Six months after a rough inspection, nobody remembers the trooper's name. The insurance underwriter remembers the number.
To a small-fleet owner, a roadside inspection feels like a one-time event — a bad morning, a few write-ups, back on the road. In the federal system it is something else entirely: a data feed. Every violation flows into the Safety Measurement System (SMS), the engine behind your CSA profile, where it is scored, ranked, and held against you for up to two years.
That score is not abstract. It decides how often you get pulled in for inspection, what a broker sees when they vet you, and what an underwriter quotes at renewal. So the real question isn't “did I get a ticket” — it's how far does one violation actually move the needle. Three forces decide it, and a major FMCSA overhaul is rewriting two of them right now.
Seven BASICs.
Every violation lands in exactly one Behavior Analysis and Safety Improvement Category.
The SMS sorts safety behavior into seven buckets — the BASICs. A violation code on your inspection report maps to one of them, and your standing in each is scored separately. The first job in reading a CSA profile is knowing which bucket a problem is filling:
Unsafe Driving
Part 392
Speeding, reckless, lane use, texting, seat belt
HOS Compliance
Part 395
Logs/ELD, driving beyond the 11- or 14-hour limits
Crash Indicator
Crash history · Not publicly displayed
State-reported crashes, by frequency and severity
Vehicle Maintenance
Parts 393 · 396
Brakes, lights, tires, defects, securement
Controlled Substances / Alcohol
Part 382
Use or possession found at the roadside
Driver Fitness
Part 391
CDL, medical certificate, qualification gaps
HM Compliance
Parts 171–180 · Not publicly displayed
Placards, papers, leaks, package integrity
◆ Source: FMCSA SMS Methodology + 49 CFR Part 385. Thresholds shown are the long-standing intervention lines for general (property) carriers; passenger and HM carriers use lower ones.
Three forces.
Severity, time, and peer group — multiplied together, not added.
- 01
Severity — how bad the violation is
1–10
Today · + 2 if OOS
Every violation carries a severity weight tied to its relationship to crash risk. Under the methodology in effect today that weight runs from 1 (minor) to 10 (severe), and a violation flagged out of service adds 2 more points. A burned-out marker lamp and a brake defect both land in Vehicle Maintenance — but they are not the same size.
- 02
Time — how recent it is
×3 · ×2 · ×1
Today · 0–6 / 6–12 / 12–24 mo
Recent violations hurt more. A violation in the last 6 months counts ×3, one 6–12 months old counts ×2, and one 12–24 months old counts ×1. After 24 months it drops out of the calculation entirely. This is the mechanism that lets a score heal — a clean two years erases a bad inspection on its own.
- 03
Peers — who you are measured against
65th / 80th
Today · intervention thresholds
Your weighted total becomes a percentile — ranked against carriers with a similar number of safety events, not against the whole industry. Cross the 65th percentile (Unsafe Driving, Crash, HOS) or the 80th (the rest) and you move into FMCSA's intervention range. The peer grouping is exactly why the same violation hits a 3-truck fleet harder than a 300-truck one.
◆ The small-fleet penalty ◆
A big fleet's bad inspection gets averaged away. A small fleet's bad inspection is the average.
Why three trucks feel it and three hundred don't.
Picture one out-of-service brake violation. On a 300-truck fleet running thousands of inspections a year, it is one bad line in a deep pool of clean ones — the Vehicle Maintenance percentile barely twitches. On a 3-truck fleet with a handful of inspections on record, that same line can be a large share of the entire sample. The severity and time weights are identical; the denominator is not.
Then the loop closes. A higher percentile raises your ISS score — the Inspection Selection System that tells the next officer whether to wave you through or pull you in. More stops mean more chances for another violation, which raises the percentile again. For a small carrier, one OOS can be the first turn of a wheel that is genuinely hard to stop.
The flip side is just as real, and it is the lever you own: a clean inspection is a data point too. It enlarges the denominator and pulls the percentile down. For a small fleet, deliberately accumulating clean inspections is one of the few fast ways to move a score in the right direction.
Read this before you panic about a number.
FMCSA is rebuilding the SMS — and two of the three forces above are changing.
Most articles you'll find describe the old system as gospel. It isn't. FMCSA has been phasing in a redesigned SMS, and as of mid-2026 the new methodology is available in preview — you can see your own scores under it at the CSA Prioritization Preview — while the old methodology remains the official one until FMCSA announces the switch. What is changing:
Today
The system in effect
- ~2,000 individual violation codes
- Severity weights on a 1–10 scale (+2 for OOS)
- Seven “BASICs”
- Percentile vs. a peer group of similar carriers
Coming
The redesigned SMS (preview)
- ~100 violation groups (codes consolidated)
- Severity simplified to 1 or 2 (OOS = 2)
- BASICs renamed compliance categories, plus a new “Vehicle Maintenance: Driver Observed”
- A proportionate percentile — smoother, and notably fairer to small fleets
For a small carrier the proportionate-percentile change is the headline: it is designed to stop a single inspection from whipsawing a tiny sample the way the old method could. Don't assume your current number and your preview number match — pull both.
◆ “My scores aren't public” is not the same as “nobody sees them”
FMCSA does not display property-carrier percentiles to the general public. But the underlying inspection and violation data is fully visible to insurance underwriters, freight brokers, and shippers, who pull it to price and vet you. The score being hidden from the public does not make it private — it just means you have to log in to see what they already see.
◆ CSA/SMS rules are FMCSA program methodology, not 49 CFR, and they are actively in flux in 2026. Verify the current state on csa.fmcsa.dot.gov before relying on any specific weight or threshold.
Four levers, one honest limit.
You cannot delete a violation that genuinely happened — no service, ours included, can. What you can do is manage everything around it:
Prevent the next one
The cheapest CSA points are the ones never scored. A current DQ file, a real maintenance program, and an on-time drug-and-alcohol program close the doors most violations come through.
Let time do its work
Every violation decays — ×3 to ×2 to ×1 to gone at 24 months. A disciplined clean streak is a score-repair strategy on its own.
Dispute the wrong ones
A violation attributed to the wrong carrier or driver, or factually incorrect, can be challenged through a DataQs Request for Data Review. Reading the report carefully — covered in the report decoder — is what surfaces the disputable lines.
Certify and close out OOS events
An out-of-service event handled well — repaired, certified, and returned on time — keeps a bad day from becoming a pattern in the data.
The carriers who win at CSA aren't the ones who never get inspected — they're the ones who treat every inspection as a data point they can shape.
Prevention, patience, and disputing the wrong lines beat worrying about a number you can't see. A compliance and CSA-profile review turns the abstract score into a short, concrete list of what to fix first.
See what the underwriter sees — and fix it first.
Your CSA score prices your insurance whether you watch it or not. We review your BASICs and inspection history, flag the violations dragging the score, surface the lines worth disputing, and build the prevention program that keeps the next inspection clean.
Disclaimer
For informational purposes only — not legal, tax, or regulatory advice. Always verify requirements with FMCSA, your state agency, and qualified compliance professionals. Regulations and fees change; verify current requirements on official .gov sources before filing.