Firstly - keep in mind most advice about SEO becomes outdated fairly quickly if it's "cheap and effective". Secondly, I've have very little to do with SEO in t

Firstly - keep in mind most advice about SEO becomes outdated fairly quickly if it's "cheap and effective". Secondly, I've have very little to do with SEO in the last 7-8 years, so most of my advice should be viewed very much through the lens of my first point. But, my advice for what it's worth - optimise for long term, which for me mostly means doing the most reliable technical thing that avoids the bulk of the SEO problems. I'd go for geographical subdomains of a keyword relevant .com - or second choice geographical sub-paths. That give you the greatest ability to easily automate new regions - and avoids manual search and selection of unique top level names each time, as well as ensuring you don't fall victim to entire odd-ball TLDs (line .online) sometime in the future becoming spam-havens and SEO "bad neighborhoods". You'd need to talk to someone reliable and reasonable who's current on the SEP impact of subdomains vs paths, but I did _very_ nicely using city and state based subdomains (over 1600 of them at peak) back int he mid 2000's (which is likely to have put it in the "cheap and effective" category - so perhaps the ranking algorithms have punished it in the meantime...)