I think the most important question is how long it takes for you to effectively win the game, and thereby, when you need your cities and your tiles to return on the investments made. Better locations return the investment quicker, and more cities in the same amount of area lets you work the tiles sooner.
So if you're winning on turn 200, it's always wrong to settle a city on Turn 160, then build Granary, Water Mill, etc. That's true no matter how good the location is. On the other hand, even marginal terrain is worth settling if it's early enough. I would much rather take a 1 Fish, 1 Deer Tundra location overlapping my Capital than walk the same Settler 20 turns to get to a double Wheat double Luxury location on a far island. But if it's early enough, I will try to take both.
I think it's also a mistake to make, for example, a second city situated just right so that you get two luxuries, three food resources, etc etc, mostly on the third ring. You end up waiting dozens of turns to work those tiles, when making two close cities on the same ground lets you grab all of it right away. A lot of the screenshots I see seem to indicate a bias for what a size 25 monstrosity looks like on turn 280, when prime tiles have gone to waste every turn up until then. I care much more what my cities look like on turn 100 at size 6, because the outcome of the game is usually decided by turn 200 at the latest.
It's also a mistake for players to want to grab too many "plain" tiles. Where I see this most often is in coastal locations. Say a city is 4 tiles from the coast, then there's a few hills maybe on the third ring in that direction, and 2 Fish off into the Ocean. Some players will earmark all the hills for the first city, and so they won't found a second coastal city there because they already think those tiles belong to them. Or maybe they just really hate seeing so much overlap. But the overlap is on plain tiles, and they're foresaking premium Fish tiles. Correct is to found a city there, grow it with the premium Food tiles until to the point where it's working 3 or so production tiles, then staff Artists and Scientists.
Quite often I see mistakes in undersettling and overgrowing, and there are very few maps I can point at and say, hey, you built way too many cities.