Comparing locations in different markets
A location scoring lower than its sibling may be performing better. Raw comparison across markets is the most common misreading in multi-location work.
The problem with a league table
Ranking locations by raw score is the obvious thing to do and it systematically misleads, because the score is partly a property of the market rather than of the branch.
A location that is one of four options in a small town and a location that is one of ninety in a city are being asked to do different things with the same evidence.
What density changes
How much evidence is needed to be named. In a thin field, modest corroboration is enough. In a dense one, the same corroboration puts a business in the middle of a crowd.
How stable the result is. Dense fields produce more volatile selection results, because many near-equivalent options mean small differences reorder the answer.
What improvement is worth. Moving from unnamed to named is worth far more in a dense market and is far harder there.
Reading it correctly
Score each location against the options a customer in that market actually has, not against its siblings. A branch at 55 in a field of ninety may be outperforming a sibling at 72 in a field of five, and a league table will tell head office to congratulate the wrong one.
What to compare instead
Position within the local field. The only number comparable across markets.
The gap to the local leader. Actionable in a way an absolute score is not.
The specific missing signal. If nine locations have local corroboration and one does not, that is a finding regardless of scores.
The exception
Anything that is a property of the business rather than the market compares directly across locations: profile completeness, description consistency, review response rate, and whether the location is retrievable at all. Those should be identical everywhere and a variance in them is always a real finding.
Multi-Location AI Visibility is largely the discipline of separating those two categories before anyone draws a conclusion.