How far does a Google Maps ranking reach? 183 geo-grid scans of UK small businesses

Key findings

• In our scans, ranking in the centre of a map grid did not mean broad visibility across the surrounding area. In the latest scans, 83.9% of centre pins were in the top three, while only 12.0% of points within one kilometre were.
• In our latest-run distance view, no sampled points beyond three kilometres appeared in the top ten. That is a pattern in these scans, not a universal radius for local search.
• The phrase and scan date were associated with different observed cell shares in these examples. Cooking It had identical shares for three phrases in the latest scans, while CH Developments had different shares for two phrases. These examples do not isolate wording effects; repeat-run shares also changed, sometimes over a short interval.

Method

We analysed Google Maps grid scans for six UK businesses, using DataForSEO Google Maps SERP data. Each grid point used one live search call on mobile. We matched the business by its Google CID and recorded its rank to depth 20. A rank absent from the first 20 results is recorded as “not in top 20”. Top-three share of cells is SoLV, or share of local voice.

The scans ran from 2026-09-02 to 2026-10-15. Across all UK runs, grid sizes were 3-by-3, 5-by-5, 7-by-7 or 9-by-9. The latest-per-business-keyword export includes grids of 25, 49 or 81 points, arranged as 5-by-5, 7-by-7 or 9-by-9. Spacing across all UK runs ranged from 0.4 to 3 km; in the latest-per-business-keyword export it ranged from 0.4 to 2 km. For distance decay, each point is grouped by its straight-line distance from the grid centre. The latest-run view uses the most recent scan for each business and keyword; the all-run view pools every scan. The two views answer different questions and should not be mixed.

Findings

Across all scans, 3.7% of sampled points placed the business in the top three, 6.8% placed it in the top ten, and 89.0% did not show it in the top 20. This pooled summary combines every run and point, so it is best read as context for the six-business dataset rather than as a typical business result.

Distance decay: a strong centre, a smaller footprint

A business can be prominent at the centre of a scan and still have a limited surrounding footprint. In our latest-run view, 91.9% of centre pins were in the top ten. Across sampled points from zero to one kilometre, that share was 26.7%; from one to two kilometres it was 13.6%; and from two to three kilometres it was 7.4%. The top-three share also fell with distance: 12.0% in the zero-to-one-kilometre band, 6.1% in one to two kilometres and 4.1% in two to three kilometres. No sampled points in the three-to-five-kilometre or five-kilometre-plus bands ranked in the top ten in this latest-run view.

That pattern is visible when all runs are pooled too, though the exact percentages change. Across all scans, the top-ten share was 22.7% in the zero-to-one-kilometre band, 6.7% from one to two kilometres and 2.8% from two to three kilometres. Pooling multiple scans gives more observations, but it also gives more weight to business-keyword combinations scanned more often. The latest-run view uses the most recent available scan for each combination; neither is a count of unique searchers or customers.

Treat the grid as a sampled picture, not a hard boundary. A point outside the top 20 does not prove that no person nearby could see the business. Grid spacing, grid size, the chosen centre and the query all shape what the scan captures.

A strong pin is not the same as broad coverage

The centre result can overstate how visible a business is across its surrounding grid. In the latest scans, 52 of 62 business-keyword combinations had a centre pin ranked in positions one to three. Yet, across those combinations, the mean top-three share was 7.3% of grid cells and the mean top-ten share was 12.2%. Five combinations had no centre result in the top 20, and their mean top-three and top-ten shares across cells were both zero.

This is why a single rank check can be misleading. It answers “where does this listing appear at this point for this phrase?” It does not answer “how far does that visibility extend?” A grid gives a more useful local view, but it still samples selected places and queries rather than every possible search.

Near-me, city and bare service terms

The scans do not establish a general advantage for “near me” or city-modified phrases. They do show that phrase choice can change the map substantially, even for one business. For CH Developments in Redditch, “builders near me” had a top-ten share of 48.0% of cells, while “builders redditch” had 60.0%. These are two individual keyword scans, not a controlled test of wording.

Cooking It had the same 30.6% top-three and top-ten shares for “cookery school bristol”, “cooking classes bristol” and “cooking classes near me”. The available examples do not provide a like-for-like test of bare service terms against modified phrases. The different locations and scan setups matter, so these comparisons are illustrations, not evidence that one wording wins in every market.

The practical lesson is to track the phrases customers actually use, including relevant service-plus-place and near-me versions. Do not assume that a broad service phrase, a location phrase and a specific service phrase share the same footprint.

Repeat runs: movement is part of the picture

Rank coverage was not fixed between scans. For Cooking It’s “cooking classes bristol”, the top-ten share changed from 22.2% on 2026-09-02 to 30.6% on 2026-09-23. Nata Beauty’s “lip blush” moved from 28.6% to 32.7% between 2026-09-06 and 2026-10-15. A multi-branch beauty and aesthetics group in Bristol had “electrolysis” top-ten cell share change from 6.1% to 12.0% between 2026-10-12 and 2026-10-15.

Those examples show why a map rank is a snapshot. A change between two runs is not, by itself, proof of a lasting improvement or decline. Search conditions, timing and the scan itself can all affect what one observation captures. Repeat the same query and grid over time, and look for a pattern before changing strategy.

What to do about it

1. Track a set of useful phrases. Include core services, specific services, and relevant city or “near me” wording. Keep the set stable enough to compare over time.
2. Measure the map, not only the pin. Record centre position alongside top-three and top-ten cell shares, with distance bands that suit the service area.
3. Keep comparisons like for like. Use the same query, grid centre, grid size, spacing, device context and scan depth when comparing periods. Keep a record of any unavoidable setup change, because it can make two maps less comparable.
4. Repeat before reacting. Run scans on more than one date and distinguish a one-off shift from a sustained change.
5. Use the pattern to guide local work. If visibility is concentrated near the centre, review the relevance and completeness of the business profile and the local signals for the services and places that matter. Treat this as a diagnostic prompt, not a promise that any single change will expand a ranking radius.

Limitations

The sample combines different industries, locations, grid sizes and spacings. A point count is not a count of people, searches or impressions. The pooled all-run percentages also reflect repeated scans of some combinations. The keyword examples are descriptive; they do not isolate the effect of adding a city name or “near me”. The scans report rank to depth 20, so “not in top 20” means absent from that observed result depth, not absent from every possible result.

Data and how to cite this study

The scan data behind every table is open: latest scan per business and keyword, distance bands, latest scans, distance bands, all scans and repeat scans. Two client names are replaced by descriptions in these files. For wider context on local and AI-search evidence, see our SEO, GEO and AEO statistics roundup and AI automation statistics for the UK.

Cite as: Aristral (2026), "How far does a Google Maps ranking reach?", aristral.com/blog/google-maps-ranking-radius-uk-geo-grid-study, scans run 2026-09-02 to 2026-10-15.