The Facilities Director and What They Are Actually Judged On

Why this matters

A facilities director will let you talk about your hourly rate for exactly as long as it takes them to be polite. Nobody has ever been promoted for having paid a low rate, and nobody has been fired for having paid a high one. They get called into a room over three things: something stopped working that should not have, a budget line closed outside its tolerance, or an inspector or an auditor wrote something down. Price only matters to them where it drives one of those three. Once you can see their scoreboard, most of what looks like irrational buying behavior in an institution resolves into something obvious.

The three scoreboards

Unplanned downtime and the complaints that follow it. Not total downtime. Unplanned downtime. A chiller down for four days on a scheduled shutdown is a success. The same chiller down for four hours in September is an incident, because it generated calls from people who outrank the director. In a school those calls come from principals, in a hospital from nurse managers, in a plant from production. The director's currency here is the count of events they did not see coming.

Budget variance, both directions. A line that closes 18 percent over gets a conversation. A line that closes 30 percent under gets a different conversation, and next year that line is cut. Directors are managing to a tolerance band, not to a minimum. This is the single most counterintuitive thing about selling to them, and it is why an unexpectedly cheap invoice sometimes lands worse than a predictable expensive one.

Audit and compliance findings. Backflow test records, fire damper inspections, boiler certificates, refrigerant leak logs, life-safety impairment procedures, infection-control documentation. These are pass or fail, they are checked by someone from outside, and a finding is written down permanently. A vendor whose paperwork arrives late is not a paperwork problem to the director. It is an audit exposure with a due date.

Notice that your unit price appears on none of the three. It appears only inside the second one, as one input to a variance the director is trying to keep bounded.

The gate

Here is the single question worth running every proposal through before you send it:

Does this reduce the number of events the director cannot predict, or does it move that number somewhere they cannot see?

That is the whole gate. It is not "is this cheaper." Predictability is the product. Two bids on the same scope will frequently resolve in opposite directions under it, and the cheaper one is not reliably the loser or the winner. Below are two, on the same building, run against the same gate.

The building both bids are for

A district with 40 rooftop units across four schools. The director's actual numbers from the year just closed, which they will read to you if you ask:

• 62 unplanned service calls across the 40 units.
• 23 of those 62 calls, or 37 percent, were after-hours, billed at a 1.5x rate multiplier.
• The HVAC repair line closed 18 percent over its budgeted amount.
• 9 written complaints reached the superintendent's office about classroom temperature. 7 of the 9 fell in the two weeks after the first hard cold snap; the remaining 2 were spread across the rest of the year.

That complaint clustering is a real signal, not noise, and it is worth reading before either bid. Heating sections sit unused from spring to autumn. The first genuine call for heat is the first time all year that ignition, flame sensing, gas valve operation, and heat exchanger condition are exercised, so accumulated faults surface as a batch on one cold week. The direction check runs the other way too: by midwinter, those same sections have been cycling daily for weeks, faults have already been found and cleared, and the complaint rate should drop back to background. It did, at 2 complaints across roughly the remaining ten months. If midwinter had been the worse period, the cause would not be seasonal changeover at all and would point instead at capacity or distribution.

Bid A: the same work, 20 percent cheaper per hour

Time and materials, no fixed visit schedule, respond when called, labor rate 20 percent below the incumbent's.

Run the gate. Event count: unchanged. Nothing in this proposal touches why the 62 calls happened. Timing: unchanged, so the 37 percent after-hours share stays. Variance: still unbounded, because the line's size is still the product of two numbers the director cannot forecast, an event count and an average event size.

Now do the arithmetic the director will do, and be careful about the base. The 20 percent cut applies to labor, not to parts, and on this repair line labor ran about half the spend. A 20 percent reduction on half the line is a 10 percent reduction in the line. Against an 18 percent overrun, that closes a little over half the gap and leaves the line still over. The director cannot walk into a budget meeting with "it will probably be about 8 percent over instead of 18."

Bid A fails the gate. It reduces the price of surprises without reducing surprises.

Bid B: more per hour, four known dates

Four scheduled visits a year on named dates, all 40 units, with a fall changeover visit deliberately placed ahead of the first cold week. Labor rate roughly 15 percent above Bid A's. Unplanned calls outside the scheduled visits are billed at the same rate, with a written per-event cap above which the director's approval is required before work continues.

Run the same gate. The scheduled portion carries zero variance by construction: four dates, one known amount, booked at the start of the fiscal year. The unplanned portion is still unforecastable in count, but each event is now bounded in size, and the director controls the boundary. And the fall visit is aimed directly at the mechanism that produced 7 of the 9 complaints.

The director's read: the predictable portion of the line is now genuinely predictable, and the unpredictable portion cannot exceed a per-event ceiling without their signature. That is a variance they can defend. It also costs more per hour, and they will not care, because they are not scored on the rate.

Bid B passes the gate.

What the gate does not excuse

Predictability sold and not delivered is worse than an honest time-and-materials relationship, because you will have set an expectation on the one axis they are measured on.

If you commit to four named dates, those dates are now on the director's calendar and probably in a board report. Missing one is not a scheduling inconvenience, it is a variance in the thing you sold. If you commit to a per-event cap, the first time you exceed it without prior written approval you have converted yourself from a predictable vendor into a source of surprise invoices, which is the exact category you were bidding against.

The fall changeover visit itself carries a hazard your proposal created, not one the building presented. Firing a gas-fired heat section for the first time since spring, in a building with students in it, means anyone at the unit wears a personal carbon monoxide monitor, confirms flue draft is established before the section is left running, and does not return the unit to occupied operation until combustion is verified. Scheduling that visit for a school day rather than a break week is a choice you made about somebody else's air.

What flips the answer

A hard mid-year spend freeze. When an institution freezes spending, a committed recurring obligation becomes the wrong shape entirely. A director in a freeze needs the ability to defer, and Bid A's structure is deferrable while Bid B's is not. Under a freeze, the gate's answer inverts and the time-and-materials structure is the better fit. Ask directly whether any freeze or spend restriction is in effect for the current period before you build a recurring proposal.

Equipment past the point where maintenance changes outcomes. If a meaningful share of the 40 units is beyond useful life, a preventive program buys much less predictability than it does on a healthy fleet, and the honest recommendation is a replacement plan on capital money rather than a maintenance program on operating money. See related: Capital Money and Operating Money Are Not the Same Money.

How to verify you read the scoreboard right

Ask three questions and listen for whether the answers are specific:

1. "What did this line do against budget last year, over or under, and by roughly what percentage?" A director who knows this to the percentage manages to variance and will respond to a predictability argument. One who does not know is being managed on complaints alone, and your proposal should lead with response time and event count instead.
2. "How many of last year's calls were after-hours?" The share tells you whether their pain is cost or disruption. A high after-hours share in an occupied building is usually a disruption problem wearing a cost costume.
3. "What did your last inspection or audit flag, and is it closed?" An open finding with a date on it outranks everything else you could offer them this quarter, including the repair you came to talk about.

If all three answers are vague, you are not talking to the person who is judged on any of it, and the proposal you write will be evaluated by someone you have not met.

References

• GASB Statement 34, capital asset reporting for state and local governments, for why public institutions separate recurring maintenance from asset improvement
• 29 CFR 1910.132, personal protective equipment hazard assessment, covering the monitoring and protection selected for a combustion changeover in an occupied space
• See related: The Budget Cycle and Why Timing Can Beat Price; Capital Money and Operating Money Are Not the Same Money; Reading Customer Budget Signals Reference