Why I Check Unsafe Driving Before I Ever Look at the Crash Indicator

There's a carrier I almost used three years ago that had a spotless Crash Indicator. Zero reportable crashes in the preceding 24 months. Four years of active authority, $1M BIPD on the L&I record, no safety rating issues. The kind of SAFER snapshot that makes you want to check the box and move on.

The Unsafe Driving BASIC was at the 84th percentile.

I didn't approve them. I got lucky — I was having a thorough day and actually scrolled down in the SMS dashboard instead of stopping at the first line that looked clean. Six weeks after I turned them down, one of their drivers was in a serious crash on an interstate run. Pre-crash investigation clocked him at 58 in a 40. When investigators pulled the carrier's inspection history, they found four speeding citations from the prior 18 months, all logged in FMCSA's system, all visible to anyone who looked. None of it showed up in the Crash Indicator because none of those incidents involved a crash.

The Crash Indicator only moves after someone gets hurt. The Unsafe Driving BASIC is keeping score the whole time.

That's the distinction most brokers miss. They open the SMS scorecard, glance at Crash Indicator first because it sounds most relevant, and if that looks clean, they relax. Crash Indicator is the output. Unsafe Driving is the input. If you want to predict how the next crash happens, don't start with the signal that only fires after one already has.

What the Unsafe Driving BASIC Actually Captures

The BASIC pulls from roadside inspection data and state police reports. The violations it measures are things an officer can observe directly: speeding citations, reckless driving findings, following-too-close violations, improper lane changes, handheld-device infractions.

Each violation goes into FMCSA's SMS with a severity weight. Reckless driving hits the top of the scale. So does texting. A driver caught doing 15-plus over the limit in a CMV pulls a near-maximum severity weight. Minor speeding — under 5 mph over — registers, but at a fraction of the weight the serious violations carry. The carrier's final percentile is exposure-adjusted, so you're seeing behavior normalized for miles driven, not raw counts.

The regulations behind these violations are specific. 49 CFR § 392.2 requires drivers to comply with all applicable traffic regulations — that's the federal anchor for every state speeding and reckless-driving citation. 49 CFR § 392.80 prohibits texting while operating a CMV. 49 CFR § 392.82 prohibits using a hand-held mobile telephone while driving. And here's the one brokers almost never cite: 49 CFR § 392.6 puts the obligation on the carrier, not just the driver. A motor carrier cannot schedule a run in a way that would require the driver to exceed the speed limit to make the delivery on time. So when you have a fleet with repeated speeding violations, you don't just have drivers making bad choices. You potentially have a carrier whose operational model is producing those choices.

The 65th percentile is the threshold where FMCSA starts sending intervention letters. Above 80th, the carrier becomes a candidate for an on-site investigation. A carrier above the 80th percentile for Unsafe Driving has been flagged by the federal government as a priority safety problem. That flag is public. It was public when you approved them.

Lagging Signal vs. Leading Signal

Think about what the Crash Indicator actually tells you. It measures crashes that have already happened. That's not useless — a carrier that crashes at twice the rate of its peer group is probably doing something wrong — but it's retrospective by definition. The crashes are in the past.

The Unsafe Driving BASIC is close to real-time. Roadside inspection data typically flows into FMCSA's system within a few days of the stop. A 24-month rolling window of Unsafe Driving violations is about as current as any public safety data you can access. You're looking at documented behavior from recent loads, not a five-year-old crash report.

There's also a survivorship problem with Crash Indicator. A carrier can have genuinely dangerous driving behavior and no crashes because they've been lucky — no one was in the intersection the three times the driver blew through it. No crash happened to be reportable in the past 24 months. The behavior was there. The paper trail just wasn't. Luck isn't a safety program, and it doesn't stay good forever.

Conversely, a clean Unsafe Driving BASIC paired with elevated Crash Indicator is worth understanding. It might mean the crashes were legitimately non-preventable — rear-ended, deer strike, black ice with other vehicles involved. The Crash Preventability Determination Program exists to flag those, but it's voluntary, so many carrier records show raw crash counts that include incidents any reasonable person would call not-at-fault. Unsafe Driving cuts through that ambiguity because it measures what the driver was actually doing. A speeding citation means the driver was speeding. There's no ambiguity to litigate.

The Legal Theory Is the Stronger One

Post-Montgomery v. Caribe Transport II, LLC — the Supreme Court's unanimous May 2026 ruling that the FAAAA doesn't preempt state-law negligent-selection claims against freight brokers — the practical question for every approval decision is: how does this look to a jury?

Crash Indicator evidence is arguable. A defense lawyer can work with it: "Three crashes in two years sounds bad, but two were non-preventable, one was ice-related, and their peer average is 2.1, so statistically the difference is noise." That argument won't always land, but it's not nothing. There are variables to fight about.

Speeding violations don't leave much to argue about. The plaintiff's theory constructs itself: "You had access to MC-2034517's FMCSA Safety Measurement System data before this load. That data showed four speeding violations in the 18 months preceding the crash, all documented in a federal database any broker can access for free. The driver who caused this crash was traveling 58 mph in a 40 mph zone. This was not an anomaly. It was a pattern. You chose not to look."

That is a much cleaner theory than "their crash count was elevated." And under 49 CFR § 392.6, a broker who approved a carrier whose operational model produces speeding violations isn't just looking at driver negligence — they're looking at carrier negligence, which means the broker's decision to use that carrier is being evaluated against what they knew about the carrier's regulatory compliance posture, not just their crash outcomes.

I'm not a lawyer. But I've read enough post-accident analyses by plaintiff's firms to have learned where the exposure concentrates. It concentrates in the signals you had access to and didn't review. Unsafe Driving is a public, timestamped, severity-weighted record of how the carrier's drivers have behaved at the wheel. After Montgomery, not checking it is a documented gap.

What I Do With a Specific Score

Say I'm evaluating DOT-4892341, running a dedicated dry van lane, four years of active authority, $750K BIPD on file. SMS shows Unsafe Driving at 71st percentile.

The first thing I do is click through to the underlying violations. The percentile is the headline; the violations are the story. A 71st percentile built from four minor speeding citations (3-4 mph over) spread across 180 inspections over 24 months is a very different carrier than a 71st percentile with one reckless-driving finding and a handheld-device citation from the past six months. Same score. Completely different risk profile.

If the violations are low-weight and spread thinly over a long inspection history, I'm probably approving this carrier. If I see any reckless driving finding, any texting or handheld-device citation, or a pattern of significant speeding (anything 11 mph or more over the posted limit), I'm calling dispatch. Not because I expect a useful explanation — dispatch rarely has the roadside details — but because the conversation is part of my diligence record, and how the carrier responds tells me something about whether they know this is a problem.

If the score is above 75th percentile with substantive violations, I'm not approving without something that explains the score. "High utilization" doesn't explain it. That's what exposure normalization already accounts for in the percentile calculation. An explanation would be something like: "We had a fleet expansion in Q3, two drivers in their first 90 days, both violations from that cohort, both drivers terminated." That's a real explanation. Give me the details and I'll document them. Give me a shrug and I'm passing on the load.

The Carrier's Obligation, Not Just the Driver's

One point worth understanding before you get on the phone with dispatch: the driver behavior feeding the Unsafe Driving BASIC isn't solely a driver problem. The 49 CFR § 392.6 obligation is on the carrier. If drivers are consistently speeding to make appointment windows, the carrier's scheduling and dispatch model may be producing the violations — not just tolerating them after the fact.

This reframes how you read the score. A carrier with four low-severity speeding citations probably has a few drivers making ordinary judgment errors. A carrier with multiple high-severity violations across different drivers, spread over different routes, suggests something systemic about how that fleet operates. The percentile treats them differently — which is exactly why clicking through to the violation detail matters. The score is a signal. The violations are the story behind it.

How I Document This

Every carrier approval that touches BASIC scores gets a written note in the carrier record, not just a checkbox.

For Unsafe Driving specifically, the format I use: date the data was pulled, the percentile at time of review, whether I clicked through to the underlying violations, what violations were present (type, severity weight tier, and date range), and the disposition — approved, approved with documented exception, or declined.

If the score is above 65th and I'm approving, I write the one-line rationale: "71st percentile from three low-weight speeding violations over 180 inspections in 24 months — volume-driven exposure score, no high-severity violations in record." That's it. Not a memo. One sentence that shows I understood what I was looking at.

If I made a call to the carrier because of the score, I log it: date, name or role of who I reached, the substance of the conversation, even if it was essentially nothing. "Called dispatch 6/1/26, asked about 71st pct Unsafe Driving, rep confirmed no awareness of specific violations, no changes to safety program noted" — that still goes in the file. It shows the check was live and deliberate, not a box-check on a form.

A plaintiff's attorney reading your carrier file three years from now is going to look for two things: that you pulled the relevant data, and that you understood what it said. The file needs to answer both questions before they ask. Unsafe Driving is too specific a signal to leave undocumented. If you looked at it and decided to approve, show your work. If you didn't look at it, that's the part that ends up in the opening argument.

— Mason Lavallet

Founder, DOTScreener.com