How to build it
How to build a company that turns building permits into contractor leads
Every construction project in the United States starts with a permit, and almost every permit is a public record the day it is filed. Somebody just told the government they are about to spend money on their house.
Most contractors never see that. They buy leads from marketplaces where the same lead is sold to four competitors, or they wait for referrals. The business is the gap between those two facts.
Why a permit is a better signal than a lead form
A lead form tells you somebody filled in a form. A permit tells you somebody has committed.
The permit record usually carries four things that matter:
- Address — so you can filter to a service radius.
- Work type — reroof, addition, service upgrade, pool, solar, demolition.
- Declared valuation — the owner's own estimate of job size, which separates a $4,000 job from a $200,000 one.
- Contractor of record, or its absence. A permit pulled by the homeowner rather than a licensed contractor is a different, and often better, prospect.
The second-order signal is where most of the value sits. A permit for a kitchen remodel is not just a lead for a remodeler — it is a forward indicator for the electrician, the plumber, the flooring installer, the countertop fabricator and the skip-hire company, in roughly that order, over the following eight to twelve weeks. One record, sold to five different trades, none of whom compete with each other.
Where the data actually comes from
There is no single national source, and that fragmentation is the moat. Anyone can get one city. Getting forty is work, and work is what you are being paid for.
In practice permits appear in three shapes:
- Open data portals. Larger cities publish permits as a queryable dataset, often with a documented JSON endpoint and a stable schema. These are the easy ones and should be your first ten.
- Permit-system web portals. Many jurisdictions run a hosted permitting system with a public search page but no export. The data is public; the convenience is not. This is the bulk of the country.
- Records request. Smaller jurisdictions may have nothing online at all. A standing request can sometimes produce a recurring export.
Start with one metro and one trade. A tool covering every county badly is worth less than one covering four counties precisely, and the narrow version is the one you can sell this month.
Who pays, and what they are actually buying
The buyer is not "contractors" in general. Pick one:
- Roofers, on storm and reroof permits — the highest-urgency category, because roofs get replaced on a deadline set by weather rather than taste.
- Solar and electrical, on service-upgrade and panel permits.
- Suppliers and yards, who want volume in a territory rather than individual jobs.
- Adjacent trades, on the second-order signal above — the ones nobody is currently selling to.
What they are buying is not the raw record. It is the record, filtered to their trade and radius, delivered before their competitor sees it, with enough context to make the first call. If your output requires the buyer to interpret a municipal permit code, you have shipped a database and not a product.
What the incumbents already do, and where the opening is
Be honest about this, because your buyer already knows. Construction Monitor has sold permit-based construction leads since 1989 and covers all fifty states. BuildZoom's database spans roughly 175 million building permits and around 3.7 million contractor licence records, drawn from thousands of jurisdictions. These are serious companies with national coverage, and pretending otherwise will cost you the first conversation.
They are also built for national and enterprise buyers. They price accordingly, they onboard accordingly, and their products assume a sales team. The one-truck roofer working three ZIP codes is not their customer and is not served well by them.
The opening is hyper-local, single-trade, and cheap enough to buy without a meeting. That is not a gap in the data. It is a gap in who the data has been packaged for.
The pricing shapes that work
- Per-lead. Easy to sell, hard to hold. The buyer counts outcomes and churns the moment a month goes badly, whether or not you caused it.
- Territory subscription. A flat monthly fee for one trade in a defined set of ZIP codes or counties. Predictable, and the exclusivity is the product — a territory sold once is worth far more than a lead sold four times.
- Hybrid. A low subscription for access plus a fee on qualified outcomes.
Territory exclusivity is the strongest position, because it is the one thing the large shared-lead marketplaces structurally cannot offer. Their model depends on selling the same lead repeatedly.
The operational parts nobody writes about
- Freshness beats completeness. A permit delivered the morning it posts is worth several times the same permit a week later. Your delivery schedule is your product.
- Deduplicate against revisions. Permits get amended, reissued and closed. Sending the same address three times destroys trust faster than missing one.
- Suppress the already-contracted. If a licensed contractor is already named on the permit, that job is usually gone. Filtering those out lowers volume and raises perceived quality, which is the trade you want.
- Deliver where the buyer already is. A daily email a roofer reads in his truck at 6am beats a dashboard he has to remember to log into.
The first ten customers
Do not build a marketing funnel. There are a finite number of roofers in a county and you can contact all of them.
- Pick the trade and the county.
- Pull the last ninety days of relevant permits.
- Find the ten contractors who appear most often as contractor of record. They are already winning this work, which means they already believe in the signal.
- Send each one last week's permits for their trade in their radius. Free, no pitch, just the list.
- The ones who reply are your customers. What you sell from that point is continuation.
The demonstration is the sales pitch, and it costs nothing but data you already pulled.
What this is, minus the romance
It is a data-plumbing business. The work is ingestion, normalisation across inconsistent municipal schemas, deduplication, and reliable delivery every single morning. The insight is available to anyone. The execution is boring and continuous, which is precisely why the opportunity persists.
That is also exactly the kind of work that should not require a person.
Or don't build it
acoco is a platform that builds businesses, funds them, and runs them. You bring an idea; it builds the business, puts capital behind it, and operates it — the site, the outreach, the customers. You own it.
Which means the playbook above is one prompt:
Questions people actually ask
Is building permit data public?
In most United States jurisdictions, yes — building permits are public records, and many cities publish them through open data portals with documented APIs. Availability and format vary by jurisdiction: some publish a queryable dataset, some only a search page, and some nothing online at all.
Can you resell public permit data?
The records are public. What you sell is the access layer — filtering, enrichment, deduplication and delivery — rather than the raw record itself. Check the specific terms of any portal you pull from, since some open data licences carry attribution or redistribution conditions.
How much do contractors pay for leads?
It varies widely by trade and market. The structural point is that shared-lead marketplaces sell the same lead to several contractors at once, which is why exclusive territory access commands a different price than per-lead pricing.
What is the hardest part of a permit-lead business?
Not the data. It is reliable daily delivery across jurisdictions whose schemas disagree with each other, every morning, indefinitely.