Most science papers are only read by a tiny circle of experts, maybe a dozen or so. The process still works out. The critical challenge is to have some quality human readership (instead of the AI bot brigade and random zero second clicks).

Theoretically the internet would have enabled the long tail, making tiny domain focused niches viable. In practice discovery is the most gamed algorithm in existence. Unless you have the resources and inclination to spend quality time towards useless "SEO" nobody knows you are alive.