Guides

What your electricity bill is actually telling you

A kWh figure on a bill and a wattage number stamped on a tool look like they should turn into a dollar amount with one multiplication. Usually they don't — here's what actually sits between the two, and how to close the gap.

By D. Shin · Published September 3, 2026

Somebody hands you a compressor, a bill, and a question: what does running this thing cost? The honest answer starts with admitting that "watts times hours times rate" is the easy 90% of the problem. The other 10% — what the meter is really counting, whether your rate is even a single number, and whether the thing you're costing out draws its nameplate power continuously or in bursts — is where estimates quietly go wrong by a factor of two or three. None of it is complicated once it's laid out; it's just rarely laid out.

What the meter is counting, and how to read it yourself

A utility meter doesn't measure power (watts) — it measures energy (watt-hours), accumulated over time. A kilowatt-hour is exactly what it sounds like: the energy used by a one-kilowatt load running for one hour. A 100-watt load run for ten hours uses the same 1 kWh as a 10,000-watt load run for six minutes; the meter can't tell the two apart and doesn't need to, because your bill is priced per kWh, not per watt.

kWh = (Watts ÷ 1000) × Hours

That's the entire arithmetic core of every appliance-cost estimate on this page, and it's also exactly what the electricity cost calculator does per row when you give it a wattage and a usage time — the calculator just saves you doing the division and the rate multiplication by hand, and lets you stack several loads to see a combined total.

If your building still has an old electromechanical meter — a glass-domed unit with four or five small dial faces instead of a digital readout — reading it correctly takes one non-obvious trick. Each dial is numbered 0 through 9, and because of the gear train linking them, adjacent dials spin in opposite directions from each other. Read left to right and write down the last number each pointer has fully passed. The trap is a pointer that looks like it's sitting right on a digit: check the dial immediately to its right. If that dial's pointer hasn't yet swept past zero, the dial you're reading hasn't actually finished that count yet either — take the lower of the two digits it appears to be between. Get this backwards on a meter with several dials and a reading can be off by a full order of magnitude, not a rounding error. Most utilities have replaced these with digital or smart meters by now, but plenty of older workshops, barns, and outbuildings still meter through one.

A smart meter solves the ambiguity but adds a different opportunity: many utilities publish 15-minute, 30-minute, or hourly interval data through an online portal or app. That interval data is the closest thing to ground truth you'll get without your own equipment — it shows the actual load curve of the building, spikes and all, rather than a single monthly total you have to reverse-engineer.

Your rate is probably not one flat number

Everything above assumes a single price per kWh, and plenty of residential accounts really do bill that way. Two structures don't, and both are common enough that it's worth checking your own tariff before trusting a flat-rate estimate:

• Tiered (inclining-block) rates. Usage up to some threshold bills at one rate; everything above it bills at a higher rate, sometimes with a third or fourth tier stacked on top. Utilities that use this structure are usually trying to price a lot of use — think a workshop with a compressor and a kiln running alongside the house — noticeably higher than baseline household use. The exact thresholds and the rate at each tier are set by your utility and revised periodically, so there's no universal number to plug in; they're printed on the bill itself or on the utility's published tariff schedule.
• Time-of-use (TOU) rates. The price per kWh changes by the hour of day and often by weekday versus weekend — a peak window (commonly a late-afternoon-through-evening block) priced well above an off-peak overnight rate, sometimes with a third "super off-peak" band cheaper still. On these plans, running a compressor or a kiln during the peak window can cost two or three times what the identical run would cost four hours later.

Neither structure fits into a single "rate" field without a bit of translation. The practical fix for a reasonable estimate is a blended rate: take the total dollar amount from a recent bill, divide by the total kWh it covers, and use that as your effective per-kWh price. It won't capture which tier or time window a specific load falls into, but it's honest about your actual average cost, and it's exactly the number the electricity cost calculator's rate field expects if you don't want to model tiers separately. For a load that genuinely shifts with the clock — say, deciding whether to run a compressor at 2 p.m. or 10 p.m. — the more accurate move is running the calculation twice, once per rate tier or TOU window, using the calculator's ability to hold several appliance rows at once, and comparing the totals directly rather than guessing which one wins.

Nameplate watts vs. what a cycling load actually draws

A space heater or an incandescent bulb draws close to its rated wattage the entire time it's switched on, so watts × hours is close to the truth. A lot of the loads people actually want to cost out don't behave that way — they're thermostatically or pressure-controlled, cycling on and off to hold a setpoint, and the nameplate number is what they draw while running, not what they average over a day.

The fix is a duty cycle: the fraction of time the load is actually on.

Average power = Running watts × Duty cycle

The gold-standard way to get the duty cycle is a plug-in power meter logging over a full 24 hours — it captures the real cycling pattern including seasonal and use-pattern quirks a formula can't. Absent that, a reasoned estimate gets you into the right neighborhood, which is usually good enough to compare options or sanity-check a bill.

Refrigerators are the classic case. A modern compact-to-mid-size fridge might draw somewhere in the 100–200 watt range while the compressor is actually running, but the compressor itself typically cycles on for roughly a third to half of any given hour — less in a cool kitchen with the door rarely opened, more in a hot garage or one that gets opened constantly. Take a fridge with a 150 W running draw and a 35% duty cycle: average continuous draw is 150 × 0.35 ≈ 52.5 W, which is the number you'd actually plug into an hours-based estimate — not the 150 W on the nameplate, and not a guess at how many hours a day it's "plugged in," since it's always plugged in.

Phantom loads: the total that never shows up on a nameplate

Plenty of equipment draws power while doing nothing you'd call "on" — a router and modem running around the clock, a cable box or DVR that stays semi-awake to keep its guide data current, chargers left in the wall with nothing attached, a microwave's clock display, a shop computer left in standby rather than fully shut down. Individually these are usually small, commonly in the low single digits of watts up to perhaps 10 W; a handful of older or poorly designed devices — some cable boxes and older set-top gear in particular — have been measured drawing 15–30 W continuously, essentially indistinguishable from just leaving a load switched on all the time. Energy-efficiency researchers studying whole-house standby draw have repeatedly found it adds up to a meaningful slice of total household use, commonly cited in the single digits to low teens as a percentage, once every always-plugged-in device in a home or workshop is counted.

You can measure your own phantom floor directly without any special equipment: switch off or unplug everything you reasonably can, then watch the meter (or, on a smart meter, the live usage figure on the utility's app or dashboard) for a minute. Whatever load remains — the stuff you couldn't or didn't switch off — is your standby floor, running 24 hours a day whether anyone's using the building or not. It's worth doing once, because it's the single easiest reduction to act on: a power strip that cuts a bench's worth of idle chargers and standby electronics with one switch removes that floor entirely when the shop's closed for the night.

Putting it together: a month for a small workshop

Here's the whole method run against a plausible one-person workshop — a benchtop table saw, a small piston air compressor, and LED shop lighting — over a 22-working-day month, billed at a flat $0.15/kWh for simplicity. Treat the wattages and duty cycles as the illustrative inputs they are; substitute your own tools' nameplate ratings and your own honest estimate of how much of the working day each one is actually drawing power, not just switched on.

That's about 4.85 kWh on a working day for the three loads together. Over 22 working days that's roughly 107 kWh, and at $0.15/kWh, about $16 for the month from these three loads specifically. Notice where the money actually goes: the saw, the tool with the highest rated wattage by far, contributes the least to the bill, because it's only actually cutting for a few minutes an hour. The compressor and the lights — lower wattage each, but on (or cycling) for most of the working day — dominate the total. That's the duty-cycle lesson from the sections above showing up directly in a real total: nameplate wattage ranks the loads backwards from how they'll actually rank on the bill.

A whole-building bill will run well above a three-load subtotal like this one — HVAC, water heating, any welding or heavier three-phase equipment, and the phantom floor discussed above all add on top. Building the estimate load by load, the way the table does, is what turns "my bill went up, no idea why" into an actual answer: run the same exercise before and after adding equipment, or across a season, and the difference between the two totals should roughly account for the swing on the actual bill. When it doesn't, the gap is almost always one of a short list — a duty cycle assumed too low, a heating or cooling load left out of a seasonal comparison, or a billing period read as a flat 30 days when it actually ran 34.