A client emails: "We're #1 on Google Maps — my wife searched from the house." Two days later the same client emails: "We've dropped off Google completely — my sales rep can't find us." Both are true. Neither tells you anything about the business.
That contradiction is the entire argument for grid-based map pack rank tracking. A single ranking number for a local query is not a measurement. It's one sample from a field that changes every few hundred meters, and whoever takes the sample decides the answer.
What "we rank #3 in the map pack" actually means
Google's local pack is built from three broad inputs — relevance, prominence, and distance — and distance does far more work than most agency clients assume. When someone searches "emergency plumber," Google resolves their physical coordinates first and builds a result set around that point.
So "we rank #3" is shorthand for: on one device, at one set of coordinates, at one moment, for one exact query string. Change any of those four variables and the number moves.
Here's a real pattern from a single-location dental practice in a mid-size metro. For the query "dentist near me":
- At the office address: position 1
- 0.8 miles north: position 3
- 2.1 miles north: position 11
- 2.4 miles south, across a river: not in the top 20
Every one of those is a legitimate "ranking." A rank tracker configured to check from the city centroid would have reported 11 and called it a bad month. A tracker checking from the business address would have reported 1 and called it a win. The practice did not change.
The four variables a single number hides
- Proximity. The dominant factor. Rankings decay outward from your pin, and the decay curve is the actual thing you're trying to improve.
- Competitor density. A grid point sitting in a cluster of twelve competing med spas behaves nothing like a grid point in a residential pocket with none.
- Query variant. "Dentist," "dentist near me," "emergency dentist," and "cosmetic dentist" return visibly different packs from the same coordinates. One number usually represents one phrase.
- Volatility. Local packs reshuffle. The same point checked Tuesday and Thursday can differ by two positions with zero cause worth investigating.
What a rank grid is, and what it actually measures
A geo-grid runs the same query from a matrix of simulated coordinates — say 7 across by 7 down, spaced 0.5 miles apart — and records the business's position at each point. You get 49 data points instead of one, rendered as a heat map: green near the top, yellow mid, red or grey where you don't appear.
Three numbers come out of it, and agencies should know the difference:
- AGR (Average Grid Rank) — the mean position across all points, usually counting "not found" as 21+. Good trend KPI, terrible standalone number because it averages away the shape.
- ATGR (Average Total Grid Rank) — a variant that treats unranked points more harshly, which punishes coverage gaps. Useful when the goal is expanding reach.
- SoLV (Share of Local Voice) — the percentage of grid points where you appear in the top 3. This is the one that maps most cleanly to leads, because positions 4–20 in a map pack generate almost nothing without a click to "more places."
A business can hold an AGR of 6.4 with 18% SoLV, or an AGR of 6.4 with 41% SoLV. The second business is making money. Same average, completely different reality — which is why AGR alone recreates the same problem as a one-location rank.
Reading a grid: four shapes and what each one tells you
Tight green bullseye, fast fade
Top-3 within half a mile, gone by 1.5 miles. The profile is relevant and the category is right; prominence is weak. Fix list: review velocity, review text that mentions services and neighborhoods, citation consistency, and local links from genuinely local sources. This is the most common shape for a healthy small business that hasn't done much off-profile work.
Lopsided green
Strong to the west, dead to the east. Usually competitor density, not your fault. Check which businesses hold the pack at the dead points — if the same two competitors own every eastern grid point and both are physically closer, you're fighting geography. The honest play is service-area content and paid coverage for that zone, not promising a green grid.
Uniform yellow everywhere
Positions 7–14 across the whole grid with no hot core. Almost always a relevance problem: wrong primary category, thin services list, or a profile that doesn't semantically match the query. Check the primary category of the top 3 at several points. If they're all "Personal Injury Attorney" and the client is set to "Law Firm," that's the whole story.
Checkerboard
Green and red alternating with no spatial logic. Either you're sampling during a local pack shuffle, or the grid spacing is too fine for the market and you're reading noise. Re-run in 48 hours before touching anything.
None of these diagnoses are available from a single number. That's the point.
Choosing grid size and spacing without wasting credits
Grid scans are the most expensive thing in most local SEO tool stacks, so configuration matters more than people admit.
- Urban storefront (restaurant, salon, dental): 5x5 or 7x7 at 0.3–0.5 mile spacing. Total coverage of 1.5–3.5 miles. Anything wider is measuring people who will never visit.
- Suburban service business (HVAC, roofing, plumbing): 7x7 or 9x9 at 1–1.5 mile spacing. Covers a 7–12 mile radius, which is where the trucks actually go.
- Multi-city service-area business: separate smaller grids centered on each target city, not one giant grid. A 13x13 grid at 3 mile spacing looks impressive and tells you nothing actionable.
On frequency: weekly is enough for most retainers. Daily scans on a 7x7 grid for twelve keywords across five clients burns budget producing noise. Run weekly for trend, and run an ad-hoc scan when something changes — a category edit, a new location, a suspected suspension, or a client panic email.
On keyword selection: three to five money terms per location beats twenty. Pick one head term ("dentist"), one high-intent modifier ("emergency dentist"), and one or two service terms that carry real revenue ("dental implants"). Track them all on the same grid geometry so comparisons are valid.
Reporting grids to clients without creating new confusion
A heat map is persuasive precisely because it's visual — which means it's easy to misuse. Three rules that have saved a lot of awkward calls:
Lead with SoLV, not AGR. "You're in the top 3 for 44% of your service area, up from 29% in March" is a sentence a business owner understands. "Your AGR improved from 8.1 to 6.9" is not.
Always show the before/after grids side by side. One grid is a photograph; two grids are a story. Clients retain the image long after they forget the number.
Annotate the geography. Mark the client's address and the top two competitors' addresses on the grid. The moment an owner sees that the red zone contains a competitor who is physically two miles closer to it, the conversation shifts from "why aren't we #1" to "what do we do about that zone." Agencies running local SEO software for agencies that ties grid scans directly into the client-facing report cut a lot of explaining time here — the grid, the review feed, and the trend line arrive in one document instead of three screenshots pasted into email.
Where grid tracking fits with everything else
Grids answer "where am I visible." They

Nick Quirk

