The algorithmic logic: recommendation systems classify your content and match it to interested viewers, and consistent topics produce confident classification. A focused account gets pushed to the right people, and a scattered one confuses the system and the audience alike. The commercial logic: brands buy specific audiences, and "the lens-cleaning guy" gets the camera-brand deal over a generalist with triple the followers.
The counterweight is real: niches can become cages, and many successful creators broaden once trust is established. The niche gets them discovered, and the personality keeps people around. Practical guidance has converged on niching by audience and problem: "helping freelancers price their work" travels across formats and platforms, while "talking-head pricing tips" stays locked to one format. Testing several topics, then doubling down where retention and comments concentrate, beats committing blind.