Daily vehicle checks drivers actually do

Ask a workshop manager what the daily check sheets tell them and you tend to get the same answer: not much. The forms come back complete. The boxes are ticked. And the defect that eventually grounded a vehicle was almost never on one of them.

This is not a story about careless drivers. A check that takes eight minutes on paper gets done properly in week one, quickly in week two, and in the cab from memory in week three — because the driver is already late, the list has thirty-one items, and twenty-nine of them have been fine for a year. The trade calls the result pencil-whipping: filling in the record without doing the work.

What makes that worse than useless is that a check nobody performed leaves the same artefact as one done carefully — a page of ticks. Silence reads as all-clear. So the design goal is not more compliance pressure; it is to make the honest answer the fast one.

Make the list shorter than is comfortable

List length is the biggest lever, and cutting it feels like lowering the bar. It is not; it decides where attention goes. Thirty items completed from memory gather nothing. Twelve completed at the vehicle gather twelve real observations.

You do lose coverage, and it is worth saying so. The items you cut do not disappear — they move to a periodic inspection at a longer interval, where somebody with a pit and a lamp does them properly. A daily walkaround is good at what changes daily: tyres, lights, leaks, load security, coupling, mirrors and glass, warning lamps. It is bad at anything needing tools or a ramp, and those items were never really being checked. They were being ticked.

Then order the list the way a person walks the vehicle — front, nearside, rear, offside, cab. A list ordered by category makes the walker double back, and a walker who doubles back soon stops walking.

Photograph the exception, not the routine

A photo on every item is the eight minutes rebuilt in a new medium. A photo on a fail is different: a defect with a picture is one the workshop can triage without ringing a driver who is by then driving. It answers is this the tyre we changed last month and can it run until Friday — questions that otherwise cost two phone calls and a guess.

The costs are real: photos take a moment, they need uploading, and signal at a depot gate is not always there — which is why they belong on the exception. And a photo is not a description. Keep a short free-text line beside it: a picture of a damaged mudflap does not say which mudflap, and a defect the workshop cannot locate is one it cannot plan.

One report, not two

What actually kills a daily check is not length. It is reporting a defect and watching nothing happen. A driver who reports the same wiper blade three weeks running learns that the report goes nowhere — and after that the list gets ticked and the problem gets mentioned at the depot instead, in a channel with no record, no owner and no deadline.

So the check has to be the workshop’s inbox, not a document somebody later transcribes into it. A failed item should open a defect carrying an owner, a deadline, and a state the driver can see next time. Two things follow.

Severity is decided where the failure is found. A cracked mirror and a brake air leak fail the same tick box and belong in different queues. A system that treats them alike either stops vehicles that need not be stopped, or stops none at all — and once operations tires of the first, it becomes the second.

Deferral needs a name and a date. Most defects are not fix before the vehicle moves, and pretending otherwise is how a check loses credibility. Deferring is legitimate; deferring into silence is not.

Rules differ by country and you should work from your own. The UK’s DVSA guidance on HGV daily walkaround checks is a useful reference for how the record is expected to behave: record and report the defects you find; record nil defects when you find none; report them in writing to the person in charge of vehicle defects; and get dangerous defects fixed before continuing the journey. The nil return matters most: checked, nothing found is an entry, not the absence of one, and it is what makes exceptions legible.

The odometer is the by-product worth the most

Here is the part most often left out, and it quietly pays for the whole exercise. Service cycles come in two kinds. A calendar cycle judges itself — the date arrives whether anyone looks or not. A distance cycle cannot be judged at all without a current reading. If the only moment a reading gets recorded is a workshop visit, the number is fresh exactly when you need it least and stale for the whole interval between visits — which is the entire period during which is this truck due? is a live question.

The daily check is the one moment a human is standing at the vehicle, every day, already entering something. It costs one numeric field. Three details decide whether the number is worth having:

• Readings only go up. Ignore one lower than the last rather than storing it. It is a typo or a stale form, and a decreasing odometer silently resets a service baseline.
• A missing reading is unknown — not zero, not a guess. A distance cycle with nothing to count from is a missing reading, not a clean vehicle. Guessing a baseline manufactures a due date, which either sends a workshop after work nobody can justify, or reads as not due yet on a truck that is.
• Ask once per shift, at the start. Twice a day trains people to type anything.

We looked at how the two halves interact in km and calendar service cycles. The short version: the calendar half works without you, and the distance half is only ever as good as your last reading.

Expect it to be slower than paper on day one

A phone check is not automatically faster than a clipboard. Typing is slower than a tick, and if you sell the change on speed, drivers will time it and you will lose the argument.

What it buys is everything after the tick: no transcription, no lost sheet, no second conversation, and a defect that already has an owner before the driver reaches the gate. Sell it as the last time anyone has to be told — then make that true, because a defect reported in the new system that still has to be raised at the desk costs you the fleet’s cooperation. More on that in getting drivers to use the app.

The specification, on one page

1. Ten to fifteen items, ordered as you walk the vehicle.
2. Three outcomes per item: pass, fail, not applicable today.
3. Photo required on a fail, available otherwise.
4. One free-text line on a fail, so the workshop knows which and where.
5. Severity chosen where the failure is found; critical grounds the vehicle.
6. Odometer once per shift — monotonic, allowed to be unknown.
7. The submitted check is the workshop record. Nothing is re-keyed.
8. Nil returns stored, and last time’s status visible to the driver.

None of this is exotic. It is mostly the discipline of refusing to ask for things you will not act on — which is also what keeps the list short enough to survive week three.

navichain was built around this shape: daily walk-around checks that open a defect when something fails, defects deferred with an owner and a deadline, and a failed inspection that grounds the vehicle, with a critical defect keeping it grounded. Maintenance is scheduled on distance or on the calendar, per maintenance type, so both halves of the interval question get asked, and the fleet view shows which truck is due a service on either. Photos, signatures and failed-delivery reasons are captured in the driver app on the spot, and that app is included with two seats per vehicle rather than priced per device. Every feature is on every plan — you pay for scale, not capabilities — from 995 kr a month; the pricing page has the ladder and the platform page shows what is inside.