What is a geogrid in SEO (local ranking heatmap)?

Three settings define a geogrid completely, and between them they decide what it costs and what it shows.

The three settings constrain each other. Widening the radius without raising density stretches the step, and the map becomes a coarse approximation. Raising density at constant radius multiplies readings — and cost — to refine an area that may not be where the customers are.

The colour convention is stable from one tool to the next, which is what makes these maps instantly readable: green for positions 1 to 3, amber for positions 4 to 10, red beyond the top 10.

Three metrics read off a single map. Top 3 share of voice is the proportion of green points: forty-two green points out of a hundred and twenty measurements is 35%. Average position is the mean of the ranks recorded — useful for a trend, misleading as a description, since it blends places with no commercial relationship to each other. The dominance area is the continuous green patch around the address: its shape and extent say far more than any single number.

One reading note that applies to all these maps: the green patch is almost never a circle. A river, a railway line, a competing retail park deform it — and that is precisely what a geolocated measurement gives you over a theoretical radius.

A geogrid generates its locations by calculation: radius, density, and the coordinates land where they land. The other approach is to declare locations one by one — the town centre, the business park, a competitor's address, the neighbouring town that sends you 20% of your customers.

The two methods answer different questions. A grid answers "how far does my visibility reach?". Declared points answer "am I visible where I want to be?". The first is a discovery question, the second a management one — and you do not ask them at the same frequency.

A grid is a geometric figure. It knows nothing about where people live, where the offices are, where the retail parks and traffic routes run. On a 7 × 7 grid around a tradesperson based on the edge of town, it is common for fifteen points to land on farmland, woodland, or an industrial estate that empties at the weekend.

Those points get measured, billed and coloured — and they describe no commercial opportunity at all. Worse, they enter the average: a grid with a third of its points where nobody searches produces a mechanically low top 3 share of voice, and a trend line that moves for reasons unrelated to the market.

The mirror problem exists too. A neighbouring town supplying a large share of the customer base can fall between two grid points, or just outside the radius. It is then not measured at all, and nothing on the map says so.

A town-centre shop. Customers come from within a mile, mostly on foot or by public transport. A tight grid over that radius makes sense: every point lands on an inhabited street, and the map describes a homogeneous territory. This is where a geogrid is most faithful.

A tradesperson with a service area. The business is registered in a village and works across thirty parishes. A grid centred on the registered address spends most of its budget on empty ground and misses the towns that generate the revenue. Here, a handful of chosen locations — the six places that matter most — describes reality better than forty-nine calculated points.

The full comparison of both methods, with cost per measurement and a five-step sampling method, is the subject of the next article in this section. For the general framework of local rank tracking, see the complete guide.