If you market a personal injury practice in Ohio, you have been pitched crash leads. The pitch is nearly always the same: fresh injury crashes, delivered daily, priced per record.
What is rarely said out loud is where those records come from. They come from the OH-1 traffic crash reports that agencies across Ohio file and make available on request.
That does not make lead vendors dishonest. Collecting, cleaning, and delivering data is real work, and work is worth paying for. But it does change the question. You are not asking "what is this data worth." You are asking "what is this labor worth, and who should be doing it."
What You Are Actually Buying
Strip a crash lead product down and there are four things in it:
- Access. Somebody requested the report.
- Extraction. Somebody turned a multi-page PDF into structured rows.
- De-duplication. One crash produces several people, and sometimes several reports. Somebody reconciled them.
- Delivery. Somebody put it somewhere you can act on it.
Only the first item is about the data itself, and it is the cheapest of the four. The other three are labor, repeated identically every single day. That is the exact shape of problem software is good at.
The Exclusivity Problem
The biggest issue with bought leads is not the price. It is who else has them.
The underlying record sits with the agency. A vendor is not in a position to make it exclusive to you — at best a contract covers their delivery of it. Another vendor can request the same report from the same agency an hour later.
The practical consequence is that a crash victim may hear from several firms in the same week, all of whom believe they bought something fresh.
Ask any vendor two questions and get the answers in writing:
- How many other buyers receive this same record?
- How many days after the crash does it reach me?
A vendor who answers both clearly is probably worth working with. A vendor who reframes the question is answering it.
Latency
Crash lead value decays quickly, and there are two clocks running.
The first is the agency's. A report does not exist until the officer writes it and a supervisor approves it, which takes days. Nobody can beat that clock — not you, not a vendor.
The second clock is the vendor's. After the agency files, someone has to pull, process, and batch the record before it ships to you. That is additional days on top, and it is the only part of the delay that is actually optional.
A pipeline you run yourself is on the agency's clock alone.
Buy vs. Build, Honestly
| Buying a list | Running your own pipeline | |
|---|---|---|
| Per-record cost | Priced per record, forever | Compute and software only |
| Exclusivity | Shared unless contracted | Yours by construction |
| Latency | Agency time plus vendor batch | Agency time |
| Coverage | The counties the vendor prioritizes | Any of Ohio's 88 counties |
| Ownership | You license the data | The database is yours |
| Upfront cost | Effectively none | One-time or subscription |
| Ongoing effort | None | Runs on a schedule |
The honest summary is that buying trades money and control for convenience, and building trades a setup step to get both of those back.
For reference on the cost side: reading a single crash report with a language model costs us around a fifth of a cent in compute. That is the actual marginal cost of extraction once the pipeline exists. Everything a per-record price adds on top of that is paying for access, delivery, and margin.
When Buying Is the Right Call
It genuinely is, sometimes. Buying makes sense when:
- You need a small volume once, in one county, and never again
- You are testing whether crash-based outreach works for your practice at all
- You have no appetite for operating anything, and that is a legitimate position
Building makes sense when you are buying continuously, when you want counties a vendor does not prioritize, when latency is costing you cases, or when you want the historical database to still be yours next year after you stop paying anyone.
The break-even is not subtle. If you are spending monthly on leads and expect to keep doing it, the arithmetic favors owning the pipeline fairly quickly.
One thing to be skeptical of: any crash lead product advertising phone numbers and email addresses. Those fields are not reliably on an OH-1. If a vendor has them, they were appended from a separate consumer data source, which is a different data-quality question and a different compliance question.
Do Your Own Due Diligence
Whichever route you take, obtaining records and contacting people are two separate questions, and only the first one is about data.
Every regulated profession has its own rules about outreach, they differ by field, and they change. Whatever you build or buy, use it inside the rules that govern your own practice — and get that answer from your own counsel, not from a data vendor.
Be wary of any vendor who volunteers that everything is fine. That is not their call to make, and it is not ours either. We sell software that assembles public information. How it is used is yours to determine.
Where to Start
If you are currently buying, do one thing before your next renewal: ask for those two questions in writing, and price out what the same volume costs to run yourself over twelve months rather than one. The comparison usually answers itself.