Map Extract

A Google Maps grid search covers a city by dividing it into small cells and running the same search in each one, because a single search only shows its top results (about 120 on the website, 60 through the Places API). Start with cells sized to the area's density, split any cell whose search comes back close to the cap, and stop splitting when a finer grid adds mostly duplicates. Then check that each result is really inside your target area and merge duplicates on place ID.

Grid search is one of the coverage methods in the Google Maps lead generation guide. If you have not met the result cap yet, read Google Maps only shows 120 results first.

Grid, town list or postcode list

MethodSet-upCoverageDuplicatesBest for
Town and city listMinutesUneven: big cities under-coveredLowRegions of separate towns
Postcode districtsMinutes, with a district listGood in cities, coarse in the countrysideModerateUK city and suburban campaigns
Uniform gridNeeds coordinates and a cell sizeEven across the areaHigh in dense cellsComplete regional coverage
Adaptive grid (split cells that hit the cap)Most effortBestControlledLarge, mixed-density areas

Postcode districts are a good default in the UK. Switch to a grid when districts are too big (rural Scotland, Wales, the Pennines), when towns blur into each other, or when the client wants every business in a region and will ask how you know none were missed.

How cell size and zoom relate

On a web map, each zoom level halves the distance shown per pixel. For the Web Mercator projection that Google, Bing and most web maps use, ground resolution is about 156,543 metres per pixel at zoom 0 at the equator, multiplied by the cosine of the latitude and divided by 2 for each zoom level.

For a browser map panel about 1,000 pixels wide at the latitude of Leeds or Manchester (about 53.5° N):

ZoomMetres per pixelWidth shown in 1,000 px
12about 22.6about 23 km
13about 11.3about 11 km
14about 5.6about 6 km
15about 2.8about 3 km
16about 1.4about 1.4 km

Further south the same zoom shows slightly more ground. The rule of thumb: pick a zoom whose visible width roughly matches the cell. Searching a 2 km cell at zoom 12 makes the map show a 23 km area, and the ranking will reflect that wider area rather than the cell.

Google does not strictly restrict website results to the visible map, so results also come from just outside a cell. That is why neighbouring cells overlap. If you use the Places API instead, Text Search (New) accepts a locationRestriction, and Google's documentation says "results outside the specified area are not returned", which gives cleaner cells but a lower cap of 60.

Choosing cell size by density

Start from what you expect, then let the counts correct you.

Area typeStarting cellTypical zoom
City centre, high street clusters1 to 2 km15 to 16
Inner suburbs3 to 5 km14
Outer suburbs and market towns5 to 10 km13
Rural areas20 to 25 km11 to 12

These are starting points, not rules. A rare category, such as a specialist engineering firm, can use large cells even in a city. A common one, such as cafes or hairdressers, may need the smallest cells everywhere.

The adaptive method: split only cells that hit the cap

A uniform fine grid wastes searches on empty fields and reservoirs. An adaptive grid spends them where businesses are:

  1. Cover the area with a coarse grid suited to the countryside.
  2. Run the search in every cell and record the result count.
  3. Any cell that returned close to the cap (about 120 on the website, 60 through the API) is incomplete. Split it into four quarter cells.
  4. Run those, and split again where needed.
  5. Skip cells that are entirely sea, moorland or water after the first pass returns nothing.
  6. Stop when no cell is near the cap, or when a new round adds very few new businesses.

This is the same "split where saturated" idea that tool makers describe for getting past the cap, for example in Apify's article on Places limits, applied with your own counts.

Duplicate rate expectations (fictional run)

The figures below are invented to show the pattern you should expect, not measured results.

A fictional campaign searches "hairdresser" across a city of about 15 km by 15 km.

GridCellsRows returnedDistinct businessesNew vs previous gridDuplicate share
1 search (whole city)1120120n/a0%
5 km cells9900410+29054%
2.5 km cells363,100620+21080%
1.25 km cells (centre only, 12 cells split)48 more2,300655+3598% of the new rows

Read the last column carefully. Duplicate share rises with every split, which is normal. The decision metric is the "new" column: the 1.25 km round cost 48 searches for 35 new businesses. Whether that is worth it depends on your margins, but it is a signal that coverage is close to what the platform will show.

How to measure it yourself:

  • Duplicate share = 1 − distinct businesses ÷ rows returned.
  • New per search = new distinct businesses in this round ÷ searches in this round.

When new per search drops below one or two, stop splitting.

Validate the geography

A grid can still return the wrong places. Before merging:

  • Drop results whose coordinates or postcode fall outside the target area, unless you deliberately want neighbouring towns.
  • Flag service-area businesses with hidden addresses rather than dropping them; their coordinates may be approximate.
  • Check the category: a "hairdresser" grid will return wholesalers and training academies too.
  • Watch for searches that localise to your own network rather than the cell; a VPN in another city can skew every result.
  • Look at cells that returned zero. Zero in a city centre is more likely a blocked or failed search than an absence of businesses.

Merge and dedupe

Merge in order of trust: place ID, then registrable website domain, then phone in +44 form, then name plus postcode. Google's place ID documentation warns that the same place can have more than one place ID and that IDs can change, so the later keys still matter. Branch handling and fuzzy name matching are in deduplicating business leads.

If you also run the grid on Bing or Apple Maps, merge across platforms the same way. Each platform ranks differently and holds some businesses the others lack; the maps scraper page covers all three.

Pace the run

A city grid can mean hundreds of searches. Run them at a steady, unhurried pace. If a CAPTCHA or "unusual traffic" page appears, stop and slow down rather than continuing; repeated challenges mean the platform wants less automated traffic from you. Collecting Maps data with automated tools may breach a platform's terms, so check the rules first.

How Map Extract helps

  • Area grid. The Google Maps scraper can split a map viewport into a grid of cells, 25 km by default and adjustable, up to 5,000 cells, with a zoom per cell. The 25 km default suits regional and rural coverage; reduce it for cities.
  • Location picker. Choose a country, regions and target areas from the bundled location catalogue, or paste your own place list, including coordinates and a radius per place. Each place becomes a job paired with each keyword.
  • Saved counts per job. Listing counts per job show which cells came back near the cap and need a finer grid.
  • Geography validation. If fewer than 60% of sampled results fall inside the place's radius (50 km by default) or name the locality in their address, the job is flagged as wrong geography or insufficient sample instead of being saved silently. Geographic features such as lakes and towns are rejected as listings.
  • Network location. An option lets Google use your machine's own IP location, and Bing and Apple searches are centred on the place's coordinates when known, via the Bing Maps scraper and Apple Maps scraper.

What it does not do: it does not pick the right cell size for you, it cannot return more than the platform shows for each search, and it does not guarantee complete coverage. A CAPTCHA pauses work for manual handling. Third-party platforms may restrict data collection or automated actions. Before running a workflow, check the platform's rules and make sure you have permission for the actions you configure.

Public business and social results the app collects are uploaded automatically to Map Extract's servers while it is licensed and running; the privacy notice explains this.

Key takeaways

  • Use postcode districts for most UK city work and a grid where districts are too large or towns merge.
  • Match zoom to cell size; a small cell searched at a wide zoom behaves like a big cell.
  • Split only cells whose search comes back near the cap.
  • Expect duplicate share to rise; judge each round by new businesses per search.
  • Validate geography and dedupe on place ID, domain and phone.

Frequently asked questions

What is a Google Maps grid search?

A grid search divides a city or region into a grid of small map cells and runs the same search once in each cell. Because each search only shows its top results, about 120 on the Google Maps website, many small searches together reach businesses that one city-wide search hides. The results are then merged and duplicates removed, since neighbouring cells return some of the same businesses.

What cell size should I use for a Google Maps grid?

Match cell size to business density. Dense city centres may need cells of one or two kilometres, suburbs five or so, and rural areas twenty or more. The practical test is the result count: if a cell returns close to the cap, split it into four smaller cells and search again; if it returns well under the cap, it is small enough.

Is a grid better than searching by postcode?

Not always. UK postcode districts are quick to set up, easy to explain and follow how people describe areas, so they suit most city campaigns. A grid gives more even coverage where towns merge, where postcode districts are very large, or where you need every business in a region. Many campaigns use postcodes first and a grid only for the dense districts that hit the cap.

Why do I get so many duplicates from a grid search?

Google does not strictly limit results to the visible map, so a search in one cell also returns well-ranked businesses just outside it. Neighbouring cells therefore overlap, and smaller cells overlap more. Duplicates are expected; remove them with the place ID first, then website domain and phone, and track how many genuinely new businesses each finer grid adds.

Sources

  1. Text Search (New), Places API, Google for Developers
  2. Place IDs, Google Maps Platform documentation
  3. Google Places API limits, Apify blog

Written by the Map Extract editorial team. Found an error? Email support@mapextract.uk and we will check and correct it. How we research guides: editorial policy.

Google Maps Lead Generation: The Practical Guide

How to find local business leads on Google Maps: pick terms and areas, beat the visible result cap, enrich websites, verify emails and stay within the rules.

13 min read

How to Find Businesses Without a Website

A method for web designers: find local businesses with no website on Google Maps, weed out false positives, add dead and parked sites, then pitch lawfully.

8 min read

See it working on your own searches.

Start with one keyword, one location and a small limit. Review the records, then decide on a licence.

Top