Grid search

Also called: Tiling, Grid scraping, Tiled search.

One Google Maps search shows a limited number of results, often cited as about 120, and the official Places API Text Search returns at most 60 results per query. In a big city, that's a fraction of the businesses in a trade. Grid search covers the area in pieces so each search is local to a small tile.

The method: divide the area into tiles sized to the density of businesses, run every keyword in every tile, and merge the results. Dense downtown areas need smaller tiles than suburbs. Results overlap between neighboring tiles, so you need deduplication on the Place ID.

The trade-off is cost and time. A city might need hundreds or thousands of searches per keyword. Combining it with a good keyword list (category names plus synonyms) matters as much as the tiles: grid search finds more of the businesses that match a keyword, and keywords decide which businesses can match at all.

Coverage is never guaranteed. It depends on the keywords, how Google files each business, and how fine the grid is. Measure it by comparing against a plain search.

Example

A single search for "coffee shop" in Portland returned 230 businesses in our measured run; a tiled search of the same city and keyword found 457. That's one city and one keyword, so treat it as an illustration, not a rule.

In Leadraft

Leadraft tiles every search past Google's result cap, runs every keyword in every tile, and dedupes on the Place ID. The coverage methodology shows the measured runs and their caveats.