I think I read somewhere the beam size for starlink is around 200km diamater. That's actually really far. You'd probably want to service rural areas fairly distant from metro areas well, rather than slam the sat with a handful of suburban city customers.
I could be wrong but my guess is they will blacklist areas near cities entirely at least to begin with and focus on very rural areas with literally no alternative (no LTE etc).
Personally I think Starlink is going to have capacity crisis quickly if takeup is anywhere near as strong as demand seems to be and I'm sure SpaceX know this, so by fcoussing almost entirely on very rural areas (and maybe niche commercial/govt uses in more urban enviroments) they have a chance at managing demand a bit. Or they will start ramping up cost of service and/or putting very punitive bandwidth caps on (like all other sat providers).
I still think Starlink is an amazing breakthrough, but it really is a solution only for the 'last'/most rural 1% of households (or less).
I assume they will eventually dynamically price based on density, to ensure it's available to whoever needs it and not where other equivalent or faster options are available for less money.
By definition I think the places that need Starlink the most won't hit density limits. But I think a lot of people will buy Starlink just because it's Starlink -- anything so their dollars aren't going to Comcast.
They may ration service to a few people in big cities but i'm doubtful they will offer it at all to be honest.