You can, you say it by grouping on only columns from the customers table, and those columns including a candidate key. Necessarily and equivalently, this means everything not from the customers table must be only referenced in an aggregate expression, but that's trivially what “I want a single row per customer, with some data not from the customers table” is asking.
> it's a consequence of how you set up your JOIN and GROUP BY keys,
Well, its a consequence of the GROUP BY. which is literally “what do you want one row for each of”, so... it's kind of weird to complain that it should be something else. GROUP BY is the dedicated construct in SQL that soecifies the thing you are looking for.