There is a mechanism in commercial electricity tariffs called a ratchet, and if it applies to you it changes what a single bad afternoon is worth. Under a ratchet, your billed demand in any month cannot fall below some fraction of the highest demand you set in the preceding months. Set a new peak in July, and you keep paying against it through the winter, whether or not you ever come close to it again.
Whether that mechanism applies to you decides something practical: if a one-off spike costs you twelve months instead of one, preventing it is worth roughly twelve times more, and a lot of demand management equipment is sold on exactly that arithmetic.
What we find in ComEd's commercial structure
On the standard ComEd commercial delivery classes, billed demand is the current month's maximum inside the on-peak window. That is the whole determinant. The tariff's definition looks at the monthly billing period and no further back, and on the commercial bills we have reconciled there is no ratchet line and no carried-forward value.
So on those classes, one bad month is one bad month. The spike costs you the difference for that billing period, and the next bill starts fresh.
Two limits on that, stated because they are real. Our reconciliations are on the standard secondary-voltage delivery classes; we have not reconciled every class, every voltage or every special contract, and a ratchet appearing in a corner of the book we have not read would not be visible to us. And absence on a reconciled bill is evidence from the bills we have, not proof across all periods and all accounts. If your bill shows a demand line whose value does not match anything your meter did that month, that is exactly the signature of a determinant we have not accounted for, and it is worth chasing.
Why this changes what prevention is worth
Run the two cases against the same event and the difference is not subtle.
A site spikes 30 kW above its usual peak during one summer afternoon. Under a monthly determinant, the cost is thirty kilowatts at the delivery rate, for one month. Under a ratchet holding, say, most of that peak forward, the same thirty kilowatts are billed again and again until the ratchet period rolls off.
Same event. Very different number. And the equipment that would have prevented it has the same price in both cases.
This is why the ratchet question belongs before the equipment question rather than after it. It is also why proposals built on a generic national model of demand charges can be badly wrong in ComEd territory in a specific direction: they assume a persistence the tariff does not have, and the assumed savings never arrive.
What to do with the answer
If your peak is set by rare events, and there is no ratchet, the honest read is often that you should tolerate them. A spike two or three times a year, priced for those months only, may be cheaper than the system that would prevent it. That is a legitimate conclusion and it is the one nobody selling the system will reach for you.
If your peak is set by your ordinary operating shape, the ratchet question is nearly irrelevant, because you re-establish the same peak every month anyway. Nothing is being carried forward that you were not going to set again. In that case the lever is the shape of the load itself, not event prevention.
If you are comparing sites across states, do not assume ComEd's structure travels. The determinant, the window and the ratchet are per-tariff, and a portfolio decision made on the Illinois mechanism can be wrong everywhere else.
How to check your own bill in five minutes
Take twelve consecutive bills and write down the billed demand on each.
If the billed kilowatts move up and down freely, tracking what you would expect from your operation, you are being billed on the month. If several consecutive months show the same billed demand while your operation clearly varied, or if billed demand never falls below a floor that appeared after one large month, that is the shape of a ratchet and you should ask ComEd or your supplier which provision produced it.
Either way the check costs nothing and uses only bills you already have. And once you know which case you are in, the next question is the one that decides the money: how far your top on-peak interval sits above your second-highest, and whether the gap is worth attacking at all.
What is a demand ratchet?
A tariff provision that sets your billed demand to at least a fraction of a peak established in an earlier period, so a single high month raises your bills for months afterward even if your usage falls.
Does ComEd use a demand ratchet on commercial accounts?
On the standard commercial delivery classes we have reconciled, billed demand is the current month's maximum inside the on-peak window, with nothing carried forward. Check your own bill for a demand line that does not match your meter's activity that month.
If there is no ratchet, is a one-time spike still expensive?
It costs the difference in kilowatts for that billing period, at your delivery rate. That is real money, but it is one month of it, and that changes what preventing it is worth.
How do I tell whether my bill is being ratcheted?
Compare billed demand across twelve consecutive bills. Freely varying billed kilowatts indicate a monthly determinant. A value that repeats or refuses to fall below a floor set by one large month is the signature of a ratchet.