What FMCSA data actually tells you about a carrier's risk.

MC 559018 — “Harbor Star Trucking” — replies to a Long Beach CA → Salt Lake City UT van at $2,650. You pull their SAFER snapshot. Authority active for six years. Insurance on file. Three crashes in the last 24 months. Driver out-of-service rate at 9.2% — above the national mean. One unsafe-driving inspection flagged 11 months ago. Is this a carrier you book or a carrier you pass on?

The honest answer is “it depends what else is in the file.” FMCSA data is rich, but most brokers either read one signal in isolation (one bad inspection → no thanks) or skip the data entirely (rate looks good → book it). The discipline is reading three or four signals together. This post is a working guide to which FMCSA data points actually matter for risk, what gets misread, and a 90-second pre-book check using free FMCSA tools.

What's in the FMCSA dataset

Operating authority

The most basic check, and the one that catches the most fraud. Is the MC's operating authority currently active? When was it granted? Has it ever been involuntarily revoked? A 60-day-old authority quoting on a premium long-haul lane is a stop sign. A 6-year-old authority with continuous active status is the baseline normal.

Insurance on file

Cargo and liability policies filed with FMCSA, with effective dates and coverage amounts. Two things to read here: is the coverage current and adequate for the cargo you are booking, and how recent are the filings? A carrier with six years of authority but an insurance policy filed three weeks ago has had a coverage gap. That gap is worth a question.

Crashes

Reportable crashes in the last 24 months, with severity and fatalities. Crash counts have to be read against fleet size — a 200-truck fleet with three crashes is roughly normal; a 5-truck fleet with three crashes is not. The crash-rate calculation, not the raw count, is the useful number.

BASIC scores

The Compliance, Safety, Accountability percentile scores across seven categories — Unsafe Driving, Hours-of-Service, Driver Fitness, Controlled Substances, Vehicle Maintenance, Hazmat, Crash Indicator. Each score is a percentile relative to similar carriers. A score above the intervention threshold (65 for most categories, lower for Crash) is a meaningful signal; a score just under is normal.

BASIC scores are aggregated over 24 months and lag real carrier behavior. A carrier that just hired a bad batch of drivers will look fine on BASIC for months. A carrier that just cleaned up will look bad on BASIC for months. Use them as trend signals, not point-in-time verdicts.

Inspection history

Every roadside inspection in the last 24 months, with results and violation codes. A single bad inspection is not a risk signal — every carrier has them. The pattern that matters is a string of out-of-service inspections, or a consistent violation code that points at a systematic problem (brake adjustment, log violations, tire condition).

Out-of-service rate

Percentage of inspections that resulted in a driver or vehicle being placed out of service. National means run roughly 6.7% driver and 21% vehicle. A carrier sitting materially above those means in either category is a maintenance or compliance problem you do not want to inherit.

What gets misread

More carrier vetting decisions go wrong by misreading a single data point than by missing data entirely. The most common errors:

• Treating one bad inspection as a risk signal. Roadside inspections happen. A single out-of-service event on a five-year-old carrier with otherwise clean data is noise, not signal.
• Ignoring fleet size when reading crash counts. Three crashes on a 200-truck fleet is unremarkable. Three on a 5-truck fleet is a serious risk indicator. Always normalize.
• Reading BASIC scores as point-in-time accurate. They are 24-month aggregates. Movement matters more than the absolute number.
• Reading authority age in isolation. A new MC on regional reefer work is normal. A new MC on long-haul spot is the fraud setup. Same data point; opposite meaning.
• Skipping the address check. An MC registered to a UPS Store or a residential address is not illegal, but it is unusual for an established carrier and worth a question.

The composite signals that actually matter

One FMCSA flag is rarely the whole story. The signals to actually act on are composites — two or three patterns showing up together that no honest carrier would naturally produce:

• New authority + fresh insurance + premium spot rate. Sub-6-month authority, insurance filed in the last 30 days, quoting within $50 of the median bid on a long lane. The classic fraud-setup composite.
• Sudden lane change. A carrier whose historical bookings cluster in the Southeast suddenly quoting on a Pacific Northwest reefer. By itself, fine. Paired with new equipment claims or a domain mismatch, a stop sign.
• BASIC trend reversal. Scores improving for two years and then jumping back across the intervention threshold in the last quarter. The carrier just changed something — driver pool, maintenance vendor, dispatching — and the change is not going well.
• Out-of-service spike on the vehicle side. A carrier whose vehicle out-of-service rate doubles quarter-over-quarter is running equipment harder than the equipment can handle. Cargo claim risk goes up before the crash risk does.
• Address or officer overlap with a shutdown MC. The chameleon trace — see the chameleon carrier deep dive.

The 90-second pre-book check

Free tools, no subscription, runnable on any phone. For each carrier you are about to book:

1. Pull the MC on SAFER Company Snapshot. Confirm authority active, insurance on file, fleet size and crash count make sense together.
2. Read the BASIC scores. Anything in the red on Unsafe Driving, Vehicle Maintenance, or Hours-of-Service is a yellow flag. Anything in the red on Crash Indicator is a stop sign.
3. Glance at the last six inspections. Pattern of out-of-service events, or a recurring violation code, is the signal. One-offs are noise.
4. Authority age + lane + rate sanity check. New MC + premium lane + tight bid is the composite to avoid.
5. Call the FMCSA-registered phone number, not the number in the email. A 30-second call confirms the dispatcher knows about the load and the carrier you think you are booking is the carrier on the other end.

For a longer walk-through with screenshots, see the FMCSA carrier lookup guide and the free FMCSA lookup tool.

What FMCSA data does not tell you

One last honest note: FMCSA data is a snapshot of compliance and safety. It does not tell you whether a carrier is about to rebroker your load. It does not tell you whether the dispatcher you are emailing is who they say they are. It does not tell you whether the truck that shows up at pickup will be the truck on the rate con. Those are different checks — email-side and post-booking — and they need the inbox and the dock, not the FMCSA dataset, to catch.

If you would rather not run this check by hand on every quote, Keelway scores every inbound reply against FMCSA on the server before the broker ever sees the ranked list — see Carrier Trust Score, our FMCSA lookup tool, and the FMCSA carrier lookup primer.