# How to Compare Carrier Records with Different Dates Canonical: https://usdotdata.com/blog/compare-dated-carrier-records Author: John Hauler Published: 2026-09-28T07:13:25.439Z Updated: 2026-09-28T07:13:25.439Z --- When two carrier records disagree, compare the same entity and field, identify the date attached to each value, and preserve both observations until you can explain the difference. A newer download is not necessarily newer source information. The useful output is a reconciliation log showing what changed, what remains uncertain and which decision needs another check. This guide helps researchers maintain carrier shortlists and verification files. It addresses conflicts across dated records, rather than deciding whether a particular fleet is growing. John Hauler is USDOTData's editorial persona; this is an independent research workflow, not an FMCSA determination. Sources reviewed September 28, 2026. ## Start with the decision the disagreement affects Write one sentence before comparing files: “I need to know whether this account still fits our service area,” or “I need current authority information for this shipment.” That sentence determines the next source to consult. A conflict about a mailing address may matter for correspondence while telling you little about truck location. A conflicting operating-status label can require an immediate official check before a transport decision. Do not apply one blanket rule such as “always keep the newest row” to every field. Keep the original exports or saved observations unchanged. Create a separate working comparison. This preserves an explanation for a later colleague who asks why the research conclusion changed. ## Confirm that both records describe the same entity Match the USDOT identifier before comparing values. Preserve legal name, trade name and any supplied identifier separately. Similar names, a common phone number or a shared address are not sufficient grounds to combine different entities. Our [legal-name and DBA matching guide](https://usdotdata.com/blog/match-carrier-legal-name-dba-usdot) explains that separate identity task. If one record describes a parent company and another describes a particular operating entity, stop the field comparison. Mark it “entity scope unresolved.” A difference between those records may be a difference in scope, not a change over time. For each matched entity, also check whether you are comparing a current profile, a historical snapshot, a provider's calculated assessment or a manually entered note. These are different kinds of evidence. ## Keep the dates in separate columns Use the dates actually supplied by each source, with the original label: - **Observed or retrieved at:** when you accessed the page or file. - **Source snapshot date:** the dataset date, when provided. - **Filing or form date:** the date associated with a particular filing. - **Event or effective date:** when a specific status or action takes effect, if shown. - **Coverage window:** the period summarized by a count or percentage. SAFER identifies the MCS-150 form date separately from the form's mileage and describes inspection totals using a rolling historical window. These labels do not mean that every field on the page shares one update date. Consult the [official field definitions](https://safer.fmcsa.dot.gov/saferhelp.aspx) for the field you are using. FMCSA's [Open Data Program](https://www.fmcsa.dot.gov/registration/fmcsa-data-dissemination-program) distinguishes datasets and publication schedules. A provider retrieval timestamp should not silently replace the underlying source date. If that date is unavailable, write “source date unknown”; do not derive it from a filename or the day you paid for an export. ## Classify the difference before interpreting it Use five practical comparison outcomes: 1. **Same value, different presentation.** Formatting, punctuation or whitespace differs, while meaning is unchanged. Preserve raw text and use a normalized working value. 2. **Different field definition.** A mailing address is compared with a physical address, or drivers with power units. These are not like-for-like measurements. 3. **Different source age.** One observation contains an older filing or snapshot. This can explain disagreement without establishing when the business changed. 4. **Comparable values differ.** Entity, field definition and units match, and both dates are known. Record an observed difference; the cause still requires evidence. 5. **Unresolved conflict.** A date, definition, identity match or corroborating source is missing. Retain the uncertainty rather than choosing the more convenient value. This classification is a research method, not a proprietary reliability score. It prevents an apparent change from becoming an unsupported claim about growth, fraud or buying intent. ## Worked example: a fresh export with an older filing The following fictional example uses no real carrier. A researcher reviews two records on September 28: - Observation A was downloaded September 28 and shows 18 reported power units, with a March 15 MCS-150 form date. - Observation B was saved September 20 and shows 24 reported power units, with a September 10 form date. - Both show the same fictional entity identifier. The researcher has confirmed that the two power-unit fields use the same definition. Observation A is the newer download. Observation B contains the later form date. The defensible note is: “The sources disagree; A may contain an older source record. Current reported count requires an official recheck.” It is not “the fleet shrank by six trucks this week.” Suppose the current official record also shows 24 and the September 10 form date. The researcher can record that current official observation and retain A as a stale-source conflict. That establishes the current displayed count at the time checked. It still does not identify which vehicles changed or prove a purchase project. Use the [power-unit interpretation guide](https://usdotdata.com/blog/reported-fleet-size-power-units-prospecting) for those limits. Now suppose both snapshots instead show the same count but their inspection totals differ. Before interpreting a change, compare the displayed reporting windows and retrieval dates. A total over a moving window is not a fixed lifetime tally; subtracting two totals alone does not count all new events. ## Use a field-level reconciliation worksheet Copy this block once for each material disagreement. Separate blocks are easier to read than a wide comparison table on a phone. ```text Decision this field supports: Entity identifier and scope: Field name, definition and units: Observation A value: A source URL/file reference: A retrieved at / source date / field date: Observation B value: B source URL/file reference: B retrieved at / source date / field date: Comparison class: Current authoritative check and timestamp: Accepted working value, or unresolved: Reason for decision: What this does NOT establish: Research owner and next recheck trigger: ``` Retain the actual source labels. A field called “updated” without a definition is not enough to establish whether a filing, a database row or a page changed. Record that ambiguity and ask for clarification when it affects the decision. ## Resolve current questions without rewriting history For current identity or profile facts, begin with the [SAFER Company Snapshot](https://safer.fmcsa.dot.gov/CompanySnapshot.aspx). For authority and insurance questions, follow the [current official-system checking workflow](https://usdotdata.com/blog/fmcsa-insurance-operating-authority-lookup-motus). Select the source for the question; one screen is not authority for every field. Record a newly checked value as another observation. Do not overwrite the old observation with today's value and leave its earlier date attached. That would create a false historical record. If the conflict affects only a prospecting preference, place the account in a research queue until the missing fact is resolved. If it affects a real shipment, escalate it to the responsible verification decision-maker. A salesperson should not turn a research spreadsheet into transport clearance. ## Checklist before accepting a reconciled value - Both observations belong to the same entity and business scope. - Field definitions and units are compatible. - Retrieval dates are separate from source and event dates. - Missing dates remain explicitly unknown. - Raw observations remain available and unchanged. - The accepted value has an identified source and check time. - Any inferred explanation is labeled as a hypothesis. - The decision owner and next trigger are recorded. For a larger file, the [lead-export evaluation guide](https://usdotdata.com/blog/evaluate-trucking-lead-list-export) covers the broader quality audit. This worksheet handles individual disagreements after that audit finds them. ## Frequently asked questions ### Should I always use the most recently downloaded record? No. Download time proves when you obtained it. Compare the field's source date and definition, then check the appropriate current source if the difference matters. ### Does a newer MCS-150 date make every value on the page current? No. Treat it as the displayed form date. Other data can have different reporting periods or effective dates. Keep each date connected to the fact it actually describes. ### Can I calculate fleet growth from two exports? You can calculate a difference between comparable reported counts with known dates. Label it as a reported-count difference. Real fleet growth, ownership changes and purchase needs require additional evidence. ### What should I do when one source has no date? Preserve its retrieval time and mark the source date unknown. Avoid a historical-change claim based on that pair. Obtain dated evidence or keep the comparison unresolved. ### Should I delete the record that appears wrong? Preserve the original observation and document why it was superseded for the current decision. Remove it from active use when appropriate, but do not erase the evidence trail needed to understand the correction. ### Does matching information in two websites count as independent confirmation? Not necessarily. They may use the same underlying dataset. Record their source relationships when known and seek evidence appropriate to the decision, such as the official record or a separately verified business confirmation.