Can ELD Data Be Used as Proof of Detention?

ELD data proves your truck was at the facility and for how long. It does not prove the driver checked in at the guard shack, was staged at an assigned door, or was ready to load. Detention is owed for the second set of things.

That gap is why carriers who assume their ELD already contains the evidence keep getting denied. The data is objective and accurate and it still loses arguments, because it answers a question the broker isn’t asking.

What an ELD actually records

Two separate streams, and they get conflated:

Duty status. On-duty not driving, off-duty, sleeper, driving. This is the regulatory record, and a dock wait logged as on-duty not driving is meaningful evidence that the driver was working, not resting.

Position and motion. GPS coordinates sampled at intervals, plus engine and movement events. This is what generates a geofence entry and exit, and it’s what most detention tooling runs on.

What neither stream records: whether anyone at the facility acknowledged the driver’s arrival, when a door was assigned, whether paperwork was in order, or whether the driver was in the cab asleep in the staging lot. The ELD knows where the truck was. It knows nothing about the transaction.

Raw ELD dwell is not billable detention

This is the part that trips people up, and it’s worth being blunt about because it cuts against the pitch you’ll hear from software vendors.

Four different clocks could start at a facility: geofence entry, arrival at the gate, check-in at the guard shack, and dock assignment. Your ELD gives you the earliest of these. Free time typically runs from the scheduled appointment or from check-in, neither of which the ELD sees.

So a truck showing five hours inside a geofence has not necessarily accrued three hours of billable detention. It might have arrived ninety minutes early, sat in a staging lot before checking in, or been released and stayed for a break. Bill the raw figure and you’ll get denied on the first one a broker actually reads, and you’ll have taught them to scrutinize the rest.

Where geofences go wrong

A geofence is a circle or polygon someone drew, and its accuracy depends entirely on how it was drawn:

• Too tight and you miss the staging lot, the queue on the access road, or the overflow yard, which is often exactly where the waiting happens.
• Too wide and you catch the truck stop across the road, so a driver’s legitimate break registers as facility dwell.
• Large sites with yards are the hardest case. A truck parked in a yard 400 metres from the dock is inside the fence and not remotely being loaded.

Position is also sampled rather than continuous, so an arrival timestamp is only as precise as the sampling interval. For a two-hour free time window that’s immaterial. For an argument about whether the driver arrived before or after the appointment time, it isn’t.

None of this makes the data bad. It makes it data that needs interpretation, and the interpretation is where claims are won or lost.

Why brokers discount it even when it’s right

The broker’s objection is structural rather than technical. They need to hand the shipper something the shipper will accept, and the shipper trusts their own gate log and their own clerk’s signature. Your telematics is your record. The bill of lading is a shared one.

But it isn’t a flat refusal of electronic evidence, and the published schedules are more encouraging than the folklore. Campbell Soup’s brokerage exhibit requires a statement of record showing check-in and check-out times, and accepts either an EDI 214 status message or a signed BOL annotated with those times, depending on the facility. So at least one major shipper treats an electronic status update as equivalent to paper.

Where ELD data is decisive

Three situations where it carries real weight:

When the facility refused to sign. This is the strongest evidentiary use. An annotated BOL noting the refusal, a notice sent during the wait, and an ELD record confirming the timeline is a considerably stronger package than any one of those alone. The ELD isn’t the proof here; it’s what makes the driver’s handwritten times credible — and it is often what turns a denial into a resubmission.

When the dispute is about duration rather than existence. If the broker accepts the wait happened but argues about how long, position data settles it.

When nobody is arguing at all. The highest-value use of ELD dwell data has nothing to do with claims. It tells you which receivers, which lanes, and which customers are costing you hours, and that changes what you quote. No broker has to accept anything for that to be worth money.

The mandate changed the economics, not just the record

The US mandate took effect in December 2017. Canada’s came into force on 12 June 2021 and reached full enforcement on 1 January 2023, with Quebec following that June. Before those dates, dock time recorded on a paper log did not consume the duty clock the way it does now. The delay still happened. It just didn’t cost the driver the same way.

Electronic logs ended that. Dock time now lands on the duty clock automatically, and the daily window can’t be paused for it. So the same regulation that made your dwell data auditable also made every hour of it expensive.

ATRI’s 2024 telematics work found that trucks drive measurably faster both before and after a detained trip. That’s the compressed schedule showing up in the driving, and it’s the safety argument that sits underneath the billing one.

What makes an ELD-based claim actually stick

• Duty status showing on-duty not driving across the wait, not just position data
• Geofence entry and exit times, with the fence drawn to match where trucks actually wait at that site
• A real-time notice sent during the wait that references the check-in time, not the geofence time
• A BOL with facility times, or annotated with the refusal noted
• Any facility-issued paper: gate ticket, door assignment, lumper receipt

The ELD record is the spine that holds the other four together. On its own it’s rarely enough.

What to do next

Pull one recent denied claim and check whether your ELD timeline matches your check-in time or your geofence entry time. If those two are far apart, that difference is either costing you claims or inflating ones you shouldn’t be filing, and it’s worth knowing which before you send another invoice.

Then check how your geofences are drawn at your five worst receivers. Most detention tooling ships with default radii that were never set for a site with a staging yard.

Reading the ELD stream, working out where the billable clock actually starts, and having the packet assembled before the notice window closes is the whole problem. It is why we built DwellWatch on top of Geotab, Motive, Samsara and the portals built on them rather than asking anyone to change how they log.

And before you argue a rate, check what normal actually looks like.