Most lead lists answer the question “who exists?” Every business of a certain type in a certain area, with a phone number. That’s useful, but it’s not the question a salesperson actually has. The real question is “who is about to buy, and when?”
For some industries, the answer is sitting in public records. Nobody has to guess. A government agency already wrote it down, because spending public money means publishing what you’re spending it on.
I built a prospect list on spec for a manufacturer whose products go into new and renovated schools, and it turned out to be one of the cheapest and most useful builds I’ve done.
Find out who really decides
The obvious customer for a product that goes into a school is the school district. They pay for it. But they mostly don’t choose it.
When a school gets built or renovated, the architect writes the specifications, including the section that describes products like this one. That happens during the design phases, often one to three years before the project goes out to bid. By the time contractors are bidding, the spec either names your product or it names a competitor’s with “or approved equal” tacked on, and you’re fighting uphill.
So the most valuable thing a list can tell this company isn’t “here’s a school.” It’s “here’s a school whose architect is writing your part of the spec right now, and here’s the firm.”
Where the records live
States that help pay for school construction track every project they fund. I used four free sources:
- Massachusetts’ school building authority publishes a project report before most board meetings. It lists every active project with its phase, budget, bid dates, and the names of the architect, project manager, and construction manager. That one document did most of the work.
- Maine’s Department of Education publishes its approved projects and a ranked list of schools with documented needs that haven’t been funded yet.
- New Hampshire’s building aid ranked list.
- The federal school directory, which adds phone numbers, addresses, and enrollment for every school and district.
None of it was hard to read, but some of it was annoying to get. Two state websites block automated downloads, so a couple of files had to be pulled by hand, and one of the older lists was a scanned PDF.
What the AI did (not much, on purpose)
I wanted to keep this one cheap, and it didn’t need much AI. The only paid step was having Claude transcribe tables out of the PDFs, including the scanned one. That cost about $1.80, once.
Everything after that is plain code. Each project gets matched to its school in the federal directory, then scored with a simple weighted formula anyone can read:
- Fit (35%). Some kinds of schools need much more of the product than others. A high school usually needs more than an elementary school.
- Timing (30%). Projects in the design phases, where the spec is being written, score highest. Projects already under construction score low, because the product decision is mostly made.
- Size (20%). Bigger enrollment, bigger job.
- Project type (15%). New buildings beat renovations.
Every row also gets a plain next step tied to its phase. For a project in early design, it says to contact the named architect’s K-12 team now. For one that’s already under construction, it says to ask the construction manager whether the package for the product is still open, and if it isn’t, to offer an inspection and service agreement after the building opens.
- 01Public recordsState project reports and lists
- 02TranscribeClaude reads the PDFs, about $2
- 03MatchEach project tied to its school
- 04ScoreNeed, timing, size, project type
- 05Next stepWho to call, based on the phase
What came out
123 active school projects across Maine, New Hampshire, and Massachusetts. In 25 of them, the spec is being written now. 72 name the architect, so the person to call is already on the page. On top of that, there’s a watch list of 93 Maine schools with a documented need but no state funding yet. Some of those will pass local bonds and turn into projects.
Most of the active projects are in Massachusetts, simply because Massachusetts publishes the most detail. That’s worth knowing about public data: what you can find depends on what each agency decides to publish, not on where the business is.
The catch: public records go stale
Here’s the mistake I made. Some of Maine’s approved projects were years old. The list said “approved,” so I treated them as active. When I went back and checked each one, two schools had already been built and opened. Several others had moved on to construction. I caught it later than I should have.
Nothing in the record was wrong. It was just old. A record shows what was true on the day it was written. So my rule now: anything older than a year gets checked against a current source before it goes on a list someone will use. Usually that means a quick look at the district’s website or a local news story. It takes minutes per record, and it’s the difference between looking like you did your homework and looking like you didn’t.
Where else this works
School construction is a niche, but the pattern isn’t. Lots of buying decisions leave a public trail before the money gets spent:
- Building permits for roofers, HVAC and solar installers, and anyone who sells into new construction.
- Property sales for businesses that serve new owners: dock builders, landscapers, movers, home services.
- New business registrations and licenses for insurance, bookkeeping, equipment, and restaurant suppliers.
- Public bid boards and meeting minutes for anyone who sells to towns, schools, or utilities.
The questions are the same each time. Who actually makes the decision? What public record shows up before they make it? How old is that record, and is it still true?
The best lead isn’t everyone who might buy. It’s whoever is about to decide, found while there’s still time to be part of the decision.