A factory drawing 200 kW can still be billed for 400

A factory budgeting electricity savings from lower peak demand can be months ahead of its bill. In the hypothetical case below, a current peak of 200 kW still produces 400 kW of billed demand, because the tariff remembers a busier month.[1]

That gap matters when a business approves new controls, shortens shifts or moves production between buildings. The equipment can perform as expected while the cash forecast disappoints. A demand ratchet makes part of the electricity bill depend on the site's recent history, so the timing of savings deserves its own calculation.

Green substation doors beside red-brick industrial buildings and a paved yard in Burley, Leeds. — A substation on an industrial estate in Burley, Leeds, photographed by Stephen Craven in July 2024. The scene illustrates the infrastructure behind electricity service; the billing scenario below is hypothetical. Photograph via Geograph and Wikimedia Commons, CC BY-SA 2.0; resized for display.[5]

The bill has a memory

Electricity consumption and demand measure different things. Kilowatt-hours record energy used over time; kilowatts measure the rate of use. For billing, demand is commonly measured over a defined interval, so the relevant peak is an interval average rather than a momentary electrical flicker. The Department of Energy's manufacturing guidance distinguishes that measured peak from the billing demand calculated under the rate schedule.[2]

A ratchet adds a historical comparison. Oncor's delivery-charge glossary describes a common arrangement: billing demand is the higher of the current monthly peak or 80% of the highest monthly peak in the preceding eleven months. It directs customers to the applicable tariff because the rule is not identical for every account.[1]

The practical result is a temporary floor. Once the historical calculation exceeds the current peak, further reductions in that peak leave this particular billed quantity unchanged. The separate energy charge can still fall with consumption. DOE's rate-selection guidance treats energy, demand and fixed charges separately for precisely this reason: a change in how a building uses electricity affects each component differently.[3]

Scenario one: the quieter month arrives first

Assume a factory has the lookback rule described above and a historical peak of 500 kW that remains inside the window. Its current measured peak is 200 kW. Ignore contractual minimums, power-factor adjustments and other demand definitions for this illustration; assume the price per billed kilowatt stays constant.

The calculation is:

Billed demand = the greater of 200 kW and (80% × 500 kW) = 400 kW.

These are illustrative inputs and arithmetic, not a reported customer's bill or a quotation for electricity service.

The factory pays twice what a forecast based only on its current peak would assign to this demand-charge component. That does not mean its entire electricity bill doubles. Energy purchases, fixed fees and any other charges have their own calculations.

Nor does the example mean the operational improvement saved nothing. Against a counterfactual in which the factory kept reaching its old peak, billed demand has fallen. The budgeting error is to assume that the whole measured reduction becomes an immediate reduction in billed demand. Part reaches the invoice now; the rest waits behind the floor.

For an equipment project, that distinction can move the payback date. A proposal may correctly estimate lower power requirements yet overstate early cash savings by pricing every avoided kilowatt as though the historical peak had already disappeared. The project forecast needs a billing calendar alongside its engineering estimate.

Scenario two: the old peak leaves the window

Now let the busy month's reading age out. Assume every remaining historical peak, and the new current peak, is at the lower operating level. In this simplified case, current demand becomes the larger term in the comparison, and billing demand finally falls to it.

The relief follows the tariff's billing periods. A manager's anniversary of an equipment installation does not itself determine when an older meter reading stops counting. Reconstructing the eligible history matters more than circling a convenient calendar date.

There is another qualification: the largest old peak can expire while the next-largest reading still supports a floor. A contractual minimum can also keep billing demand elevated. Alberta's Utilities Consumer Advocate explains that measured demand, historical demand and minimum requirements can interact, and that different charges on the same account may use different demand measures.[4] Its examples concern Alberta; the broader lesson is to inspect each charge's actual rule before forecasting a release.

Scenario three: a restart extends the wait

Suppose a production restart brings enough equipment online together to establish another high interval-average peak. That new observation enters the rolling history. Even as the original busy month approaches expiry, its replacement can preserve the billing floor.

This is why preventive controls can retain value while immediate savings are constrained. DOE's manufacturing guidance identifies load shedding and load shifting as ways to manage demand.[2] In this scenario, keeping a replacement peak out of the record protects the later reduction. Simply moving the same factory-wide peak to another hour would not defeat a ratchet based on the maximum at any time; the relevant peak itself has to fall.

The strongest counterweight to the delayed-savings concern is therefore operational. Energy savings can begin immediately, and sustained peak control can improve future bills.[2][3] A ratchet changes the timing and marginal value of those improvements. It does not make them pointless.

There is also a physical reason for the charge. The Alberta advocate explains that distributors must provide infrastructure capable of meeting a customer's high demand even when that capacity is needed only briefly.[4] Lower activity this month does not remove the poles, transformers and other equipment already required to serve the site.

The three events that decide the forecast

The delayed-savings thesis has a clear falsifier for a real account: the applicable tariff and reconciled invoice show that this demand-charge component follows the current measured peak, with no binding historical floor. In that case, carrying the hypothetical ratchet into the cash forecast would be wrong.

Watch three events:

The next invoice after the operating change: reconcile the measured peak, billed demand and rate for each demand-based charge. This tests whether the anticipated savings have reached the bill.
The billing cycle when the old maximum expires: recalculate the floor from all remaining eligible readings and any applicable minimum. Expiry alone is insufficient evidence of full relief.
The next restart or equipment commissioning: inspect interval data for a replacement peak. Its absence supports the forecast reduction; a new high reading can postpone it.

cronfeed.work