A facility manager I worked with had 11 years of service tickets on a rooftop unit. Every visit was documented. Every invoice was paid. When the unit failed and the manufacturer asked for evidence of annual maintenance before considering a goodwill claim on the compressor, none of it was usable. The tickets proved the building had been billed. They did not prove the unit had been maintained.

That distinction is the whole problem, and it is not a paperwork failure. It is a design outcome.

A Service Ticket is an Accounts Receivable Instrument

Before I started my current company I ran a mechanical contracting business. The work order we used was not designed by a technician or by anyone responsible for an asset. It came out of the accounting software, and every field on it existed to answer one question: can this be invoiced, and will the invoice survive a dispute?

That question is answered by the date, the address, the labor hours, the parts consumed, and a description of the work sufficient for the customer to recognize it. Once those fields are complete, the ticket has done its job. Everything else is optional, and in the field, optional means it usually does not get filled in.

Look at your own tickets through that lens and the omissions become predictable:

The model and serial number. Nobody needs it to bill a repair. It is needed for warranty claims, parts sourcing, recall checks, and any replacement forecast more considered than a guess from the install year. On most tickets it appears once, on the original install paperwork, and never again.

What was measured, as opposed to what was concluded. A technician who writes that a capacitor was tested and found within tolerance has recorded nothing you can use next year. A technician who writes 41.8 microfarads against a rated 45 has given you a number that becomes a trend the moment it is recorded a second time. Two readings on two visits will show you a component drifting toward its limit. Pass and fail will never show you that, however many times it is entered.

Repaired or replaced. These can cost nearly the same and appear almost identically on an invoice, and they have entirely different consequences for remaining service life. A contactor that was cleaned is not a contactor that was changed. Twelve months later, the ticket reads the same either way.

Found, as distinct from done. Tickets describe actions, because actions are what get billed. They rarely describe the condition of anything the technician looked at and left alone. An inspected heat exchanger found in acceptable condition is a fact worth having, and it costs nothing to capture at the moment somebody is standing in front of it.

The Technician is Not the Problem

It is tempting to read that list as evidence that trades cut corners. That is not what is happening.

A service technician is measured on calls completed in a day. Time spent writing is time not billed, and under most compensation structures it is time not paid. The form on the tablet asks for what the office needs in order to invoice. Asking a technician to record more, without changing the form or the schedule, is asking for unpaid administrative work. They will do it for one visit because you asked politely. They will have stopped by the fourth.

The contractor loses from this too, which is the part that rarely gets said out loud. A maintenance contractor whose tickets record only actions has no evidence that their program worked. At renewal, when a lower bid arrives, they cannot demonstrate that four years of preventative service kept a unit out of failure, because nothing in their own records shows the condition trend that would prove it. They end up competing on price because they have no data with which to compete on outcome.

Both parties want the same thing. The document standing between them was designed for neither.

What To Specify, Inside the Agreement You Already Have

The useful part is that this is contractual rather than technological. Maintenance agreements come up for renewal on a regular cycle, and the service ticket is a deliverable under that agreement like any other. A handful of clauses change what arrives.

Require nameplate capture once per unit. Make, model, serial, and install date for every piece of equipment under the agreement, recorded on the first visit and carried forward from there. This is a one-time cost per asset rather than a recurring burden, and it is the highest value field on the list.

Require measured values, with units and the rated value beside them. Specify the readings that matter for your equipment class: amperage against nameplate, capacitance against rated, static pressure, temperature split, operating pressures. Exclude pass and fail as a standalone entry.

Require an explicit repair or replace designation, with part numbers. One additional field, no additional site time, and it is the difference between knowing remaining life and estimating it.

Require condition on inspected components, not only on changed ones. A short list is enough. What was looked at, and what it looked like.

Specify the format, and who owns it. Ask for the data in a structured export you can keep, and state in the agreement that the record belongs to the building. Records that live only inside a contractor’s system leave when the contractor does, and contractors rotate.

Then do two things that determine whether any of it holds. Price the first visit capture as a line item, so the contractor is paid for the additional field time rather than absorbing it. And make ticket completeness a condition of invoice approval, because that is the only enforcement mechanism in this relationship that has ever reliably worked.

The Ticket Will Still Be a Billing Document

None of this asks the industry to redesign the work order, and none of it asks your contractor to adopt anything new. The invoice will go on doing what it was built to do. What changes is that the asset requirements get stated alongside the billing requirements, in a document both parties already sign, so that the information you will need in year six is captured in year one by somebody who was already standing in front of the equipment.

The cheapest data you will ever collect is the data recorded by a technician who is on site anyway. The most expensive is the data you try to reconstruct after the unit has failed.