, and sleep stories score 10. Cooking tutorials and fitness score 3. Ask: can this be produced with AI voiceover + stock footage + B-roll alone?
Inverse of saturation. More competition = lower score. Search your 5 seed keywords on YouTube. If the top results are all channels with 1M+ subscribers, score 3–4. If you're seeing channels with 50K–200K getting strong views, score 7–9. Niche is a spectrum — "finance" scores 3, "finance for truck drivers" scores 8.
Pro move: Never score "finance" — score a sub-niche. "Personal finance for millennials" is a different market from "options trading explained simply." Sub-niches have lower competition, more loyal audiences, and often identical RPM to the parent niche.
RPM ranges are based on typical YouTube AdSense performance. Actual RPM varies by geography, seasonality, and audience quality — but these ranges are realistic benchmarks for US-majority traffic.
Financial advisors, brokers, and fintech apps bid aggressively for this audience. Text-to-speech + stock charts = zero face required.
Fastest-growing search category. Screen recordings, UI walkthroughs, and B-roll of terminals. Advertisers: SaaS companies with fat CAC budgets.
Explainer-style content works perfectly faceless. Pharma, supplement brands, and health apps all advertise here at premium rates.
How companies succeeded or failed. Decision-maker audience means premium ad targeting. Illustrated timelines + narration = fully faceless.
B-roll of devices + screen recordings. No face required. High-intent audience drives click-through rates that push RPM up significantly.
Massive audience, high watch time, stock footage + AI narration. RPM is mid-tier but volume compensates. Evergreen content drives passive views for years.
One of the purest faceless niches. AI-generated visuals, narration, and map animations. Timeless content earns views indefinitely with no updates needed.
Evergreen content that never expires. B-roll of nature/cities + licensed audio + AI voiceover. Highest Faceless-ability score in the list.
Aspirational content targeting high-net-worth adjacent audiences. Stock footage libraries cover yachts, cars, architecture. Premium ad categories.
No face, no visuals needed — just audio. Lowest production cost in this list. RPM is low but watch time is extraordinary, boosting algorithm ranking.
The RPM trap: Finance has the highest RPM but also the most competition. A finance channel with 500K views at $30 RPM earns $15K. A true crime channel with 5M views at $6 RPM earns $30K. Volume beats rate at scale. When starting out, pick the niche with the best Niche Score — not the highest RPM.
You don't need to master all of these on day one. Start with scripting + voice. Add visuals when you're publishing consistently. Add automation last.
Prompt with your niche, target keyword, hook formula, and desired length. Claude tends to produce cleaner narrative structure; ChatGPT is faster for high-volume batch scripting. Use a custom system prompt that locks in your channel's tone and audience level. Output is a full script with timestamps, B-roll notes, and chapter markers.
ElevenLabs is the current standard. Pick a voice that matches your niche's energy — not the default robot voice. Clone a voice if you want consistency. A 10-minute script takes about 3 minutes to generate and sounds indistinguishable from a real narrator at normal playback speed. Pro tip: run audio through a subtle compressor in CapCut to add broadcast warmth.
Midjourney for hero images and scene illustrations. Runway and Pika to animate stills into 4-second video clips. Storyblocks or Artgrid for high-production stock footage when you need real-world visuals. The combo of stock + AI-generated fills any content gap. Finance and history channels run almost entirely on AI visuals + stock charts.
CapCut handles full video assembly: drop voiceover + B-roll + music, auto-sync, export. Submagic adds word-by-word captions styled for reels (critical for Shorts and mobile). Descript is the power tool — lets you edit video by editing the transcript. For long-form explainers, Descript saves 60% of edit time vs. traditional timeline editing.
Build a Canva template with your channel's color palette and font. Generate the hero image in Midjourney, drop it in, add bold text overlay. The thumbnail decides whether someone clicks — invest time here even when everything else is automated. Test two variants per video using YouTube's built-in A/B testing once you hit 1,000 subscribers.
VidIQ shows monthly search volume, competition score, and trending videos for any keyword. Use this to validate your Niche Score and find specific video ideas that have proven demand. Don't script a video without checking the keyword first. A video with no search intent is a lottery ticket; a video targeting a 20K/month keyword with low competition is a business decision.
Not "AI" — that's too broad and too saturated. "AI tools for solopreneurs" is a defined audience with specific intent and purchase behavior. Here's how it scores.
Reasoning: Demand is high and growing (AI interest accelerating into 2026). RPM is strong at $12–$25 — SaaS companies advertise heavily here. Faceless-ability is nearly perfect: screen recordings, UI demos, and AI-generated visuals cover 100% of content needs. Open Field docks 4 points because "AI for productivity" is competitive — but "solopreneurs" specifically is still workable for new channels.
Notice the pattern: Every title has specificity (who it's for + what they get), a clear benefit or curiosity gap, and works as screen-recording or B-roll content. You could script all five in under 2 hours using Claude with a properly formatted prompt.
Lessons 1–2 are free. Lessons 3–12 cover everything else: script writing, AI production, algorithm optimization, monetization stacking, and the system that scales it.
The exact spreadsheet, prompt library, and 30-day posting schedule we use to launch faceless channels from zero. Free, forever.