Methodology & data sources

Independent policy 2.0.0

This is a transparent screening heuristic, not an FMCSA rating, crash prediction, insurance warranty, or dispatch approval. A–F summarizes scored observations within the available public extract. It does not certify a carrier as safe. NR means not rated; it is not an F.

Data and dates

Source datasets on data.transportation.gov: Motor Carrier Census (az4n-8mr2), SMS property-carrier measures (4y6x-dmck), insurance filings (c5y8-a4uz), and crash records (4wxs-vbns). A static rebuild recalculates our policy but does not itself download new records. A scheduled weekly or manually requested refresh downloads and validates a new snapshot; failure retains the previous snapshot without changing its data dates.

Dataset last-updated time, retrieval time, reporting-period end, and page build time are different. Where the publisher does not supply a verified reporting-period end, we show it as unavailable. The API may return records from a different update than the static comparison benchmark. Inspect official current records before relying on a report.

Comparison benchmark

The benchmark is 100 × summed out-of-service inspections ÷ summed inspections across the public SMS property-carrier extract. It is not necessarily the same population or time window as a SAFER national figure. Current stored benchmark: driver 5.47%, vehicle 21.55%; SMS dataset last updated 2026-10-13; 676,951 carrier rows. State comparisons cover indexed carriers, not every carrier or inspection in that state.

Scoring rules

Both driver and vehicle categories need at least five valid inspections and a valid benchmark. A required crash event count, fleet denominator, registration status or applicable docket status that is unavailable withholds the overall numeric score and produces NR. This five-inspection gate is our product rule, not a statistical guarantee. An unknown value is never displayed as active, insured, below average, or free of alerts.

• Driver OOS: −10 above the benchmark; −25 at least 2× the benchmark. Rates are rounded to one decimal before the comparison.
• Vehicle OOS: −10 above the benchmark; −25 at least 1.5× the benchmark.
• Public acute/critical indicators: −15 per indicated category among Unsafe Driving, Hours of Service, Driver Fitness, Controlled Substances / Alcohol, and Vehicle Maintenance. These investigation indicators are not a full SMS percentile/alert assessment. Missing v2 indicator values withhold a grade. Legacy extracts combined missing and negative flags; these are labeled “not indicated in legacy extract,” not “no alerts.” Legacy scores consider only positively observed indicators and must not be read as full indicator clearance.
• Crash records: −20 when fatal crash records reach 0.5 per 100 power units; otherwise −15 above 8 total records per 100 units or −5 above 4. These are unvalidated product thresholds, not official national risk standards. Counts describe the source extract and do not assign fault, preventability, miles traveled, or comparable exposure. Different source windows or fleet changes can affect interpretation.
• Registration and docket: −40 for an explicitly inactive USDOT registration; otherwise −20 for an explicitly inactive docket. Missing is not inactive. No docket located does not establish whether authority is required. Active USDOT registration does not establish for-hire authority.
• Insurance: no deduction for a missing filing. Only a BIPD BMC-91/91X record is labeled a liability filing. Cargo and bond records do not establish liability coverage. A filing, including a historical or future-effective filing, is not proof of current coverage or shipment-specific limits.

A = 90–100; B = 75–89; C = 60–74; D = 40–59; F = 0–39. Deductions are clamped to 0–100 only when the required inputs are sufficient. Observed concerns remain visible on NR reports.

Crash events are not people

One record with two fatalities and three injured people contributes one fatal crash record and one injury crash record, plus two fatalities and three injured people. Fatal and injury categories can overlap. A legacy positive person total cannot reconstruct the event count: we show Unknown and withhold the overall grade rather than substituting people for events.

Official verification and corrections

Use SAFER for the official company snapshot, FMCSA Licensing & Insurance for current authority and filings, and SMS Help Center for public-data limitations. Confirm coverage directly with the insurer. A private carrier may have a USDOT number without MC operating authority; see FMCSA's private-carrier definition.

Report a site mismatch through Contact with the USDOT number, source and date. Requests to review underlying government data go through DataQs; correction or removal is not guaranteed.