Billed for Your Worst 15 Minutes: How Summer Demand Charges Inflate Cooling Costs

The two numbers on every commercial power bill

A homeowner's electric bill is simple: it counts kilowatt-hours and multiplies by a rate. A commercial demand-metered account is billed twice over — once for how much electricity it used, and once for how fast it used it at the building's worst moment.

That second number is the demand charge. It is set by the highest average kilowatt draw across any short interval — typically 15 minutes — in the entire billing cycle. Hit 220 kW for one 15-minute stretch on a single August afternoon, and you pay the demand rate on all 220 kW, even if the building spent the rest of the month humming along at 120. You are not billed for the month. You are billed for the spike.

For a typical mid-size South Florida commercial property, the demand charge runs 30 to 50% of the total monthly bill. And most property managers never look at it. It sits below the energy charge, in kilowatts instead of the kilowatt-hours everyone recognizes, and it quietly does half the damage. It is the most overlooked — and most controllable — line on the invoice.

Why HVAC owns your peak

In a cooling-dominated climate, air conditioning is the single largest electrical load in almost every commercial building — frequently 40% or more of summer electricity. It is also the load most likely to spike. Lighting draws a steady, predictable amount. Plug loads drift gently. But HVAC compressors, chiller plants, and fan motors pull their heaviest current the instant they start, and again whenever they run flat-out against a hot afternoon.

The classic peak-setting event is the morning startup stampede. The building sits in overnight setback. At 6:30 a.m. the building automation system releases every zone at once. Every rooftop unit and the chiller energize within the same few minutes — into a building that has drifted warm overnight — all pulling startup current simultaneously. Add the lights and elevators coming online with them. The building can set its entire month's demand charge before the first tenant badges in.

The second danger window is mid-afternoon, when outdoor temperature, solar load, and occupancy all peak together and the equipment runs continuously. If anything in the system is undersized, dirty, or low on refrigerant, it runs at maximum draw for hours — and any one of those hours can contain the 15 minutes that defines your bill.

The detail that makes it worse: the ratchet

Many commercial rate schedules include a demand ratchet. Under a ratchet, your billing demand for a given month cannot fall below a set percentage of the highest demand you have recorded over the previous several months — often an 11- or 12-month look-back.

In plain terms: one bad 15 minutes in August can keep inflating your bill through the fall and into winter, long after the heat is gone. The worst quarter-hour of the year can cost you for the next eleven months. Ratchet provisions vary by utility and rate class, so check your specific tariff — but if yours has one, peak control stops being a summer project and becomes a year-round one.

The six fixes that pull the peak down

No single upgrade solves this. Demand is a system behavior, not a component, and pulling it down is a sequence of moves — most of them low-capital or no-capital.

1. Stagger equipment startup. Replace the simultaneous morning release with a staged or optimal-start sequence. Optimal start learns how long each zone actually needs to reach setpoint and brings units online in a rolling order rather than all at once. The building still reaches comfort by the time tenants arrive — but the combined draw is spread across 45 minutes instead of stacked into 5. This alone can shave a meaningful slice off the monthly peak, and it costs nothing but controls programming.
2. Tune the building automation schedule. Most BAS schedules are set once at commissioning and never revisited. Units run full-bore on holidays, on weekends, and in suites that emptied two tenants ago. Tightening schedules to actual occupancy, widening temperature deadbands a degree or two, and enabling genuine night and weekend setback removes runtime that produces no value — and every hour a compressor does not run is an hour it cannot set a peak.
3. Restore lost capacity with maintenance. A coil caked with dirt, a clogged filter bank, a refrigerant charge that has drifted low, a failing capacitor — none of these stop the building from cooling. They make it cool inefficiently. The equipment simply runs longer and pulls more amps to deliver the same result. Clean coils, correct charge, and healthy electrical components mean the system reaches setpoint faster and at a lower draw. Maintenance is not just reliability insurance; it is direct demand reduction.
4. Control outdoor air and ventilation. Every cubic foot of hot, humid South Florida air pulled through the outdoor-air damper has to be cooled and dehumidified. Many systems bring in far more than code requires because the dampers were set conservatively at startup and forgotten. Demand-controlled ventilation modulates outdoor air to actual occupancy using CO₂ sensors, and on milder shoulder-season mornings a working economizer can cool the building with outdoor air and the compressors off entirely.
5. Pre-cool ahead of the peak. The building structure itself stores cooling. By cooling the space a degree or two below setpoint before the hottest part of the afternoon, then easing equipment back as outdoor temperature climbs, the building coasts on its own thermal mass and the HVAC peak shifts out of the most expensive window. This is a controls strategy, not an equipment purchase.
6. Upgrade to variable-speed equipment. When equipment reaches end of life, the structural fix is variable speed: VFDs on fans and pumps, and inverter-driven compressors. A constant-speed motor draws nearly full power whether the building needs 100% of capacity or 60%. A variable-speed motor's draw falls roughly with the cube of its speed — running at 80% speed uses about half the power. Across a plant, that reshapes the entire demand profile. It is a capital decision, but it is the one that permanently lowers the floor.

What the math looks like

Consider a mid-size South Florida office or retail property billed on a commercial demand rate. Say it sets a summer billing demand of 200 kW, and the demand charge falls in the range of $10 to $13 per kW — check your own tariff, since rates vary by class and change over time. That is roughly $2,000 to $2,600 every month in demand charges alone, before a single kilowatt-hour of energy consumption is counted. Over a year, $24,000 to $31,000.

Now trim that peak by 15% — a realistic result from staged startup, schedule tuning, and restored maintenance, with no major capital spend. That is 30 kW off the top: roughly $300 to $390 a month, or $3,600 to $4,700 a year. If the rate schedule carries a ratchet, the savings compound, because you have also lowered the floor every future month is measured against. A controls-and-maintenance program that costs a fraction of that figure pays for itself inside the first cooling season.

The same logic scales up. For high-rise and multi-property portfolios, demand management belongs in the capital plan alongside equipment replacement — a connection we cover in our high-rise capital-planning primer.

Where ABC Mechanical fits

ABC Mechanical works with South Florida property managers and facility directors to do exactly this. A commercial maintenance agreement keeps coils, charge, and electrical components in the condition that keeps draw low. Our controls and building-automation work addresses the startup sequencing, schedules, and ventilation strategy that shape the peak. And when equipment reaches replacement age, we plan variable-speed upgrades around your demand profile — not just your tonnage.

It starts with something simple: reading one summer bill together and finding where your peak is really being set. From there, see the full scope of our commercial HVAC services across South Florida — Fort Lauderdale, Miami, and the Palm Beaches.