The Work Order System and Why the Ticket Is Your Only Record

Why this matters

On institutional work the customer's ticket, not your invoice, is the document that authorizes payment. A facilities director cannot pay for work that has no ticket number, because their controls do not permit it, and no amount of proving you did the work fixes that after the fact. Shops lose more money on institutional accounts to work performed outside a ticket than to work performed badly. The interesting part is what the ticket does not do, because that is where every unpaid hour hides.

What the ticket is, in their system and in yours

A work order in a facility's system is an authorization object with a lifecycle. It is opened by a requester, assigned a priority and an account to charge, routed to an in-house shop or to a vendor, worked, closed with a completion record, and then reviewed. Your job number, your dispatch record and your invoice are a parallel set of documents that exist entirely inside your shop. The two are joined by exactly one thing: the ticket number you quote.

That single join is the whole relationship. Everything you can prove downstream, and everything you can bill, hangs off whether that number is on the record at the time the work happened.

The negative space: five things the ticket will not do for you

This is the section worth reading twice, because the ticket is a genuinely useful document and its usefulness makes people assume coverage it does not provide.

It does not expand to cover what you found. A ticket says "no heat, room 214." You arrive, fix the no-heat condition, and while you are in the room you also repair a leaking valve you could not responsibly leave. The valve repair is outside the ticket. It is real work, it was the right call, and there is no authorization object attached to it. In most systems it will be paid only if somebody opens a second ticket, and the person who can do that is usually not the person standing in front of you.

It does not carry your scope, only their symptom. The requester describes what they noticed. They are frequently not a tradesperson and are describing an effect. A ticket reading "smells musty in the east corridor" may be a drain trap, a coil, an envelope leak or a housekeeping issue, and the ticket's text will never grow to say which. Your notes are what add the scope, which is why the notes field matters far beyond the immediate job. See related: What a Maintenance Management System Does With Your Notes.

It does not stop the clock. A ticket left open on your side is a ticket the facility sees as unresolved work in their aging report. Aged open tickets are the single most visible vendor performance metric in most systems, and yours are visible next to the in-house shop's. A ticket you finished but never closed reads identically to one you never started.

It does not survive a change in requester. The nurse manager who called it in transfers. The ticket text remains, the context does not. If the only record of what was agreed lives in a hallway conversation, it is gone.

It does not authorize access. A ticket number gets you paid. It does not get you into the wing, past the badge desk, or onto the roof. Those run on a separate track, and a shop that assumes the ticket is a universal pass will spend its first morning learning otherwise.

Ticket states, and which ones let you invoice

Systems differ in their vocabulary, but the states are functionally the same everywhere, and only one of them is a billing event.

The trap is the gap between Complete and Closed. Your crew considers the job finished at Complete. The facility's payment cycle starts at Closed. Whatever calendar time sits between those two states is time your shop is financing, and it is invisible on your side because your own system shows the job as done.

Priority codes are a promise you are being held to

Most systems carry three to five priority levels, and each has an implied response and completion window written into the vendor agreement even when nobody discusses it. Read them out of the contract and put them somewhere your dispatcher sees, because the measurement runs against the timestamp on the ticket, not against when you actually got the call. If a ticket sat unassigned in their queue for two days before it reached you, many systems still start the clock at creation. Ask which timestamp governs before you sign, because the answer changes what a two-hour emergency response actually requires of you.

The ticket you cannot finish today

Half of institutional work does not close on the first visit, and how you park it decides whether the next visit is paid. Three parking states exist in nearly every system, and they are not interchangeable.

On hold, awaiting parts keeps the ticket open, keeps your labor attached to it, and keeps it aging in the facility's report. It is correct when the remaining work is yours and a delivery date is known. Put the expected date in the ticket, because a hold with no date is the state most likely to be administratively closed out from under you during a queue cleanup.

Awaiting owner decision moves the aging clock off your performance and onto theirs in most systems, which matters when your response metrics are being scored at re-bid time. Use it when the next action genuinely belongs to the facility: a cost approval, an outage window, an access decision. Do not use it as a place to hide your own backlog, because the person who reviews that queue can see who touched it last.

Split to a new ticket is the right move when what you found is a different asset, a different trade, or a different funding source. This is the state that recovers the found-work problem described above, and it is the only one of the three that generates a fresh authorization.

The failure mode is using a single "in progress" state for all three. It produces a queue of long-running tickets that all look like slow work from the outside, and at renewal nobody separates the ticket that waited eleven weeks on a compressor from the ticket that waited eleven weeks on you.

The emergency call inside a bureaucracy

An after-hours emergency is the one case where the ticket comes after the work, and it is the case most likely to end unpaid. The sequence that survives is: take the call, act on the hazard, and get a ticket number retroactively from the person with authority to issue one, in writing, within the same shift.

Where the emergency involves a real hazard, act on the hazard first and document second. If there is a gas odor, everyone leaves immediately, nothing is switched on or off, no lights and no phone used inside, and the call goes out from outside the building. If water is entering an area with energized equipment, keep people back and have the circuit de-energized and locked out under 29 CFR 1910.333(b)(2), proved dead with the live-dead-live sequence in NFPA 70E-2021, 120.5, before anyone approaches to assess. Documentation is a same-shift obligation, not a same-minute one.

Then get the number. "I will sort the paperwork Monday" is where the money goes, because on Monday the person who authorized you verbally is off, and the person at the desk has no record of a call that never generated a ticket.

Worked example: the same quarter, two ways of counting

A shop holds a service agreement with a school district. Over one quarter they run 46 tickets. Their own dispatch system shows 52 site visits.

The 6-visit gap is the whole lesson. Four of those visits were second trips to a ticket that had already been closed, because a closed ticket cannot absorb more labor and nobody reopened it. Two were work found and performed on site with no ticket at all.

Count the hours. The 46 ticketed visits averaged 2.4 hours each, so 110.4 hours are attached to an authorization. The 4 return trips averaged 1.8 hours, so 7.2 hours. The 2 unticketed jobs averaged 3.0 hours, so 6.0 hours. Total unattached labor is 13.2 hours against 110.4 attached, which is about 12 percent of the ticketed labor performed with no path to payment, or 11 percent of the 123.6 total hours worked on the account.

Now the second number, the one that hurts more. Of the 46 tickets, 9 sat in Complete for more than 20 days before the district moved them to Closed. If the shop's terms are net 30 from a valid invoice, and the invoice cannot be raised until Closed, those 9 tickets carry an effective wait of 50 days or more, against the 30 the shop planned its working capital around. Nine of 46 is about 20 percent of the quarter's tickets sitting at roughly 1.7 times the expected wait.

The fix for the first number is procedural: nobody performs work outside a ticket without a number in hand or a written commitment to issue one. The fix for the second is relational: ask the facility what causes verification delay, and the answer in this case was that the district's verifier was a single assistant director who reviewed in batches. The shop moved to submitting completion notes with a photo and the asset tag attached, which let the verifier close from the desk instead of walking the building, and the batch stopped being a bottleneck.

Note what did not change. The shop did not shorten its terms, did not add staff, and did not renegotiate rates. It changed the completeness of a record so somebody else's approval step got cheaper. That is the general form of most institutional cash problems.

How to verify your ticket discipline is real

Once a quarter, pull two counts and compare them: the number of site visits your own system recorded for that account, and the number of tickets the facility's system shows for the same period. They should match within your known count of legitimate multi-visit tickets. Every unexplained visit is either unpaid labor or an unclosed ticket, and both are worth finding while somebody still remembers the job.

Second check: for every ticket you invoiced in the period, confirm the ticket number appears on the invoice exactly as the facility writes it, including any prefix. A transposed or reformatted number is the most common reason an otherwise valid invoice is returned, and a returned invoice restarts the payment clock rather than pausing it.

References

• 29 CFR 1910.333(b)(2), electrical lockout and verification; NFPA 70E-2021, 120.5, live-dead-live proving sequence
• See related: What a Maintenance Management System Does With Your Notes
• See related: How to Close a Work Order So It Survives an Audit
• See related: Asset Tags and Equipment Lists and Why They Matter to You