Skip to main content

U.S. Location Intelligence

Decide where to open next. Then check the arithmetic.

GISmetric scores every comparable U.S. market on six factors, then shows the measurement behind each one. Every input names a public dataset and the release it came from, so a number you disagree with is a number you can take apart.

Free plan, no card required. See what each plan includes.

gismetric.com/analysis

Showing an illustrative ranking weighted for Coffee Shop.

Weighted for Coffee Shop

5 of 1,486 tract areas

1Clarendon, VA71.4
2Silver Spring, MD69.3
3Columbia Heights, DC68.0
4Bethesda, MD66.0
5Del Ray, VA64.7

Map: real census tracts, shaded by their published population-density percentile. Ranking: illustrative factor scores, ordered by the same weighted sum the engine uses and the real published weights for this business type.

Business types
18
Scoring factors
6
Score range
0–100
Records loaded
13,357,179
Census and Overture, as loaded

Coverage

Every comparable area in the country, not the ones you thought of.

An analysis does not rank a shortlist. It ranks every area of the same kind inside the frame you chose, which is why the answer can be somewhere you had not considered.

0–100Shaded by population density — each county’s percentile among the 3,222 scored, from ACS population over TIGER land area. One of the six factors, not the opportunity score: that one depends on what you are opening and cannot exist before you say.

Resolution

You choose how finely the ground is cut.

The same city, divided three ways and shaded by the same measurement. A county answers a question about a market; a census tract answers one about a corner. Finer is not automatically better — a block group carries wider survey margins, and the product says so on every run at that level.

All three are population density percentiles against every comparable area in the country, which is why the same ground changes colour as it is cut more finely: a dense corner of an average county is still a dense corner.

County

A whole market. Wide margins on nothing, and no answer about a street.

Census tract

The default. Roughly a neighbourhood, and the level most estimates are reliable at.

Block group

A few blocks. Available, flagged for reliability, and never the quiet default.

The decision

A lease is five years. Most of them are signed on a feeling.

The tools that answer this question properly are sold to chains with a real estate department, priced by quotation, and built on panel data nobody outside the vendor can inspect. If you are opening your second location, or your first, you get a map and an opinion.

Ask for a price

Every established location-intelligence platform answers with a contact form. Nobody publishes what it costs, which means nobody publishes it to you.

Take the number on faith

Scores built on a proprietary movement panel cannot show their arithmetic, because the arithmetic is the product. You are told the answer and asked to trust it.

Get sold to somebody else

The models are trained on chains with fifty sites and a marketing budget. A single independent business is not the customer they were designed around.

How it works

Three steps, and the third is the one the rest of this market skips.

  1. Say what you are opening, and where to look

    Pick one of the business types and an area — a named place or a point and a radius. Optionally say who your customers are: an age range, an income band.

  2. Every comparable area gets scored

    Not the handful you thought of. Every tract, block group or ZIP area of the same kind in the frame, ranked against each other on the six factors.

  3. Open any number and take it apart

    Each score decomposes into the measurements under it, each measurement into its position among peers, each peer group into the dataset and release it came from.

What goes into a score

Six factors, each a position among comparable areas.

Not a mark out of a hundred. Each factor is where an area sits among the areas it is being compared with, then weighted by business type — and the weights are visible and yours to move, because a coffee shop and a dental practice do not care about the same things.

What you get

A ranking, a comparison, and a document you can hand to someone.

A ranked market

Every comparable area scored and ordered, on a choropleth you can recolour by any single factor, with the ranking beside it. Move a weight and both update without a request.

What is measured

Two to four areas, side by side

The gap between two areas decomposed into the factors that produced it, exactly — the parts sum to the difference, and the residual is printed to prove it.

How the gap is split

A standing report

One area in depth: the findings behind its score, the areas it was measured against, every source with its release date, and what the figures cannot tell you. Downloadable as a PDF.

What each plan includes

Why this one

Everyone else says trust us. This one says check it.

Every figure traces back to a named source

Each factor is built from published federal data, and every input carries the dataset and the release it was loaded from. An analysis records the weighting it applied and a report quotes the releases behind it, so a number you disagree with is one you can take apart rather than one you have to accept. Where the data runs out, it says so.

What goes into a score

No sentence in a report was written by a language model

Findings are composed by a fixed procedure out of the figures themselves, so no report can contain a number that is not in the data or a claim the data does not support. In a market that spent the last two years attaching a chatbot to everything, this is the harder promise and the one worth making.

It says what it cannot tell you

Fourteen stated limitations, four of which appear on every report and ten when they apply. A factor that could not be computed for an area is dropped and declared rather than filled with a zero, because a zero is a measurement and absence is not.

The price is on the website

Three plans with monthly allowances, counted from the analyses and reports themselves rather than a separate tally, so the usage you are shown is the usage you can list. Reopening or reusing work is free. Nothing is deleted when a plan ends.

See the plans

Provenance

What is loaded so far

Read from the database on every refresh, not from a number typed into this page. If an ingest is half finished, this table says so.

The currency column is the result of asking each publisher whether it has released anything newer. Nobody sends a notification when a Census file appears, so this is a question that gets asked daily and recorded — which is why it can say “not confirmed” rather than quietly implying everything is fine.

Datasets currently loaded, with their currency, record count and licence
SourceCurrencyRecords
Census geographyStates, counties, tracts, block groups and ZIP areas, with the tiles to draw themNewer release available398,112
DemographicsAmerican Community Survey estimates, carrying their published margins of errorCurrent12,208,183
Establishment benchmarksCounty Business Patterns: how many businesses a market this size normally supportsCurrent210,746
Competitor locationsOpenly licensed points of interest, matched to sectors, chains split from independentsCurrent540,138

Score a market in about a minute.

Pick a business type and somewhere to look. The free plan is enough to score a real market and read the findings behind it, and it does not ask for a card.