CSA SMS BASIC Scores: Methodology, DataQs Challenge Workflow, Inspection Data Reconciliation, and Severity and Time Weighting

What the Safety Measurement System Is Measuring

The Compliance, Safety, Accountability program, usually shortened to CSA, is the FMCSA enforcement model that uses roadside inspection results, crash reports from the states, and investigation findings to decide which motor carriers get attention first. The data engine behind it is the Safety Measurement System, or SMS. SMS takes 24 months of inspection and crash records, sorts the violations into seven Behavior Analysis and Safety Improvement Categories, computes a measure for each category, and then ranks each carrier against similar carriers to produce a percentile from 0 to 100. A higher percentile means worse relative performance. When a percentile crosses an intervention threshold, the carrier is flagged for prioritization, which may lead to warning letters, targeted roadside inspections, offsite or onsite investigations, and in serious cases a safety rating process.

Carriers often treat the scores as weather, something that happens to them. That is the wrong mental model. SMS is a calculation over a set of records, and the carrier has standing to review every one of those records. Some records are simply wrong: a violation written against the wrong USDOT number, a crash that was not reportable, a citation that was later dismissed in court, an inspection that lists a trailer the carrier never owned. Other records are correct but misunderstood by the carrier, because nobody in the office connected the inspection report to the trip it happened on. The data defense job is to close that gap. You cannot change what the officer saw at the scale, but you can make sure the record reflects it accurately, and you can make sure your own systems explain why it happened.

The Seven BASICs and Where Violations Land

Each violation code in the FMCSA violation list maps to exactly one BASIC. Dispatch, safety, and maintenance staff should know the seven categories well enough to predict where an inspection will land before the SMS update arrives. The Unsafe Driving BASIC collects moving violations such as speeding, reckless driving, improper lane change, following too closely, and texting or handheld phone use. The Crash Indicator BASIC uses state reported crashes that meet the federal recordable definition, generally a fatality, an injury requiring transport for immediate medical treatment, or a vehicle towed from the scene due to disabling damage. The Hours of Service Compliance BASIC collects log and ELD violations, including false logs, driving beyond the limit, and missing records of duty status. The Vehicle Maintenance BASIC covers brakes, lights, tires, cargo securement, and the inspection, repair, and maintenance requirements of 49 CFR Part 396. The Controlled Substances and Alcohol BASIC covers possession, use, or being under the influence. The Hazardous Materials Compliance BASIC covers placarding, shipping papers, packaging, and loading under Parts 171 through 180. The Driver Fitness BASIC covers licensing, medical certificates, and driver qualification file requirements under Parts 383 and 391.

The normalization column matters more than most carriers realize. Unsafe Driving and Crash Indicator are normalized by exposure, meaning the number of power units with an adjustment for vehicle miles traveled when the carrier has reported it on the MCS-150. The other five are normalized by the count of relevant inspections. A relevant inspection for Vehicle Maintenance is one that included a vehicle examination; a relevant inspection for Driver Fitness is one that included a driver examination. This is why a clean inspection has real value. It adds to the denominator without adding to the numerator. A carrier that is missing clean inspections from its record, or that has not updated its power unit and mileage figures, is carrying a worse measure than its actual performance supports.

Severity Weights, Out of Service Points, and the 24 Month Window

SMS does not count violations equally. Each violation code carries a severity weight on a scale from 1 to 10 that FMCSA assigned based on its judgment of crash risk. Within a single inspection the weights in one BASIC are capped so that a long list of related defects on one truck does not swamp the category. For BASICs where an out of service order is possible, an out of service violation receives additional severity points on top of the base weight. The exact weight for each code is published in the SMS Methodology document and its violation appendix on the FMCSA CSA website, and the carrier should pull that list into its own database rather than relying on memory.

Time weighting is the second multiplier. The SMS window is 24 months. Under the published methodology, events in the most recent portion of the window receive the highest time weight, events in the middle portion receive a lower weight, and events in the oldest portion receive the lowest weight. The widely described version of this is a three, two, one pattern across the window, which is a useful way to think about it even though the carrier should confirm the exact boundaries in the current methodology document. The practical effect is that a violation from last month does roughly three times the damage of the same violation from a year and a half ago, and every violation eventually ages out entirely at the 24 month mark. Crash Indicator uses a similar time weighting combined with a severity weight for the crash itself based on whether it involved a tow, an injury, or a fatality, with an additional weight when hazardous materials were released.

Model this in your own system. A violation table with columns for event date, BASIC, base severity, out of service flag, and a computed time weight lets you reproduce the numerator for each BASIC and, more importantly, lets you forecast it. You can see that two speeding violations will fall from the highest time weight band to the middle band next quarter, and you can see that a single out of service brake violation from this week outweighs several older defects combined. That forecast is what tells a safety manager whether to spend effort on a DataQs challenge, a training intervention, a maintenance change, or simply patience while old events age out.

Safety Event Groups, Percentiles, and Intervention Thresholds

A raw measure is not a score. To produce a percentile, SMS places each carrier into a safety event group with carriers that have a similar number of inspections, crashes, or power units, depending on the BASIC. A carrier with 8 relevant inspections is compared against other carriers with a similar small count, not against a national fleet with thousands. Inside each group the measures are ranked and converted to percentiles. A carrier also needs a minimum number of events in a BASIC before a percentile is assigned at all, which is why a small carrier can see measures with no percentile for several categories.

FMCSA publishes intervention thresholds for each BASIC that differ for general freight carriers, hazardous materials carriers, and passenger carriers. In the methodology that has been in effect, the thresholds for Unsafe Driving, Crash Indicator, and HOS Compliance are lower, meaning a carrier is flagged at a lower percentile, than the thresholds for Vehicle Maintenance, Controlled Substances and Alcohol, Driver Fitness, and Hazardous Materials Compliance. Hazardous materials and passenger carriers are held to tighter thresholds across the board. Rather than quote a specific number here that may be superseded, carriers should confirm the current threshold table in the SMS Methodology document and record it in their own configuration table so that the internal dashboard uses the same figures FMCSA uses.

Two visibility points cause confusion. First, under the FAST Act of 2015, FMCSA stopped displaying property carrier BASIC percentiles and alerts to the public, although the carrier can still see its own percentiles when logged into the SMS portal, and enforcement still uses them. Second, the Crash Indicator and Hazardous Materials Compliance percentiles have long been treated differently in public display. Shippers, brokers, and insurers who say they are looking at your scores are often looking at the raw inspection and violation counts and at out of service rates, which remain public, or at third party interpretations of them. The carrier's logged in view is the authoritative one, and a monthly export from it should be archived.

Proposed Methodology Changes and What to Watch

FMCSA has published a proposed revision to the SMS methodology, sometimes described as the CSA prioritization changes, and has made a preview site available so carriers can see how their data would look under the proposed approach. In general terms the proposal reorganizes the categories, including separating maintenance defects a driver should have observed on a pre trip inspection from other maintenance defects, consolidates the long list of individual violation codes into a smaller set of violation groups so that multiple related violations on one inspection count once, simplifies the severity weight scale, and refines how carriers are segmented and compared. FMCSA has stated that the intent is to focus enforcement on carriers with the highest crash risk and to make the results more stable and easier to understand.

The data defense approach described in this monograph does not depend on which version of the methodology is in force. Inspection records, violation codes, crash reports, and the DataQs process remain the inputs. What changes is the weighting layer, which is exactly why the weighting should live in a configuration table in your database rather than hard coded in a spreadsheet. When FMCSA finalizes a change, you update the table and rerun the forecast. Carriers should check the FMCSA CSA website for the current status of the proposal and the preview, and should not assume a specific effective date until FMCSA announces one.

Reconciling Inspection and Crash Records Against Your Own Data

The core discipline is a reconciliation, and it works like a bank reconciliation. On one side is the FMCSA record set: every inspection and crash in the SMS portal download for your USDOT number. On the other side is your operating record: drivers, power units, trailers, and dispatched trips in the TMS, plus ELD and telematics location history. Every FMCSA record should match exactly one trip in your data. Where it does not, you have found either an error in the FMCSA record or a gap in your own data, and both are worth fixing.

Build the match in layers. Start with the inspection report number, date, time, and state. Then match the driver license number on the report to your driver master, the power unit plate or VIN to your unit master, and the trailer plate to your trailer master. Then find the dispatched trip for that driver and unit on that date and confirm the inspection location sits on or near the route using ELD GPS history. A match on all layers is a confirmed record. A match on driver and unit but not on location suggests a date or time error. A plate or VIN that does not exist in your master, or a driver license number you have never employed, is a strong signal that the inspection was written against your USDOT number in error, which happens with leased equipment, recently sold trucks, and similar carrier names. Crash records get the same treatment with the added step of comparing the state crash report to your own incident report and determining whether the event actually met the federal recordable definition.

• Confirmed - All layers match; move on to coaching, maintenance, or acceptance.
• Wrong carrier - Driver, unit, or trailer not in your masters; candidate for a DataQs challenge.
• Wrong facts - Matched trip but the violation detail conflicts with ELD, work order, or paperwork; candidate for a challenge with evidence.
• Court outcome - Citation dismissed or reduced; candidate for an adjudicated citation request with the court document.
• Own data gap - Record is right but the TMS has no trip; fix dispatch or ELD process.

This reconciliation is a natural SQL data mining project. The SMS download is a flat file, the TMS has driver, unit, and trip tables, and the ELD vendor has a location history API. A set of stored queries that runs the match and writes an exception list takes far less time than the manual review most safety departments do today, and it produces the same result every month without depending on one person's memory.

The DataQs Request for Data Review Workflow

DataQs is the FMCSA system for requesting a review of federal and state data that FMCSA releases to the public, including inspection and crash records. A carrier submits a Request for Data Review, often shortened to RDR, identifies the record, states the reason it believes the record is wrong, and attaches supporting documents. The request routes to the agency responsible for the data, which for roadside inspections is usually the state enforcement agency that conducted the inspection. That agency reviews the request and either corrects the record, removes it, or denies the request with an explanation. FMCSA also operates a Crash Preventability Determination Program through DataQs for certain crash types, where a carrier can submit evidence that a crash was not preventable; crashes found not preventable are treated differently in the Crash Indicator calculation. Carriers should read the current DataQs guidance for the eligible crash types and documentation requirements because they have changed over time.

The quality of the evidence package determines the outcome more than the quality of the argument. A request that says the driver was not speeding is an opinion. A request that attaches the ELD speed history for the fifteen minutes around the inspection time, the GPS breadcrumb showing the posted limit segment, and the dashcam clip is evidence. For a maintenance violation, attach the work order that shows the defect was repaired, the pre trip inspection record, and the most recent periodic inspection for the unit. For a wrong carrier record, attach your unit list showing the plate is not yours and, if the equipment was sold or the lease ended, the bill of sale or lease termination. For a court outcome, attach the certified court disposition. For a crash, attach the police report, your incident report, the ELD data, and any video. Keep the narrative short and factual and let the attachments carry the weight.

Track every request. A DataQs tracking table in the TMS or safety system should hold the RDR number, the inspection or crash it concerns, the challenge type, the filing date, the responsible agency, the current status, the decision date, the outcome, and a link to the evidence package stored in document imaging. Status should be checked weekly. Requests that are denied can sometimes be appealed to a higher review level within the responsible agency, and the tracking table should record whether an appeal was filed and its result. Without this table, carriers lose track of open requests and cannot tell management how many challenges were filed, how many succeeded, and how much measure they recovered.

The Data Model Behind a Durable Program

A durable program lives in tables, not in a shared drive full of PDFs. At minimum the model needs an inspection header table keyed by report number, with date, time, state, inspection level, location, and the matched driver, unit, trailer, and trip identifiers plus a reconciliation status. A violation detail table keyed by report number and sequence holds the violation code, the regulation section, the BASIC, the base severity weight, the out of service flag, the time weight as of the current SMS snapshot, and any link to a DataQs request. A crash table holds the state report number, date, location, severity indicators for tow, injury, fatality, and hazmat release, the matched driver and unit, the recordable determination, and the preventability status. A DataQs request table holds the fields described above. A configuration table holds the severity weights, time weight bands, event group boundaries, and intervention thresholds for the methodology version in force, with an effective date so you can rerun history under a new version.

The joins to the operating data are what make it useful. Driver master, unit master, trailer master, and trip or load tables already exist in any TMS. The ELD and telematics location history usually sits with the vendor and should be pulled on a schedule into a location table with driver, unit, timestamp, latitude, longitude, and speed. Our ELD and EDI integration work typically establishes that pull once and leaves it running, so that the evidence for a DataQs request is a query rather than a frantic request to the vendor's support desk weeks after the fact. The TMS implementation side of the work adds the reconciliation status and DataQs link fields where dispatch and safety can see them on the driver and unit records.

Retention deserves a sentence. Evidence packages should be kept for at least the 24 month SMS window plus the time an appeal might take, and in practice most carriers keep them longer because the same records support insurance renewals and litigation holds. Access should be restricted to safety and management, and every export should be logged, which overlaps with ordinary data protection controls.

KPIs and the Monthly Cadence

The program needs a small set of indicators that management can read in a minute. Track the count of inspections and crashes in the current 24 month window by BASIC. Track the time weighted severity total by BASIC, computed from your own tables, and compare it to the measure FMCSA shows to confirm your model is reproducing theirs. Track the clean inspection rate, meaning inspections with no violations divided by total inspections, because it drives the denominator. Track the reconciliation rate, meaning confirmed matches divided by total FMCSA records, and the count of unmatched records older than 30 days. Track DataQs requests filed, open, granted, and denied, and the severity points recovered through granted requests. Track the projected measure for each BASIC at 90 days out assuming no new events, which shows what aging alone will do.

The cadence is monthly because SMS updates monthly. Within the first week after the SMS release, download the inspection and crash files and load them into the inspection and crash tables. Run the reconciliation queries and review the exception list with safety and dispatch. Decide on each exception: file a DataQs request, correct an internal record, or accept the record and route it to driver coaching or maintenance follow up. Assemble and submit evidence packages for the requests you decided to file. In the second and third weeks, check status on open requests and update the tracking table. In the fourth week, produce the KPI summary and the 90 day forecast for management, and update the configuration table if FMCSA has published any methodology change. A business process optimization review usually finds that this entire monthly cycle takes one person a few hours once the queries exist, compared to days of manual work when it is done from PDFs.

For example, a 100 truck fleet that averages a few dozen roadside inspections a month would see a few hundred inspection records in the window, of which a small share might be reconciliation exceptions in a typical month. That is a manageable list for one safety coordinator, provided the match runs automatically and the evidence is one query away. The carriers that struggle are the ones that let twelve months of records pile up and then try to challenge everything at once, after the ELD vendor has purged the location history and the driver has left.

Related practice areas for this work include SQL Data Mining ELD and EDI Integration TMS Implementation Business Process Optimization. Each practice page describes scope, deliverables, and how an engagement starts with a short discovery call.

How Strong Tower Consulting Helps

Strong Tower Consulting helps carriers build the reconciliation queries, the DataQs tracking table, and the weighting configuration described here inside the TMS and the reporting database they already run, instead of in a spreadsheet that only the safety manager understands. The work starts with a download of the current SMS data and a match against driver, unit, and trip records, which usually surfaces the first round of challengeable records within days. From there the monthly cycle is documented, the evidence pulls from ELD and imaging are automated, and management gets a one page view of where each BASIC is heading. Strong Tower Consulting, LLC and Exspeedite LLC are separately organized sister companies under common ownership. That separation keeps consulting advice vendor neutral and focused on the carrier operation.

To discuss your fleet, your current scores, and the systems you already have, call 763-244-1373 and ask for Scott Elliott. A short call is enough to scope the first reconciliation and identify which records are worth challenging before the next SMS update.

Frequently Asked Questions