Everyone's chasing 'faceless YouTube' right now, but most people pick the wrong lever. They obsess over AI voice quality or B-roll when the thing that actually decides whether a channel survives is the hook-to-niche fit.

Here's the breakdown I use before touching any editing tool:

• CPM range tells you the ceiling, not the floor. Finance/legal niches pay $15-45 CPM but need researched scripts — most beginners rage-quit before month 3 because the writing bar is higher than they expected.

• Automation potential tells you the floor. Mythology, motivation, and true crime are 'Very High' automation niches — you can run 3-4 channels solo because scripting is formulaic (setup → stakes → twist → payoff).

• The hook is 90% of retention. If your first 3 seconds don't promise a specific payoff ('Rome had a machine no one could rebuild for 1,000 years'), no amount of AI voice polish saves the video.

My honest advice if you're starting today: pick ONE formulaic niche, write 10 hooks before you write a single full script, and only scale to a second channel once the first one has a repeatable format that hits.

Happy to share the exact hook framework I use if people want it — just ask below.