What is the Opportunity Score?
A percentile against all 935 US metropolitan and micropolitan markets for one service line — not a rating out of 100 in the school sense. It combines seven components: market size, housing age, ability to pay, growth, storm exposure, climate load and how under-served the market is. The weights are published in full on the methodology page.
It is identical for every reader of a given market and service line. Your own numbers never move it — they produce a separate Fit score instead, so the market read stays comparable across metros.
Where does the data come from?
Four public federal sources, all public domain: the Census Bureau's American Community Survey, the Census Building Permits Survey, the Bureau of Labor Statistics' Quarterly Census of Employment and Wages, and NOAA's Storm Events Database and Climate Normals. Every figure on the site traces to one of them, and each report lists which measures came from which publisher and in what vintage.
Nothing is purchased from a data broker, scraped from a competitor, or estimated by a model where a measurement exists.
How often is it updated?
Twice a year in practice, following the publishers rather than a calendar of our own: most sources land in February and June. The vintage of each measure is printed in every report, and the site says “data as of” rather than implying it is live.
Is anything here written by AI?
No. Every sentence of analysis on the site and in the reports is computed from thresholds in code — the same inputs always produce the same words. There is a check that reads every figure back out of the generated prose and refuses to publish anything containing a number that is not in the underlying fact sheet.
That is a deliberate constraint rather than a limitation we are apologising for: an invented statistic is indistinguishable from a real one, and one of them would make every other figure negotiable.
What is free and what costs money?
Reading is free and always has been: the map, every market file, the comparison, the market read against your own numbers, the score, the demand model, the county detail, the trends and the national operator footprint. No account, no card.
One thing costs money. A report built around your company — your ticket, your crews, your base, your fit score — is a one-time purchase. Nothing that was free was moved behind it.
What does a purchase actually buy?
One market and one service line, permanently. Dallas roofing and Dallas windows are two separate purchases; the same market re-opened next March is not.
The presentation deck, the printable report, the workbook, the CSV and an emailed copy are five ways of holding one report and cost one credit between them — including re-downloading it later. A failed render or a bounced email never costs you anything.
Why does the model overstate a sprawling metro?
Because it treats a whole metropolitan area as one service area, and a real branch serves a drive time. In Riverside or Phoenix that gap is large. The county detail in every report is the honest partial answer — it shows where inside the metro the aged housing actually sits — but it is not a drive-time model and does not pretend to be.
What can the model not see?
Four things, and they are listed in every report rather than buried here. What competitors pay for a lead, and how hard they bid. Insurance behaviour after a storm, which moves roofing demand more than anything we measure. Whether you can hire in a given metro. And anything about a specific neighbourhood.
Advertising cost per lead lives inside Google and Meta and is not obtainable honestly, so it stays on that list permanently.
Do you sell my data?
No. There is no advertising on this site, no tracking pixel sold to anyone, and nothing about you is passed to a data broker. What is collected and why is set out in the privacy policy.
Can I get a refund?
Ask. A report is a file that cannot be un-downloaded, so there is no automatic process — but if the document was not what the page described, say so and it will be put right.