Minimum evidence set
Minimum evidence set explained: how much evidence is enough
A minimum evidence set is the smallest defined group of images, measurements and readings that must be present before an inspection record counts as complete, chosen so the record can still answer the questions asked of it months later by someone who was never there.
Why the size of the set is a real decision
It cuts both ways, and teams usually only feel one edge of it.
Set it too small and the record cannot answer the question that gets asked. A single close-up of a scratched wing proves a scratch existed on something. Six months into a damage recharge dispute, nobody can tie it to that vehicle, that day, or that side of it, and the recharge quietly gets written off.
Set it too large and something worse happens, because the failure is invisible. Twenty two mandatory shots on a five minute handover means the person on site starts gaming it. Same photo submitted twice. A tyre shot taken from four feet away because kneeling down takes too long. Fields ticked without looking. You do not get more evidence from a longer list. Past a certain length you get worse evidence, collected less honestly, and a completion rate that quietly falls off. The organisation reads the compliance dashboard, sees 94 percent completion, and has no idea the last eight items are noise.
So the set is not a wishlist trimmed for practicality. It is the point where marginal evidentiary value stops exceeding marginal completion damage, and that point is a lot earlier than most templates assume.
How to work out the minimum
Work backwards from the questions the record must survive, not forwards from what the camera could capture. In practice five questions cover most asset inspections.
- Identity. Which specific asset is this, provable from the evidence alone rather than from the file name.
- Coverage. Can a reader see that the whole area in scope was looked at, including the parts with nothing wrong.
- Condition. The specific findings, close enough to assess and with something in frame giving scale.
- State. The readings that fix the asset in time: hours, mileage, levels, counters.
- Attestation. Who says this is the record, and what they are asserting.
Every candidate item has to be defended against one of those five. If it does not answer one, it is optional, and optional items belong outside the mandatory set where they cannot hurt completion.
A minimum evidence set for a van coming off hire
Take a light commercial vehicle returning at the end of a hire. A defensible minimum is eight artefacts, and a competent driver gets through it in about four minutes.
- 1. Registration plate readable in one frame, with part of the vehicle body visible in the same shot so the plate cannot be a different van.
- 2. Dashboard with ignition on: odometer reading, fuel or charge level, and any warning lights, in a single image.
- 3 to 6. Four corner shots at 45 degrees. Four images cover all body panels with overlap, which two side-on shots do not.
- 7. Load area, wide, doors open.
- 8. For each damage item found, a close-up with a scale reference and one wide shot placing it on the panel. Two frames, and no damage claim without both.
What is deliberately not in there: individual tyre tread photographs, the roof, under the bumpers, the service book, a walkaround video. Every one of those has been argued for in some fleet somewhere. Each is worth adding only if your actual dispute history says it earns its place, which is a question you can answer from your own files rather than from a template. The output feeds the check out report and the off hire record.
Small mandatory sets outperform long ones
The clearest evidence for this comes from outside inspection entirely. The World Health Organization's surgical safety checklist is 19 items across three pause points. When it was studied across eight hospitals worldwide, the inpatient death rate after major surgery fell from 1.5 percent to 0.8 percent and major complications fell from 11 percent to 7 percent (Haynes et al., New England Journal of Medicine, 2009). Nineteen items, in an environment with vastly more that could be checked, because a set people complete every time beats a set they complete when there is time.
Regulated inspection minimums are built the same way. ISO/IEC 17020:2012 specifies seven mandatory elements for an inspection report, and Schedule 1 of the UK Lifting Operations and Lifting Equipment Regulations 1998 lists a similarly tight set for a report of thorough examination. Both are floors, deliberately short, with everything else left to the inspecting body. ISO/IEC 17020 was revised in 2026, so check which edition an accredited body is assessed against before quoting a clause number.
How to test whether yours is the right size
Two exercises, both using data you already hold.
First, pull ten disputes from the last year that went against you and check, item by item, whether your current minimum set would have answered them. The items that would have made a difference are your real minimum. Second, check which mandatory fields nobody has opened in twelve months. An item that is captured every time and read never is pure completion tax, and cutting it makes the rest of the set more likely to be done properly.
Then write the set down as a rule with a defined severity rating attached to findings, rather than leaving it as local custom. A minimum that lives in people's heads is not a minimum.
Building the set into the capture itself is what stops it drifting. Venta Capture, a product of VentaVid, lets the organisation define the required shots and questions in the flow, so the person on site is guided step by step and the submission cannot be completed with the mandatory items missing, which is a more reliable control than asking people to remember a list. The broader question of what that material has to withstand is covered under inspection evidence.
