Every year the same thing happens. You ask for a rooftop unit replacement. It doesn’t get funded. Fourteen months later the unit fails in August, the emergency replacement costs 40% more than the planned one would have, and somebody asks why this wasn’t anticipated.

It was anticipated. The problem is that it was anticipated in a form nobody could act on.

What a losing request looks like

RTU-3 is at the end of its useful life and should be replaced. Estimated cost $38,000.

Read that the way a finance director reads it. It contains one fact — the price — and one assertion offered without support. “End of useful life” is a phrase you used, not a thing they can verify. The unit is currently running. From where they sit, this is a facilities preference competing against a roof repair, a bus, and a salary line, all of which are described the same way.

It’s not that they don’t believe you. It’s that you gave them nothing to defend the decision with.

The four numbers that change the conversation

All four come out of records you already have, assuming your work orders name the equipment and carry their costs.

1. Cumulative repair spend. Every dollar of labor, parts, and vendor invoice charged to that asset since you started tracking. Not an estimate — the actual total.
2. Repair frequency over time. How many service events per year, year over year. A unit that went two, then three, then seven is telling a story that a total alone doesn’t.
3. Age against expected service life. Manufacturer or ASHRAE reference life versus installed age. This is where “end of useful life” stops being your opinion and becomes an external benchmark.
4. Consequence of failure. What stops working, for how long, affecting how many people. Not drama — specifics. Which rooms, which operations, what the emergency replacement lead time is in July.

The same request, rebuilt

RTU-3 (installed 2007, 19 years old, 15–20 year expected life) has cost $21,400 in repairs over four years. Service events went from 2 in 2023 to 7 in 2026, and the last three were compressor-related. Planned replacement is $38,000. Emergency replacement in cooling season runs 30–50% higher with a 6–8 week lead time, during which the fellowship hall and four classrooms have no cooling. We have spent 56% of replacement cost keeping a unit alive that is past its service life.

Same request. Same equipment. Same $38,000. The difference is that the second version can be defended by someone who isn’t you, in a meeting you’re not in — which is where the decision actually gets made.

The number that does the most work

Repair spend as a percentage of replacement cost. It converts your maintenance history into a single figure finance already knows how to interpret.

Common thresholds people use: past 30–40% of replacement cost, replacement deserves serious evaluation. Past 50%, you’re subsidizing a failure. Those aren’t hard rules and they shift by asset class, but they give the conversation a scale, and a scale is what an unsupported request is missing.

If you don’t have the history

Most departments can’t produce these numbers today, and the reason is almost always the same: work orders were never tied to specific equipment, so costs pooled into a general maintenance line where nothing can be attributed.

That’s fixable, and it takes one habit: every job that touches equipment gets linked to that equipment, with its real cost attached. Labor time, parts used, vendor invoice. No exceptions for small jobs — the small jobs are the frequency signal.

Twelve months of that discipline gives you a defensible case on any asset you own. Three years gives you a trend line, which is what turns a capital request into a capital plan.

One more move

Bring the request before you need it. A replacement proposed two years out, with the trend attached, reads as planning. The identical request made in the same month the unit is failing reads as a crisis you didn’t see coming — even when the evidence has been in your files the whole time.

The records aren’t the point. Being able to say what they mean, early, in a room where the money moves — that’s the point.