Most explanations of a rising commercial bill start with usage. This one cannot, because usage went down. A distillery in ComEd territory, two meters, three years of 30-minute data, used less electricity in the sixteen months from August 2024 than in the nineteen months before, and paid for nearly twice the demand. Nothing on the kilowatt-hour line would ever have shown it. The month-by-month record is below, with what the meter can say about what changed and what only the operator can.
The before and after
The billed demand here is ComEd's rule, the highest site-coincident half hour between 9 AM and 6 PM on a weekday, applied to the meter data. For this site the rule was checked against a real bill for September 2025 and reproduced the billed 40.91 kW and the $596.88 Distribution Facilities Charge to the cent, so the reconstruction is the one the bill uses.
| period | months | average monthly usage | average billed demand | typical load factor |
|---|---|---|---|---|
| January 2023 to July 2024 | 19 | 3,851 kWh | 26.3 kW | 21 percent |
| August 2024 to November 2025 | 16 | 3,254 kWh | 46.5 kW | 10 percent |
Usage fell 15 percent. Billed demand rose 77 percent. Load factor, the ratio of average to peak demand, halved, which is the arithmetic signature of a building that started spiking harder while running about the same.
The same months, a year apart
Averages can hide a seasonal shift, so here are the four months where the change first appears against the same months the year before.
| month | 2023 billed kW | 2023 kWh | 2024 billed kW | 2024 kWh |
|---|---|---|---|---|
| August | 28.6 | 4,955 | 62.9 | 3,619 |
| September | 27.0 | 4,161 | 54.0 | 3,142 |
| October | 26.8 | 3,716 | 50.7 | 2,566 |
| November | 28.4 | 3,019 | 40.1 | 2,211 |
August 2024 billed 2.2 times the demand of August 2023 on 27 percent less energy. September and October, twice the demand on a quarter to a third less energy. The new level held: from August 2024 through November 2025 the billed demand never dropped back below 31 kW, and reached 64.8 kW in May 2025, against a 2023 range of 19 to 34 kW.
What the meter can say, and what it cannot
The meter can say when. The first month at the new level is August 2024, and its peak, 62.9 kW, landed on a Friday morning at 10 AM, the same hour this building had always peaked. It can say the shape did not change, only the height: the peak stayed in the morning and midday hours, and the overnight floor stayed near 1 to 2 kW. And it can say that the extra demand was not extra work, because the energy went down.
The meter cannot say what was installed or changed. A larger still, a new chiller, a second piece of equipment now starting in the same half hour as the first, a control that no longer staggers loads: each would produce this signature, and the interval data does not distinguish them. That is an operator question, and on this site it has not been asked yet. What the data does is narrow it to one question with a date on it, which is a different thing from a bill that just went up.
Why the bill would not have explained it
A commercial bill prints the billed kW and the kWh. It does not print when the kW happened, and it does not compare either to last year. A business owner reading the August 2024 bill would have seen a demand line about twice the previous month's and a usage line lower than usual, and would have had no way to tell whether the demand figure was a one-off or a new normal. It took fifteen more months of the same to settle that, and nobody was watching.
What a kilowatt of billed demand is worth at this site is on the one bill we reconciled: September 2025, 40.91 kW at $14.59 per kW, $596.88, which was 60 percent of that month's bill. We have not priced the 2024 months, because the 2024 rate for this class is not on file and we do not estimate rates. The direction is not in doubt. A demand charge that rose 77 percent while usage fell is a building paying more for less, and the only line on the bill that would have shown it is the one nobody reads against last year.
This is one of two ways a business bill jumps without the business changing. The other is the opposite shape: the overnight floor rises and the energy line doubles while the demand line moves only by the size of the new load.
Can a demand charge go up when electricity usage goes down?
Yes. The demand charge is set by the single highest half hour of the month, not by total energy. A building that uses less overall but starts more equipment at the same moment will bill higher demand on lower usage. This site did exactly that for sixteen consecutive months.
How do I know if my demand charge jumped because of a new normal or one bad day?
The bill cannot tell you; it shows one month. Interval data can, because it shows whether the peak repeats. Here the new level repeated every month from August 2024 onward, which is the signature of a change in the building, not a bad day.
What usually causes billed demand to rise with flat usage?
Loads that start together. New equipment that starts alongside existing equipment, a control change that stopped staggering starts, or a process that now runs its heaviest steps at once. The meter shows the timing; the building shows the cause.
Does a lower load factor mean the site is doing something wrong?
No. It means the peak grew relative to the average. Whether that peak is worth attacking depends on what set it, and this is the case where the answer is worth finding, because the same energy is being bought with twice the demand.