Demand response is frequently presented as free money for load a site was going to reduce anyway. It is a legitimate revenue stream and it is not free money, because what is being sold is not electricity. It is a commitment to reduce load on request, backed by a penalty if you cannot.
That framing puts the important clauses in the right order. The payment is the easy part. The obligation, the measurement and the penalty are where the value is actually decided.
The two structures
Capacity-style programs pay for availability. You commit a quantity of reduction, you are paid for standing ready across a season, and you are called a limited number of times. Revenue is predictable; performance obligations are strict, because the system operator has planned around your commitment.
Energy-style programs pay for delivery. You reduce when called and are paid for what you actually reduced. Revenue is uncertain; obligations are lighter, because nothing was planned around you.
The choice between them is a risk decision rather than a revenue one. A site whose ability to curtail is reliable and quantifiable should prefer the first. A site whose flexibility depends on what is in production that week should prefer the second.
The clauses that decide the outcome
- Notification period. Ten minutes, thirty minutes, a day ahead? This determines whether the commitment is operationally realistic at all.
- Event duration and annual limits. Maximum hours per event, maximum events per season, maximum total hours. A program with no cap is an open-ended commitment.
- Season and hours. When can events be called? A summer weekday afternoon program is a different proposition from one that can call at any hour of the year.
- Baseline methodology. How your reduction is measured. This is the single most consequential clause and it is usually the least read.
- Penalty structure. What happens on non-performance, and whether it reaches back to capacity payments already received.
- Test events. Whether the program can call a test, how it is treated, and whether a test can itself set a peak on your tariff.
- Metering and telemetry. What equipment is required and who pays for it.
- Exit terms. Notice to withdraw, and any clawback.
Why the baseline clause matters most
Your reduction is not measured directly. It is measured as the difference between a baseline — an estimate of what you would have consumed had no event been called — and what you actually consumed.
The baseline comes from a formula in the program rules, typically an average of your consumption during the same hours on a number of recent comparable days, sometimes with an adjustment based on the hours immediately before the event.
Two consequences follow, both practical:
A site that has already flattened its load has a low baseline and therefore little measurable reduction to sell. This is the direct conflict between demand response revenue and demand charge reduction, and it is real: the same flexibility cannot always be counted twice.
Frequent events depress the baseline. If recent curtailment days are included in the baseline calculation, each event lowers the reference against which the next one is measured. Programs handle this with exclusion rules — check that yours does.
Pricing it properly
The revenue is not the value. The value is revenue minus the cost of performing, minus the risk-weighted penalty exposure.
Net value of a capacity-style enrollment
Same site as elsewhere on this section.
- Committed reduction800 kW
- Capacity payment, per kW-year$42.00
- (Gross capacity revenue)$33,600
- Events called in the season6
- Average event duration3 hours
- Production cost per hour of curtailment$900
- (Cost of performing: 6 × 3 × 900)$16,200
- Energy payment for delivered reduction$2,400
- Telemetry and metering, annualized$1,800
Net annual value$18,000
Gross revenue was $33,600 and net value is $18,000. Programs are usually sold on the first figure. Whether $18,000 is worth the operational commitment is a judgment, but it is at least the right number to make it on. Figures illustrative.
The interaction with your own demand charge
This is the part most often handled badly, because the two benefits are evaluated by different people.
Curtailing during an event may also reduce a coincident determinant. If events tend to be called at times of system stress, and your capacity or transmission obligation is measured at exactly those times, one action produces two benefits. That is genuine stacked value.
But it may not touch your non-coincident demand charge at all. If your own monthly peak occurs on a different day at a different hour, curtailing during an event does nothing for it.
And the restoration after an event can create a peak. Everything shed comes back, and if it comes back together the following interval can exceed anything the event avoided. Restoration needs staggering, exactly as a cold start does: staggered startup.
A test event is a peak risk in itself. Where a ratchet is in force, an ill-managed test or restoration can set a floor that costs more than a season of program revenue: ratchet clauses.
Where storage changes the calculus
A battery removes the production cost of performing. Instead of stopping a line, you discharge, and the operational cost line in the calculation above collapses toward the round-trip energy cost.
That changes the answer twice over: it makes participation far more attractive, and it lets you commit a larger and more reliable reduction, which is what capacity-style programs pay best for. It also means the battery earns from more than one source, which is frequently what makes the capital work: sizing a battery for peak shaving and total cost of ownership for a behind-the-meter battery.
The caution is that stacked benefits are not automatically additive. A battery discharged for a called event may be unavailable for the site's own peak later the same afternoon, and program rules govern whether a single reduction can be claimed twice. Both need checking against the specific contract rather than assumed.
Before enrolling
Establish what you can genuinely curtail, for how long, at what operational cost, agreed with the people who run the process rather than estimated from a drawing. That figure is the input to everything above, and it is the same figure needed for how to reduce peak demand charges. A site that has done that work once can evaluate any program offered to it in an afternoon.
