Demand charges, explained for school facilities teams
Most people read an electric bill as “we used this much, so we owe this much.” For schools on commercial rates, that’s only part of the story. A large share of the bill is set not by how much energy you used, but by the single worst 15 minutes of your month.
The anatomy of a commercial bill
A school’s electric bill typically decomposes into a few distinct kinds of line items:
| Line item | Billed as | What drives it |
|---|---|---|
| Energy | $ per kWh | Total consumption, often varying by time of use |
| Demand | $ per kW | Your highest 15-minute average draw in the billing period |
| Customer charge | $ per day | Fixed — exists to be connected at all |
| Adders and credits | $ totals | Program charges, taxes, riders bundled into the bill |
The demand line is the one that surprises people. The utility looks at every 15-minute window in the month, finds the one where your average draw was highest, and bills you for that peak — every kilowatt of it, at somewhere between roughly $10 and $30+ per kW depending on your tariff and season.
Run the numbers for a mid-sized high school: 100,000 kWh at $0.18 is $18,000 of energy. A 350 kW peak at $25/kW is another $8,750. One 15-minute window cost half as much as the entire month’s consumption.
Why schools are especially exposed
School buildings peak hard because their loads are synchronized. The classic school peak is the first real heat wave of late spring: every rooftop unit is fighting the afternoon sun, the kitchen is finishing lunch service, the gym is conditioning for a rally, and the pool pumps never stopped. Everything stacks between about 1 and 4 p.m., and that stack sets the demand charge for the whole month.
The demand charge is also unforgiving in a way the energy charge isn’t. Using less energy all month helps the energy line proportionally. The demand line doesn’t care about your good behavior for 30 days — it only remembers your worst quarter-hour.
The good news: peaks are schedulable
That same synchronization is why demand charges are one of the most controllable line items on the bill. The peak isn’t caused by how much cooling the building needs in total; it’s caused by everything happening at once. The standard moves:
- Precool. Bring the building down a degree or two in the late morning, before the afternoon peak window, so units can coast through the hottest hours instead of all running flat-out at 2 p.m.
- Stagger starts. After lunch, or after a power blip, don’t let every unit call for cooling simultaneously. Sequenced restarts alone can shave double-digit kW.
- Shift the shiftable. Pool pumps, irrigation, EV charging, and some kitchen prep can move outside the peak window without anyone noticing.
None of this is exotic. The hard part is knowing which afternoon matters.
Where prediction earns its place
Here’s the operational problem: you can’t run the building in peak-shaving mode every day. Precooling and curtailment have costs — comfort risk, staff attention — so you want to spend them only on the handful of days that will actually set the monthly peak.
That’s a forecasting problem, and it’s one models are genuinely good at. Given weather forecasts, the school calendar, and the building’s own history, a model can flag peak-risk days in advance: “tomorrow is likely to be this month’s peak — precool and hold setpoints through 4 p.m.” A morning heads-up on five or six days a month is an intervention a real facilities team can actually execute.
Two honest caveats. First, verify the model against the bill: the demand line item shows the billed kW, so after a month of peak management you can see exactly what you paid for and whether the called peaks matched reality. Second, read your own tariff before investing effort here — demand charges vary a lot between rate schedules, and some districts will get more from fixing after-hours schedules than from shaving peaks. The bill’s line items will tell you which problem you have; start there.