Ask most business owners where they rank and you get a single number. "We're number two for coffee shops." It feels like an answer. It is closer to an average temperature for a country.
Why one position is the wrong unit
Local results are built around the searcher's location. Two people standing four streets apart, typing the same words into the same app on the same afternoon, routinely get different results. Proximity is not a tiebreaker in local search — it is one of the primary inputs.
So "we rank number two" is really "we ranked number two, at whatever point Google inferred when that check ran." Move half a kilometre north and you might be eleventh. Nobody in the business would know, because nobody ever measured from there.
That gap is where customers quietly go elsewhere.
What a grid actually does
A geo-grid takes the area around your location and lays a lattice of measurement points over it. Each point is a separate query, run as though the searcher were standing exactly there. A 5×5 grid is 25 measurements. A 9×9 is 81.
The output is a map rather than a number: your position at every point, which reveals shape. Most businesses find one of a few patterns.
A tight core. You rank well within a few hundred metres and fall away sharply. Common for businesses whose profile is fine but whose prominence is thin — few reviews, little linking, no local coverage.
A lopsided field. Strong on one side, weak on the other. Almost always a competitor sitting on that side with a stronger profile, not a problem with yours.
A uniform middle. Mediocre everywhere, no sharp centre. Usually a relevance problem: category, services or description do not clearly say what you sell.
Each pattern implies different work. A single averaged position implies nothing at all.
Choosing size and radius
The number that matters is not grid size or radius on their own — it is the spacing between points, which is the radius doubled and divided by one less than the grid size.
A 5×5 across a 2 km radius puts points 1 km apart. In a dense city that is coarse: entire neighbourhoods fall between measurements. A 9×9 across the same radius drops spacing to 500 m, which resolves properly.
The tradeoff is cost. Every point is a separate lookup, so a 9×9 costs roughly three times a 5×5. Dense urban catchments justify the density. A rural service area with customers spread over 20 km does not — there, a wide radius with coarse spacing genuinely reflects how people search.
Reading the result honestly
Three numbers are worth more than the map at a glance.
Average position across points where you appear at all. Useful for trend, misleading in isolation, because it ignores everywhere you did not rank.
Top-three share — the percentage of points where you appear in the local pack. This tracks discovery far better than average position, because the pack is what most people actually see.
The weak edge — which direction is losing. This is the one that turns into work.
What a grid cannot tell you
It cannot tell you why. A grid is a measurement, not a diagnosis. It will show you that you fall away to the north; it will not tell you whether that is a competitor, a category mismatch, or simply a lake.
It also cannot promise repeatability. Search personalisation, history, device and the moment of the query all move results. Any single scan is a sample. The value is in the series — the same grid, the same keyword, measured repeatedly, so that a change you made has something to be measured against.
Anyone selling you a grid as proof of causation is overselling it. Used properly, it is the most honest picture of local visibility available.