There is a specific moment that starts most energy projects. Somebody in finance opens the electricity statement, divides the total by the kilowatt-hours, gets a number that is nothing like the rate in the supply contract, and asks what happened.
Nothing happened. The calculation is wrong, because the bill is not measuring one thing.
A residential electricity bill answers a single question: how much energy did this household consume. An industrial or large commercial bill answers several questions at once, and prices each answer separately. Consumption is one of them. The others are about the rate at which that consumption arrived, the quality of the current drawn, and a set of costs the regulator has allowed the utility to recover through surcharges that have nothing to do with either.
Until you can name every line and say which determinant it is priced on, you cannot tell whether a bill went up because the plant worked harder or because a compressor started at the wrong moment on a Tuesday in July.
The four blocks
Almost every commercial and industrial statement in the United States, whatever the format, resolves into four blocks.
Energy. Kilowatt-hours multiplied by a price per kilowatt-hour. This is the block everyone recognizes, and on a demand-heavy tariff it can be a minority of the total. It measures volume: total work done over the month.
Demand. Kilowatts multiplied by a price per kilowatt per month. It measures the highest rate of consumption reached during the period, not the total. It is the block that surprises people, and it has its own article because there is a great deal in it.
Riders and surcharges. A stack of separately named charges the state commission has authorized: fuel cost adjustments, transmission cost recovery, energy efficiency program funding, nuclear decommissioning, storm recovery, low-income assistance. Each has its own basis — some on kilowatt-hours, some on kilowatts, some as a percentage of the subtotal. Together they are frequently large enough that ignoring them makes any savings calculation wrong.
Taxes and fees. Applied to what is above them, and largely outside your control.
The useful observation is not that there are four blocks. It is that two of the four respond to when you use power rather than how much. Energy is a volume charge. Demand is a timing charge. Several riders attach to demand rather than to energy. Change the shape of a load without changing its size, and two of the four blocks move.
Start with the determinants, not the dollars
A billing determinant is a quantity the utility measures or derives, to which a rate is applied. Every dollar on the bill is a determinant multiplied by a price. If you learn to read the determinants first, the dollar column stops being mysterious.
The determinants you will meet:
| Determinant | Unit | What it measures | Typical charge | | --- | --- | --- | --- | | Energy | kWh | Total consumption over the period | ¢/kWh | | Billing demand | kW | Highest interval average, after any adjustments | $/kW-month | | On-peak demand | kW | Highest interval average inside a defined window | $/kW-month | | Apparent power | kVA | Demand including the reactive component | $/kVA-month | | Power factor | ratio | Real power divided by apparent power | penalty or adjustment | | Customer charge | account | Fixed cost of being connected | $/month |
Note the third row and the second. On-peak demand and facility demand are different determinants, and a tariff can bill both. A site can be excellent on one and poor on the other. The distinction is set out in facility demand versus on-peak demand, and it is worth being certain which one your tariff uses before spending money to reduce it.
The customer charge
The first line, usually the smallest, and the only one that does not move. It recovers the cost of the connection, the meter and the billing itself. There is nothing to manage here; note it and move on.
The energy block
Kilowatt-hours consumed multiplied by a price. The complications are three.
Seasonality. Many schedules define a summer season and a winter season with different prices, and the season boundary is set by the tariff rather than by the calendar. A bill that spans the boundary is split, with each part priced separately.
Time-of-use periods. On a time-of-use schedule the kilowatt-hours are divided among on-peak, mid-peak and off-peak buckets and priced separately. This is where a great deal of money moves for sites with any operational flexibility, and it is covered in time-of-use rates for commercial accounts.
Blocks. Some older schedules price the first block of kilowatt-hours at one rate and the remainder at another, sometimes with the block size itself set as a multiple of billed demand. A declining block structure means the marginal rate on the last kilowatt-hour is not the average rate, which matters enormously for an efficiency business case.
The Energy Information Administration publishes the average revenue per kilowatt-hour that utilities actually collect from industrial and commercial customers, broken down by state, which is a useful sanity check on whether your blended cost is unusual for your territory rather than merely unwelcome.
The demand block
Here is the block that justifies most of this site.
The meter does not record a single instantaneous maximum. It records the average load over each successive interval — commonly fifteen minutes, sometimes thirty — and the demand charge is applied to the highest of those averages during the billing period. A two-second inrush when a large motor starts is invisible to it. A steady elevated load for a quarter of an hour is not.
The consequences are unintuitive enough to be worth stating directly:
- Two facilities consuming identical annual energy can face materially different demand charges, because one drew its energy evenly and the other in bursts.
- Reducing consumption does not necessarily reduce the demand charge at all. Turning off a light that was already off during the peak interval saves energy and saves nothing on demand.
- One badly sequenced startup can set the determinant for the whole month, and under a ratchet clause, for much of the year afterward.
That last point deserves emphasis, because a ratchet is the mechanism most likely to be present on a bill and absent from the reader's mental model of it. Ratchet clauses establish a floor under billing demand based on a previous peak, so that a single afternoon continues to be billed long after the load that caused it has gone.
The riders
This is the block that gets skipped, and skipping it is how savings projections end up wrong by a third.
Riders are charges the state commission has approved for the utility to recover specific costs outside the base rate. The names vary by jurisdiction, but the categories recur: fuel and purchased power adjustments, transmission cost recovery, distribution investment recovery, energy efficiency and demand-side management program funding, renewable portfolio standard compliance, storm or wildfire recovery, decommissioning, and various public benefit funds.
What matters for arithmetic is the basis. A rider levied per kilowatt-hour scales with energy. A rider levied per kilowatt of billed demand scales with the peak — and therefore a demand reduction project saves more than the headline demand rate suggests, because it also reduces every demand-based rider stacked on top of it. A rider levied as a percentage of the subtotal amplifies everything beneath it.
List every rider on your statement with its unit. Add the demand-based ones to the headline demand rate before you calculate the value of a kilowatt of avoided peak. The article on what a kilowatt of avoided peak is actually worth works this through with the arithmetic on the page.
Riders and their bases are covered in more depth in riders and surcharges.
The power factor line
Not every tariff has one, and the ones that do handle it in one of three ways: an explicit penalty when power factor falls below a threshold, billing on apparent power in kVA rather than real power in kW, or a multiplier applied to billed demand.
These are not equivalent, and a correction that pays back under one may not under another. If your bill shows a kVA figure, or a power factor percentage, or a line with the word "reactive" in it, power factor penalties is the article to read next.
The numbers that are not printed
Two useful quantities appear on almost no bill and have to be calculated.
Load factor. Energy consumed divided by what would have been consumed had you run at your peak demand continuously for the whole period. A low load factor means a spiky load, and a spiky load is punished by any demand-heavy tariff. It is the single most diagnostic number you can derive from a statement, and it takes two figures already printed on it.
Load factor from figures already on the bill
Substitute your own kWh, kW and days.
- Energy consumed in the period420,000 kWh
- Billed demand1,400 kW
- Days in the billing period30
- (Hours in the period: 30 × 24)720 h
- (Energy at constant peak: 1,400 × 720)1,008,000 kWh
Load factor: 420,000 ÷ 1,008,00041.7%
A site at 42% is drawing, on average, well under half of the capacity it is being billed to have available. That gap is where demand charges live. Figures are illustrative — run it with your own.
Blended cost per kilowatt-hour. The total bill divided by kilowatt-hours. It is not a rate and it should never be used as one in a savings calculation, but it is a fair summary figure for a budget, and tracking its movement month over month is a fast way to notice that something structural has changed.
Reconciling the bill against the tariff
The bill is a summary. The tariff is the contract. When they appear to disagree, the tariff wins, and reading it is not optional for anyone about to spend money on the basis of the bill.
- Find the rate schedule name printed on the bill, usually near the account number.
- Get the filed version from the state utility commission's site, not a summary of it. Commissions publish approved tariffs as public documents.
- Confirm the demand interval length and whether demand is measured over the whole month or only inside a defined window.
- Confirm whether a ratchet applies, at what percentage, and over what look-back period.
- Confirm how power factor is treated, if at all.
- Recalculate each block from the determinants on the bill and the rates in the tariff, and check they reconcile to the printed subtotals.
- Where they do not, check the meter multiplier before assuming the utility is wrong.
That last line catches more errors than any other. A meter multiplier or current transformer ratio is applied between what the meter register shows and what the bill uses, and when it is wrong every figure downstream is wrong by the same factor. Meter multipliers and CT ratios covers how to check it.
What to do with this
Reading the bill is diagnostic work, not an end in itself. What it produces is a short and specific list:
- The proportion of the bill that is demand-driven, including demand-based riders. If it is small, peak management is not your project. If it is large, it is probably the highest-return work available.
- Your load factor, and therefore whether you are on a tariff structure that suits your load shape at all.
- Whether a ratchet is in force, which changes the value of avoiding a peak by a large multiple.
- Whether there is a power factor exposure you have been paying without noticing.
- Which determinants your riders attach to, which sets the real value of every kilowatt you avoid.
With those five answers, the rest of this site becomes actionable rather than interesting. The natural next step is demand charges explained, and after that, getting hold of your interval data — because everything past this point is done with fifteen-minute readings rather than with a monthly summary.
