Shift Coverage Math and Overtime Spend in a 24/7 Central Plant
A 24/7 central plant serving a 900,000 sq ft mixed-use complex requires two engineers on duty on each of three shifts; one of the two must hold a valid high-pressure steam and refrigeration license. The plant is currently staffed by six licensed engineers and two maintenance technicians. Department overtime has reached $186,000 year-to-date against a $70,000 budget, and two engineers logged more than 60 hours in the last payroll week. HR data show each engineer averages 15 vacation days, 11 holidays, 5 training days, and 6% sick time. Base wage is $52/hour with time-and-a-half overtime and a 32% benefits load. Last month a shift ran with a single engineer for four hours when a relief call-out could not be filled; the chief engineer covered it personally and did not record the time. Turnover in the group has been two engineers in eighteen months.
1. Calculate the number of FTE actually required to cover both positions around the clock, using a relief factor. Show the arithmetic and state the staffing gap in FTE and in hours.
Read the worked answer
Step 1 - Hours to cover. Two positions × 24 h × 365 = 17,520 duty hours per year.
Step 2 - Net available hours per FTE. Paid hours 2,080, less:
- Vacation 15 d × 8 h = 120 h
- Holidays 11 d × 8 h = 88 h
- Training 5 d × 8 h = 40 h
- Sick 6% of 2,080 = 125 h
Net available = 2,080 − 373 = 1,707 h per FTE.
Step 3 - Coverage factor. 8,760 ÷ 1,707 = 5.13 FTE per round-the-clock position. Two positions require 2 × 5.13 = 10.3 FTE, so 11 people to allow any cushion.
Step 4 - Gap. 10.3 − 6 = 4.3 FTE short; in hours, 17,520 − (6 × 1,707) = 7,278 uncovered duty hours per year.
Interpretation. That is roughly 14 hours a week of coverage that must come from overtime, a single-engineer shift, or the chief covering unrecorded. The reported $186,000 year-to-date overtime (~$372,000 annualized at 7,278 hours would be far higher) does not match 7,278 hours of premium pay - which tells you the shortfall is partly being absorbed by understaffing and unrecorded time, not by overtime. That is the more dangerous finding: the books say overtime is the problem when the real problem is that the plant is running with one qualified person on duty.
Round up, not down. Fractional bodies cannot be scheduled.
2. Compare the fully loaded annual cost of closing the gap by hiring versus by paying overtime. What costs and risks does the overtime comparison usually leave out?
Read the worked answer
Overtime route (7,278 h):
- Premium wage: 7,278 × $78.00 = $567,684
- Benefits at 32%: ≈ $181,659
- Fully loaded: ≈ $749,300 per year - and rising, because overtime is paid at the top step and breeds fatigue and turnover.
Hire 4.3 FTE (call it 5 hires including a relief engineer):
- 4.3 × 2,080 × $52 = $465,000 base
- Benefits at 32% = $148,800
- Fully loaded: ≈ $613,800 per year, and it buys 4.3 bodies who can also do preventive maintenance, respond to emergencies, and cover for each other.
Hiring is cheaper by roughly $135,000 a year even before counting turnover.
What the overtime comparison usually omits:
- Fatigue error rate. An engineer 60 hours into a week making a burner or chiller decision is a risk, not a savings. Two engineers in eighteen months leaving is the leading indicator.
- Licensure exposure. If the license holder is the one who calls out, the shift has no qualified operator. That is a code and insurance problem, not a payroll problem.
- Deferred maintenance. Overtime hours go to covering shifts, never to PM. The backlog compounds silently.
- Turnover cost. Recruiting, onboarding, and lost institutional knowledge on a 900,000 sq ft plant run $30,000–$60,000 per departure.
- Recordkeeping. Unrecorded time by the chief is a wage-and-hour liability and distorts the very analysis being used to justify the staffing level.
Recommendation: hire to 11, hold a small overtime allowance for genuine emergencies, and stop treating coverage as a variable cost.
3. Identify the supervision and compliance failures visible in the setup, and propose the policy changes you would put in place in the first 60 days.
Read the worked answer
Failures visible in the facts:
- Licensed coverage was not maintained for four hours. Running with one engineer on a high-pressure steam and refrigeration plant breaches the intent of the operating certificate and voids assumptions in the insurance policy. The chief's personal coverage does not cure it.
- Unrecorded time. The chief covering a shift without a record is a wage-and-hour violation, a false staffing record, and it hides the true gap from the very people who approve headcount.
- No relief protocol. A call-out that cannot be filled is a foreseeable event with no escalation path, so the default became "the chief does it."
- 60-hour weeks. Sustained hours at that level impair judgement on equipment that can kill people.
- Overtime managed as a budget line rather than a symptom. A $70,000 budget against a structural coverage gap was never achievable.
60-day actions:
- Adopt a written escalation ladder for every unfilled shift: relief list → cross-trained technician → supervisor → contractor. Every step logged.
- Every hour worked is recorded, including salaried staff, on the time system. No exceptions, stated in writing from the property manager, not the chief.
- Establish a relief pool: cross-train the two maintenance technicians toward the license and license two more engineers to the second ticket so the qualified-operator requirement can be met by more than a handful of people.
- Move to a rotating 12-hour or 4-on/4-off schedule if it produces more predictable coverage than five 8s; model it before committing.
- Post the gap in writing to ownership with the numbers above, so the staffing decision is documented as a management decision rather than an operating failure.