Carrier Record History: What You Knew, and When

Keep the observation separate from the current view

A current-status screen answers what the service now reports. A saved review answers what information the broker considered at a particular time. Updating the screen must not silently replace the evidence attached to an earlier decision.

For each retained observation, save the source identifier, entity or filing key, source date, retrieval timestamp and original values. Keep the interpretation and decision alongside those facts. Retain documents under the brokerage’s access and retention policy; avoid collecting personal information unrelated to the review.

A gap gives you an interval, not an exact event time

Consider this invented sequence: Monday’s successful snapshot contains Address A; Tuesday’s import fails; Wednesday’s successful snapshot contains Address B. You observed a different address on Wednesday. Without a source-reported change date, you cannot establish that the move occurred on Wednesday.

The uncertainty is broader than the missed import: the source may itself publish after the underlying event. Store “first observed in our records” separately from an agency’s effective date. Do not backfill Tuesday with B or label it unchanged simply because the service has no observation.

If a broker reviewed the carrier on Tuesday, preserve which dated evidence was actually available and the known refresh failure. A later successful import should not make the earlier review appear to have used Wednesday’s information.

A full snapshot and a change file preserve different evidence

FMCSA publishes both full MOTUS datasets and daily differences. A retained full snapshot can support a comparison between observations; a difference file describes the changes included in that batch. Neither proves that your archive contains every preceding batch.

For a proposed archive, keep a batch ledger with source URL, retrieval outcome, covered period, row count and a content fingerprint. Record missing batches explicitly. If a new baseline repairs the current view, retain the historical gap: knowing the latest state does not recover every intervening event.

Replaying a batch should preserve one event rather than create another apparent change. Decide whether a record is a revised version of an existing event or a separate event using the source’s identifiers and definitions, not its position in a downloaded file.

Sources: DOT DataHub: Motus Carrier daily differences

Corrections should explain the old evidence

A disputed inspection can be corrected after the brokerage made its original decision. Keep the earlier observed value, the carrier’s dispute, the official outcome when available and the later observation of the changed record. A dispute submission alone does not establish the correction.

Link the new evidence to the affected review so another person can distinguish “the source later changed” from “the original reviewer copied the source incorrectly.” Those are different explanations and may require different follow-up.

DataQs provides a route for reviewing potentially incorrect safety data. The downstream record still needs to be checked after the outcome; preserve the correction’s provenance rather than manually erasing the earlier observation.

Sources: DataQs help center

Test whether another broker can reconstruct the decision

Give a reviewer the load reference, approval date and retained evidence. They should be able to identify the carrier, the source versions, unresolved gaps and the decision conditions without consulting today’s profile.

For a later alert, add a new review entry that points back to the relevant evidence. Do not imply that an alert’s arrival time is the legal effective date of the underlying action. Preserve both so operations can understand what changed and when it matters.

This is a recommended evidence design, not a claim that a particular product retains a complete historical archive. Ask an archive provider for its coverage start, missing periods, correction policy and exportable provenance before relying on a reconstructed history.