← All posts · May 10, 2026 · 6 min read

Replacing $25k/yr vendor data with public records: the playbook

If you're an LBM operator, a trade contractor, or a construction-supply distributor, there's a good chance you're paying — or being pitched — $15-30k per market per year for a vendor data subscription. Zonda, Hanley Wood (BDX), MetroStudy. The pitch is uniform: builder lists, market reports, competitive intelligence.

And the pitch is largely true. Those products work. They have data you can't get for free. But for trade operators specifically, a lot of what you actually use the vendor data for is information that already exists in public records, free, if you know where to look. This post walks the practical playbook.

What you're paying for, broken down

A typical vendor-data subscription bundles four things:

1. Building permit data by county and project
2. Builder identification + contact info
3. Community / subdivision data with lot counts, status, pricing
4. Sold-vs-unsold lot data + closings (the MLS-licensed piece)

Items 1, 2, and (partially) 3 are derived from public records the vendor aggregates for you. Item 4 is the MLS-licensed proprietary piece — that's the part that genuinely can't be replicated from free sources.

What's free vs what isn't

The 80/20 question

Now the question that matters: of the four data buckets, which ones do you actually use?

Most trade operators we talk to use buckets 1, 2, and 3 every week and bucket 4 (sold-vs-unsold) maybe once a quarter, mostly for context. The high-frequency, decision-driving data is the public-records side. The proprietary MLS side is nice-to-have but rarely the binding constraint.

If you can replicate 80% of what you use Zonda for from public records at $1,499/mo instead of $25,000/yr, the ROI math is obvious. The 20% you give up is the sold-vs-unsold piece, which is usually relevant for a quarterly board deck — not a Monday morning call list.

The playbook to replicate it yourself

If you want to build this in-house, here's the order:

1. Pull Census BPS monthly. ~400KB file per month, one cron job. Gives you nationwide county-level permit volume, broken into 1-4 unit and 5+ unit.
2. Wire 2-3 county ArcGIS feeds for your most important markets. Each takes 2-4 hours of engineering. You get per-permit detail (address, date, contractor name).
3. Cross-reference contractor names against state DBPR / equivalent. Florida is the cleanest (140k licenses, simple CSV download). Other states are similar.
4. Add drive-time isochrones via OpenRouteService (free tier covers most operators) or Mapbox (paid, scales).
5. Build the UI for filtering, ranking, watchlists, and outreach tracking. This is the biggest single time investment.

Time estimate from scratch: 3-6 months of engineering for one developer who knows what they're doing. Maintenance: roughly 5-10 hours/month per county feed (portals change, formats shift, scrapers break).

Or buy the prebuilt version

We built PermitFocus exactly because the playbook above is real but not most trade operators want to do it themselves. PermitFocus is $1,499/mo for the Pro tier — single-state $499/mo, multi-state $1,499, custom enterprise.

Math: if you're currently paying $25k/yr for a vendor that delivers buckets 1-3 plus the MLS-licensed bucket 4, swapping to PermitFocus saves you ~$7,000/yr (Pro tier annual) and you keep 80% of the data + workflow utility. If your team uses bucket 4 enough to justify the difference, keep the vendor. Most teams don't.

Either way, the point isn't that one is "better" — it's that the cost ratio of the data you actually use vs the data you're paying for has shifted dramatically as public-records access has improved. Pricing yourself out of vendor data when you'd use the data daily is bad. Paying for vendor data you barely open is also bad.