How demand ratchets quietly inflate cold storage bills
One spike can set your billing demand for the next 11 months. How ratchet clauses work, why refrigeration triggers them, and how to find a false one.
Vikram Rao · September 2, 2026 · 7 min read
Key takeaways
- A demand ratchet sets a minimum billing demand as a percentage of the highest demand recorded in a trailing window, often the prior 11 months.
- Compressor restarts after an outage or a defrost overlap can create a one-interval spike that the ratchet then bills for a year.
- The fix is a three-part check: confirm the clause applies, confirm the peak was real, and confirm the meter and multiplier were right.
What a ratchet clause actually does
Demand-metered commercial tariffs charge for the highest rate of power draw in a month, measured in kilowatts over a short interval, usually 15 minutes. That is the demand charge. A ratchet adds a floor. Billing demand for the month becomes the greater of the actual demand this month or a stated percentage of the highest demand set in a trailing period.
The percentage and the window are set by the tariff and differ by utility and rate schedule. A common structure is a percentage between 50% and 90% applied to the highest demand in the prior 11 months. If a site peaked at 1,000 kW in July and the ratchet is 80%, then in January the site is billed for at least 800 kW even if it only drew 400 kW that month.
Why refrigeration sites trip ratchets more than most
Cold storage load is large, motor-driven and cyclical. Compressors, evaporator fans, defrost heaters and dock doors all create inrush and coincident peaks. After a power interruption, a facility that tries to bring every compressor and defrost cycle back at once can set a peak far above anything in normal operation, and it can set it in a single 15-minute interval.
Under a ratchet, that one interval is not a one-month problem. It becomes the reference point for the following 11 bills. A cold storage operator can pay a floor of 70% to 90% of a restart spike for most of a year while the building runs at half that load.
Step one: confirm the ratchet applies to this account
Not every demand-metered schedule has a ratchet, and the ones that do differ in how they define it. Some apply it only to certain seasons. Some apply it to a customer's contract capacity rather than metered history. Some tariffs ratchet on kVA rather than kW. The first check is to read the tariff section for the account's actual schedule and compare it to the billing demand logic printed on the bill.
We see accounts where a ratchet is applied under a schedule that does not contain one, usually because a billing system carried the rule over from a previous rate assignment. We also see ratchets calculated on the wrong window, for example 12 months where the tariff says 11. Either one is a clean, documentable error.
Step two: confirm the peak was real
If the ratchet is valid, the next question is whether the demand reading that set it was valid. Interval data answers this. A genuine operating peak has a shape: load climbs, holds, and falls as equipment cycles. A faulty reading often shows as one interval far above its neighbors, with no matching increase in energy use.
Utilities will generally adjust a demand reading when there is evidence of a metering fault, a utility-side switching event or a billing error. Peaks caused by the customer's own equipment restarting are harder to dispute, which is why the evidence matters. Outage logs from the utility, building automation records of the restart sequence and the interval export belong in the same file.
Step three: check the meter, then the controls
A wrong meter multiplier or current transformer ratio inflates every demand reading by the same factor, and it can be invisible on the bill because the factor is baked in. If a site's demand looks high relative to its connected load and square footage, compare the multiplier on the bill with the nameplate ratios on the meter and instrument transformers.
Once the billing is corrected, the operational fix is worth doing. Staggered compressor restart, defrost scheduling that avoids overlap, and a demand limiting setpoint in the refrigeration controls all reduce the chance of setting a new false peak. The refund covers the past. The controls protect the next 11 months.