Finding your next client in public construction data.

Two kinds of business development.

The firm's earlier guide covers inbound: subscribe to the right public channels, and let bid notices come to you. That is the right first move, and the free Florida bid channels that beat ConstructConnect lays it out. But inbound has a ceiling. You can only respond to what someone chose to advertise, and the highest-value relationships often start before anything is posted.

Outbound is the other half. It means deciding who to approach first, on your own initiative, which is a harder problem because it needs a reason. Cold contact with no signal is spam, and it converts like spam. Cold contact timed to a real event, sent to a business that actually fits, is a different activity entirely. The whole game in outbound is finding the signal, and for construction and the trades around it, the signals are sitting in public records that most firms never read.

A permit is the first signal of construction.

Before a contractor shows up in a directory or an association roster, they file a building permit with the local government. The permit-data company Shovels makes this its core argument, and it is correct: the permit is the clearest trigger event in construction. The public record confirms the project type, the estimated value, the contractor of record, and the status, and the status tells you whether work is underway and whether budgets are committed.

The most useful field is recency. Active contractors pull multiple permits in a rolling window of roughly 30 to 90 days; a contractor with no recent permits may have narrowed focus, shifted to non-permitted work, or left the market. Shovels recommends prioritizing contractors with at least one permit in the last 60 days, which materially improves lead quality before any other filtering. That single filter, recent permit activity in your trade and geography, turns a static directory into a list of businesses you know are operating right now.

The data is available at a range of price points and shapes. Shovels is API-first, built for teams that want to wire permit and contractor data into their own tools, and reports more than 130 million permits and over 2.3 million contractors with roughly 5 million new permits added monthly. BuildZoom runs a comparably large national database, advertising 350 million-plus permits across 25 years and more than 6 million licensed contractors, with tens of thousands of new permits and hundreds of thousands of status updates daily. Construction Monitor has aggregated national permit data since the 1990s and built its product around a weekly list of new permit filers. Many county building departments also publish permit search portals directly, free, which is enough to run the manual version of this for a single market.

Agency notices say where the work is going.

Permits tell you who is building. Public agency notices tell you where the next wave of work is headed. Florida Department of Transportation lettings, the state's water management district project notices, and Department of Environmental Protection construction notices are all public, and together they map where infrastructure money is about to flow. That is a leading indicator for an outbound list: the contractors and subs in a corridor that is about to see funded work are the ones who will soon be busy, hiring, and in need of the services that orbit a construction ramp-up.

For a firm with civil credentials, reading these notices is not generic scanning. Knowing which letting matters, which agency owns which scope, and which kind of project pulls which trades behind it is the difference between a list of names and a list of timed, relevant opportunities. The agency feeds are demand signals; permits are activity signals. The two answer different questions, and the second half of a good outbound list comes from combining them.

Weak digital presence is a different kind of signal.

Some businesses are a fit not because of what they are building but because of what they are missing. A surprising share of small businesses still have almost no web presence. BrightLocal's 2025 SMB marketing study found that only 40 percent of small and medium businesses have a dedicated website, and only 35 percent have a Google Business Profile, despite most of them saying search matters to their business. The trades skew further in that direction than professional services do, with industry surveys putting the no-website rate for local contractors and tradespeople far above the cross-industry average.

Pair that with permit data and a precise target appears: a business that is clearly operating, pulling permits, doing real work and making real money, but is nearly invisible online, with no website or a thin and unclaimed Google Business Profile. That is a business whose lack of web presence is costing it work it cannot see, and it is exactly the buyer profile one of the firm's forming-startup clients is built around. That client's entire model is to find businesses with weak digital presence, build the replacement before any sales conversation, and approach them with the finished product in hand rather than a pitch. The weak-presence signal is what makes that model targetable instead of a guess.

Combining the signals into a ranked list.

On their own, each signal is noisy. Together they rank. A scan pulls recent permits filtered to the trades and geography you care about, cross-references the agency notices that say where funded work is heading, and checks each candidate for a website and a claimed Google Business Profile. The output is a scored, de-duplicated call list: active businesses, in the right place, sorted by a weighting that reflects what you actually pursue, with the presence gap noted so the outreach can be specific.

The scan is the AI-systems part: turning fragmented public records into a clean, ranked, deduplicated list is a data pipeline, and the leverage is that it runs continuously instead of one afternoon a quarter. The civil-vertical tuning is what makes it land: the weights encode which trades fit, which counties and transportation districts are reachable, which project values are worth the effort, and which presence gaps signal a real opportunity. A general outbound tool can rank any list. The tuning that knows a stormwater scope from a structural one, and which one your firm wins, is the wedge.

The honest limits.

Public data is noisy and it lags. Permits do not capture non-permitted work, and a contractor of record on a permit is not always the decision-maker you want. A missing website is not always a missing business; plenty of healthy trades run entirely on social media and word of mouth, so a presence gap is a hypothesis to check, not a verdict. And cold outbound has a low conversion rate even when it is done well, which means the list is the easy part and the outreach is the hard part. A ranked list of perfect-fit targets approached with a generic, off-key message converts no better than no list at all.

There is also a discipline question that sits underneath the whole approach. Targeting real businesses from public records is legitimate, but the targeting list, who you are approaching and why, is sensitive, and it stays internal. Public signals make a list; they do not make the people on it public. The scan is a way to spend attention where it is most likely to pay back, not a license to blast everyone with a permit and a weak homepage.

What an operator can do.

You can run a manual version of this for one market with no paid tools.

• Start with one county's permit portal. Most building departments publish a free permit search. Filter to your trade and to permits filed in the last 60 days, so every name on the list is currently active.
• Layer in the agency notices for your area. Transportation lettings, water management and environmental project notices. Note the corridors about to see funded work; those businesses are about to get busy.
• Check each candidate for a website and a claimed Google Business Profile. The presence gap is your second signal and the hook for a specific, useful first contact.
• Rank by activity and fit, not by alphabetical order. A short scoring rubric (trade match, geography, recent permit volume, presence gap) beats working a raw list top to bottom.
• Keep a scored list and a win/loss ledger. Outbound only improves if you measure which signals actually convert. Without the ledger you are guessing, and the rubric never gets better.

If the manual version proves out and the wrangling becomes the bottleneck, that is when a permit-data subscription earns its cost. Until then, the free county portals plus a disciplined rubric are enough to test whether outbound works for your firm at all, which is the question to answer before paying for volume.

How Rarefied Earth thinks about this.

This extends the firm's general business-development tooling with civil-specific data sources: permit feeds, public agency construction notices, and weak-presence detection tuned to the trade-business buyer. The firm runs it on its own pipeline first, the standing rule for anything it builds, and the same engine sits behind a forming-startup client's weak-digital-presence targeting. That is the firm's operating model in one example: build a capability for Rarefied Earth, run it until it earns its keep, and only then package it for someone else. The dogfood-and-productize model is the discipline that keeps this honest, and a scan like this is one of its candidates rather than a finished product on a shelf.

Sources and further reading.

Related work.

This is the outbound companion to the firm's inbound guide, the free Florida bid channels that beat ConstructConnect, and to reading a bid package with AI, which covers what to do once a solicitation lands. The operating model underneath all three is in research once, harvest three ways.