DEV Community

OpenGMB
OpenGMB

Posted on

Geo-grid rank tracking for Google Maps on your own DataForSEO key

A business does not have one Google Maps ranking. It has a different rank on every street, because Maps results depend on where the searcher is standing. A dentist can be #1 two blocks from the clinic and nowhere three kilometres away. A single "we rank 4th" number hides all of that.

Geo-grid rank tracking is the usual fix: search the same keyword from a grid of points around the business and plot the rank at each point on a map. This post walks through how it works in OpenGMB, the open-source Google Business Profile manager we build at Synectus, and why the "bring your own DataForSEO key" part matters if you run it yourself.

What a scan does

You set up a tracker for a location:

  • Up to 5 keywords per tracker.
  • A grid from 3x3 to 13x13 points. 7x7 is the default.
  • A radius from 100 m to 50 km, measured from the centre to the edge of the grid.
  • Run it now, or on a schedule — daily, weekly or monthly. Every scan is kept in a history, so you can compare it with the one before.

Each point on the grid gets a rank from 1 to 20, coloured on a heatmap. The same ranks are also available as a table, with north at the top, for anyone who prefers numbers to colours.

Three numbers that summarise the grid

A heatmap is easy to look at but hard to compare month to month. So every scan also produces three numbers:

  • Average rank — the mean rank across the grid. A point where the business is not in the top 20 counts as 21, so gaps pull the average down instead of being ignored.
  • Top 3 share — the share of points where the business ranks 1, 2 or 3.
  • Found share — the share of points where the business shows up at all.

Found share is the one we watch first. If it is low, the business simply does not exist for searchers in part of its service area, and no amount of tuning the top-3 points will fix that.

Who owns the area?

Next to the grid, a table lists the top 25 businesses seen across all points, with your own business marked. For each one you get its address, rating and review count, how many points it appears at, its average rank and its best rank. In practice this is the fastest way to answer "who are we actually competing with here?" — it is often not the competitor the client names.

There is also a quick scan for a business you do not manage: paste its Google place id, pick a keyword and a grid, and run it once. It never repeats on a schedule. Agencies use it to scan a prospect before a pitch.

Why your own DataForSEO key

The rank data itself comes from DataForSEO's Google Maps results. In the self-hosted community edition you add your own DataForSEO key, either in the app (Rank tracking, then Usage and cost) or in .env. Scans on your own key are billed by DataForSEO at cost, with no markup.

Until a key is connected, heatmaps show sample data — and every heatmap says so with a "Sample data" badge, so nobody mistakes a sample for a real rank. The DataForSEO sandbox returns made-up data too, and it is labelled the same way. We like this rule: a demo screenshot that looks like real data is how bad decisions get made in client meetings.

For a sense of scale: a 7x7 grid for one keyword is 49 points per scan.

A simple way to use it

  1. Track the 3–5 keywords that bring calls, not vanity terms.
  2. Pick a radius that matches the real service area. A clinic that patients drive 15 minutes to needs a bigger grid than a café.
  3. Run monthly and compare Found share first, then Top 3 share.
  4. When a competitor dominates one side of the grid, look at what they have on that side — a second location, more reviews, a better-matched category — before changing anything.

Rank tracking sits next to the rest of the tool (reviews, posts, profile edits, performance reports), and like everything else in OpenGMB it is also available as an MCP tool, so an AI client can pull a scan or summarise a heatmap for you.

The details, limits and screenshots are on the OpenGMB rank tracking page. The code is MIT-licensed on GitHub at synectus/OpenGMB.

Disclosure: this post is from the OpenGMB team at Synectus.

Top comments (0)