Electricity Demand Charge & Ratchet Clause Calculator

What is a demand charge and why is it so large?

Two things drive a commercial electricity bill. Energy charges pay for how much you consumed, in kWh. Demand charges pay for your highest rate of draw, in kW — usually the peak fifteen-minute interval in the month. You pay for the capacity the utility had to stand ready to deliver, whether you used it for fifteen minutes or all month.

Rates run from under $5/kW to over $50/kW depending on the utility territory, with $10 to $25 a common band, and demand charges typically represent 30 to 70% of total commercial electricity costs.

The calculator's defaults model a facility peaking at 720 kW in July on a $15/kW tariff. Annual demand charges come to $104,985 against $172,800 of energy charges — 37.8% of a $277,785 bill, sitting squarely in that band.

Load factor is the diagnostic. Divide annual kWh by peak kW times 8,760 hours and you get how evenly you draw power. The defaults come to 38.05%, meaning average demand is barely over a third of peak. A low load factor is the signature of a facility paying heavily for capacity it uses briefly — and the strongest indicator that peak management will pay.

The uncomfortable part is that a single fifteen-minute interval sets the entire month's demand charge. And under a ratchet clause, it sets far more than that.

How does a ratchet clause actually work?

Your minimum billable demand in any month is set at a percentage — typically 50% to 90%, most commonly 80% — of your highest recorded demand over the preceding six to twelve months. You're billed on whichever is greater: this month's actual peak, or that floor.

That means the ratchet is a rolling window, not a one-off multiplier, and it's why hand calculations get it wrong. Every month has its own floor, derived from a different twelve-month lookback.

Follow the defaults month by month. The prior year's 660 kW peak sets a floor of 528 kW, so January through May are billed at 528 despite actual demand of 370 to 510. June and July exceed the floor and are billed on actual. August's peak of 695 is billed on actual, but July's 720 kW now sets a new floor of 576 — and September through December are all billed at 576 kW against actual demand as low as 385.

Nine of twelve months are billed on the floor, not on what the facility actually drew. Total billing demand is 6,999 kW against 5,910 kW of actual peaks. The difference — $16,335 a year, 15.6% of demand charges — is pure ratchet penalty, and it traces back to a single July afternoon.

Check your own bills for the tell: if the demand charge is identical for several consecutive months, you're on a ratchet.

Will peak shaving save what the calculators say?

No, and the gap is worth understanding before you sign a contract on the strength of a proposal.

Nearly every demand charge tool and vendor page uses the same formula: kW reduced × $/kW × 12 months. Shave 90 kW off a $15/kW tariff and it reports $16,200 a year.

The calculator says $13,770 — the shortcut overstates by $2,430, about 18%.

Here's why. In a month where the ratchet floor is binding, you are already paying the floor. Shaving actual demand from 380 kW to 350 kW saves nothing that month, because you were never billed on 380 in the first place. What the shave does is lower next year's floor — by the ratchet percentage of the reduction, not the whole reduction.

So in the nine floor-bound months you capture 72 kW of a 90 kW shave, exactly the 80% ratchet. In the three months where actual demand binds, you capture the full 90. Blended, the effective saving is $12.75 per kW-month against a $15 headline rate.

None of this means peak shaving is a bad idea — the defaults still show a 5.1-year simple payback on $58,000 of capital after $2,400 of annual operating cost. It means the vendor proposal in front of you is probably 15 to 20% optimistic, and you should ask which formula produced it.

Does a ratchet make peak management more or less valuable?

Both, and the industry commentary usually only tells you one half.

The half you'll read everywhere is correct: a ratchet magnifies the cost of a single bad interval. One equipment startup that pushes you 200 kW above normal doesn't cost one month — it locks in an inflated floor for the following eleven. Preventing a spike is worth far more under a ratchet than without one, and that's a genuine argument for real-time monitoring and interlocks on large motor starts.

The half that's usually missing is that reducing demand you were never billed for is worth less under a ratchet. Once the floor is set, ordinary month-to-month trimming in low-demand months returns only the ratchet fraction. The two effects pull in opposite directions and the net depends on your load shape.

The practical consequence is a sequencing rule. Prevention beats reduction. Spend first on avoiding new peaks — soft starts, staggered equipment sequencing, interlocks, alarms on approach to the current floor — because those dollars protect eleven future months. Spend second on trimming baseline demand, because those dollars are discounted by the ratchet.

The calculator reports the floor currently in effect — 576 kW at the defaults — which is the number to set an alarm at. Anything approaching it is an ordinary month. Anything exceeding it is a decision with a twelve-month tail.

How do I use my own data with this?

Take twelve months of bills and pull the billed demand figure from each — it may be labelled "billing demand", "peak demand" or "maximum demand". Enter those, plus the highest demand from the twelve months before that window, which seeds the rolling floor for early months.

Then find your demand rate and ratchet terms in the tariff document, not the bill. The bill shows the result; the tariff shows the rule. You're looking for the ratchet percentage and the lookback period.

Three limitations to know.

The model assumes an eleven-month lookback. Tariffs vary from six to twelve months, and some reset annually on a fixed date rather than rolling. If yours differs, the shape of the answer holds but the month-by-month floors won't match your bill exactly.

Time-of-use demand is not split out. Many tariffs bill on-peak and off-peak demand at different rates, and some apply coincident demand charges tied to grid peaks rather than your own — ERCOT's four-coincident-peak method being the well-known example. This calculator models a single demand rate. For a TOU tariff, run it once per demand period.

Seasonal rates are not modelled. Where summer and winter rates differ, use a weighted average or run the seasons separately.

None of these change the central point, which is that the ratchet is a rolling floor and treating it as a flat multiplier will misprice your project.