Getting Work Orders Completed on Site

A maintenance record is only as good as its close-out, and close-out happens on a phone, in a plant room, by someone who wants to get to the next job.

Every design decision should follow from that.

Why close-out fails

The form is too long. Fifteen fields on a phone is a form that gets abandoned or filled with single characters.

It requires a login the technician does not have. External contractors are not going to maintain accounts in your system unless it is trivially easy.

It requires an app install. For a contractor working across many clients, another app is another thing they will not do.

No signal. Plant rooms, roofs and walk-ins have poor connectivity. A form that requires a live connection loses the entry.

It duplicates their own system. A contractor already completing their own paperwork will not do yours twice, and if forced they will do one of them badly.

Nobody checks. Requirements that are not enforced stop being met within a quarter.

What actually works

A link, not an app. A URL that opens a form in a browser, ideally reached by scanning a code on the equipment.

Under two minutes. Which means very few required fields.

Offline tolerant. Captures and submits when signal returns.

Camera-first. Photographs are quicker to provide than typing and more informative.

Prefilled. Asset, site, reported fault, priority. The technician confirms rather than enters.

Voice input tolerated for the free-text fields, because typing in a plant room is unpleasant.

The minimum field set

Ruthlessly short:

What was wrong — free text or a selection from common faults for that asset type.

What was done — the same.

Parts used, with numbers.

Time on site.

Permanent or temporary, as a toggle. If temporary, what is needed.

Photographs.

Anything observed — one free-text field, optional but prompted.

Seven items. Anything beyond this trades completion rate for detail, and an incomplete record with more fields is worse than a complete one with fewer.

Accepting the vendor's system

Where a contractor has their own platform and will not use yours, the options are:

Their report, your record. They send a report; your team enters the essentials. Costs admin time and preserves the history.

Integration, where their system can export or connect. Available with larger vendors, rarely with small ones.

A contractual requirement to complete yours, which works with vendors who want the volume and not with those who do not.

The non-negotiable is that you end up with the history in your own system. A record that lives only in a vendor's platform disappears when the relationship does, and the asset history is the asset.

Verification by the site

The site manager confirms the equipment works before the work order closes.

Separate from the technician's completion. A vendor closing their own work with no site confirmation is a system that reports success.

A simple confirmation — fixed, partially fixed, not fixed — is enough, and it catches the visits where the equipment failed again an hour later.

Enforcing it

Payment conditional on a complete work order. This is the only mechanism that reliably works, and it must be in the agreement.

Monthly completeness report by vendor. Percentage of work orders with the required fields.

Raise it as a pattern, not per incident.

Accept that it will never be perfect. Eighty percent completion with useful content beats a mandatory system at thirty percent, and pushing for total compliance frequently produces forms filled with "repaired" in every field.

The test

Take a work order closed last month and try to answer three questions: what was actually wrong, what was done about it, and would you know if it happened again.

If the record cannot answer those, the close-out process is not working regardless of what the completion rate says.

Testing the close-out flow

Before committing to a system, run the close-out yourself under realistic conditions.

On a phone, not a laptop.

In a plant room or a walk-in, where the signal is poor.

Wearing gloves, which is how technicians work.

Without an account, as an external contractor would.

Time it. Over two minutes and the completion rate will be poor.

Count the taps. More than fifteen and it will be abandoned.

Try it offline and see whether the entry survives.

Then ask an actual contractor to try it and listen to what they say. Their willingness to complete it is the single property that determines whether your asset history has any content, and it cannot be assessed from a demonstration.