A local rank tracker tells you that you rank #3 for "emergency plumber." A customer two miles north opens Maps and doesn't see you at all. Both things are true, and the reason is that local search results are computed per location — so a single number describing "your rank" is describing one point on a map you have not looked at.
What single-point tracking actually measures
Classic rank tracking sends one query from one place and records where you appear. That place is usually the geographic centre of the city, or wherever the tracker's proxy happens to sit.
That's fine for organic web results, which barely vary by a few miles. It falls apart for the map pack, where Google weighs proximity heavily. Move the query origin a mile and the three businesses shown can change completely.
So the number isn't wrong — it's just a sample of one, presented as though it described a territory.
How a geo-grid scan works
Instead of one query, a geo-grid scan runs the same keyword from a grid of real GPS coordinates laid over your service area:
- 3×3 for a tight radius — a single storefront serving a neighbourhood.
- 5×5 for a typical city service area.
- 7×7 for citywide or multi-suburb coverage.
Each point returns the map pack a customer standing at that point would see. Forty-nine points means forty-nine independent answers to "where do we rank," which you can then read as a map rather than a number.
The output is usually rendered as a coloured grid: green cells where you're in the top 3, yellow in the top 10, red where you're nowhere. The first time most businesses see theirs, the surprise isn't the average — it's the shape. Rankings almost always fall off in one direction faster than the others, and that direction is a competitor.
The three metrics, and which one to trust
Once you have a grid, you need to summarise it. Three numbers do the work:
| Metric | What it counts | What it hides |
|---|---|---|
| AGR (Average Grid Rank) | Average position across cells where you rank in the top 20 | Every cell where you don't rank at all |
| ATGR (Average Total Grid Rank) | The same average, with unranked cells counted as position 21 | Nothing — that's the point |
| SoLV (Share of Local Voice) | Percentage of cells where you're in the top 3 | Depth of ranking below position 3 |
The trap is AGR. Because it only averages the cells where you already rank, it improves when your weak areas get worse. Drop out of the top 20 in your five worst cells and your AGR goes up. Any metric that rewards losing coverage does not belong in a client report.
ATGR is the honest version. Unranked cells count as 21, so losing coverage costs you, and the number only rises when you actually expand.
SoLV is the one that tracks revenue most closely. The map pack shows three results before a customer scrolls, so top-3 coverage is roughly the fraction of your area where you're visible at all. If you report one number to a client, report SoLV; if you report two, add ATGR.
Reading the map, not the number
The real value of grid data isn't the summary — it's the diagnosis. A few patterns come up repeatedly:
Uniform mid-rank across the whole grid. Proximity isn't the constraint; profile strength is. Work on categories, services, and review velocity.
Strong centre, sharp falloff in all directions. This is normal proximity decay. Expanding it means more citations and more locally-relevant content, or a second location.
Strong everywhere except one lobe. A competitor owns that area. Pull the competitor grid for the same keyword and compare — usually they have more reviews, an older profile, or a physically closer address.
Patchy, no clear geography. Often a data problem rather than an SEO one: inconsistent NAP across directories making Google unsure where you are. Start with a citation audit.
How often to scan
Weekly is the right default for an active campaign. Daily produces noise — map-pack results fluctuate on their own, and you'll spend your time explaining variance rather than trends. Monthly is too slow to connect a change you made to a change in rank.
The exception is right after a significant change: a category edit, an address change, or a suspension recovery. Scan before and 7–10 days after, so the attribution is clean.
Getting a baseline
Geo-grid rank tracking in SeoNexus runs on a schedule and keeps history, so you can put a before-and-after grid next to each other. If you just want to see your current shape once, the free local check runs a scan around your business and shows the grid without an account.
Either way, look at the map before you look at the number.