After a Collision

A collision produces an insurance claim, possibly a dispute, and occasionally a prosecution. The vehicle recorded something throughout.

What the data can establish

Where the vehicle was and when.

Speed before the event, where the device reads it from the vehicle rather than deriving it from position.

Braking and impact, on units with accelerometers.

Whether the vehicle was stationary, which resolves a substantial share of disputed claims in the fleet's favour.

Fault codes before the event, which occasionally matter — a brake fault raised the previous week is material.

Preserving it

Place a hold immediately, before routine deletion runs.

Which means a hold mechanism must exist, scoped to a vehicle and period, surviving the retention schedule.

Export a copy and record who did, when.

And note the accuracy and the source — read or derived — because an insurer's engineer will ask and an unqualified speed figure will be challenged.

Using it fairly

Tell the driver the data is being used, at the time, not after a decision.

Share the export with them.

A driver who learns afterwards that their data was used against them will object to the handling, and the objection will be sound.

This costs nothing and is the difference between a process and an ambush.

Reading it carefully

Speed derived from position over short intervals carries substantial error.

Impact thresholds produce false positives on potholes and kerbs.

A single position is not a trajectory.

Which means the data supports an investigation rather than concluding one, and presenting it as conclusive is how it gets discredited.

The maintenance question

A collision sometimes has a mechanical contributor, and the vehicle's own history is where that is established.

Which is an argument for keeping condition data longer than position data — it is both less sensitive and more useful after the fact.

Check the fault history and the service record as part of the investigation, not only the driving data.

The measure

Collisions per distance, trended.

Locations that recur, which is a route or site finding.

Whether data was preserved and shared each time, which is a process check.

And whether anything changed as a result, because an investigation producing no change teaches everyone that reporting is pointless.

Keep condition data longer than position

Because it is more useful after the fact and less sensitive.

A brake fault raised the week before a collision is material.

Position two months later is not, and holding it carries obligations.

Which is another argument for splitting the retention, and this is the case where the split pays for itself.

Turn the boundary into a test

For a concrete product reference, this transfer example can help turn this boundary into a test. Verify the current behaviour with representative fields and record what reaches the destination.

Independent reference

For a thematic point of reference, see NHTSA. Its current material provides useful context beyond product documentation.