If you open the electric bill and the demand line makes no sense, this is for you
If you run a plant, a hospital, a school system, a city, or a big commercial facility — and every month you see a huge line for demand that keeps jumping around and never quite makes sense — you are exactly who this guide is written for.
You never wanted to be in the energy business. You got dragged into it because of budget and risk. You have done projects, you have tried to use less, and the bill still does not move the way it should. That is not a you problem. The system is genuinely convoluted — and it has been quietly pushed onto operators like you.
How can your facility actually manage and reduce demand charges in a way that lowers your all-in cost per kilowatt-hour and takes risk out of your budget — without pretending your entire operation revolves around what the utility wants?
By the end of this guide you will understand what demand charges really are, why one bad fifteen-minute window can wreck a whole month, what load factor is, and how to tell whether a vendor actually understands your bill when they promise to "cut your demand."
A typical commercial or industrial bill has three big pieces. Two of them most people understand intuitively. The third is where things go sideways.
Customer or connection charges to be tied to the grid. You pay them whether you use a lot, a little, or nothing.
Cost per kilowatt-hour — power over time. A 4 kW load running 8 hours is 32 kWh. Most people get this.
Demand is the size of the pipe. Usage is what flows through it.
This one distinction is the whole foundation. Get it, and the rest of the bill starts to make sense.
Buy an eight-inch pipe but only run a trickle through it, and you still pay for eight inches of demand. The utility charges you for the size of the pipe you bought during the worst fifteen- or thirty-minute window in the billing period.
The meter watches how high your draw gets during those windows. The highest average you hit becomes your billable demand — for the entire month.
Demand is like being billed on what it would cost to eat everything in the buffet — even if you walk in and order soup. You pay for what you could pull in that one interval, not just what you did pull.
Demand stacks — it is not averaged over the month
Run an A/C unit pulling 4 kW and a water heater pulling 4 kW at the same time, and your stack is 8 kW. Run them one after the other and it never tops 4 kW. Same total energy by the end of the day. Very different billable demand.
Full demand charge.
Half the demand charge.
It is the highest coincident stack in a single measurement window that sets the number you get billed on — not your average, not your total.
Why the utility cares so much about that stack
From their side, they have to build and maintain enough capacity for everybody's worst-case peak at the same time — not the average Tuesday mid-morning, but the worst hour of the hottest day when every chiller, every press, every air handler is hammering at once.
So they build generation and wires for the peak, not the average. Demand charges are how they fund that ready, dispatchable capacity. In theory, the customers who put the most volatile and stressful load on the system pay more for it.
That is the theory. In reality, many utilities are reducing baseload capacity and reserve margin while still collecting rising demand charges that were meant to fund that margin. The "equitable recovery" story is weaker today than the textbooks suggest. But you and I don't get to vote demand charges off the bill — so the real game is understanding how they work on your specific rate, and using that to control your all-in cost per kilowatt-hour.
Load factor is the ratio of the kilowatt-hours you actually consumed in a month to the potential kilowatt-hours implied by your peak demand. It is the single most important number for controlling commercial and industrial energy cost — and the easiest way to see it is the twelve-passenger van.
Big van, three riders. You bought huge capacity and barely fill it — a poor use of funds, and a high cost per rider.
Smallest van you can get away with, filled close to full. High utilization — and a low cost per rider.
In bill terms, the size of the van is your peak demand in kilowatts. The riders are the kilowatt-hours you actually run through that capacity. Balancing the two is the key: drag demand up without a matching rise in kWh, and your cost per unit climbs. Hold demand down while still running plenty of kWh through the smaller pipe, and your cost per unit falls.
Put a number on it
Two facilities can burn the same energy and pay wildly different rates. Here's the same 120,000 kWh month at two different peaks:
| Scenario | Peak demand | Hours | Potential kWh | Actual kWh | Load factor |
|---|---|---|---|---|---|
| Peaky operation | 400 kW | 720 | 288,000 | 120,000 | 42% |
| Same energy, lower peak | 300 kW | 720 | 216,000 | 120,000 | 56% |
Nothing about the work changed — same 120,000 kWh of useful output. Shaving the peak from 400 to 300 kW spreads the demand charge across the same energy, so your all-in cost per kWh drops. That is load factor working in your favor.
The flavor of demand charge you're on changes the game
Find yours on your tariff sheet. The structure decides which behaviors actually move your bill.
| Structure | How it works | What it means for you |
|---|---|---|
| Max / non-coincident | The single highest 15- or 30-minute stack anywhere in the month sets the charge. | One bad interval — anywhere, anytime — sets the whole month. |
| Time-of-use demand | Same idea, but only inside a defined window (e.g. 4–9 pm) that matches the system peak. | Behavior in that window matters far more than at 3 a.m. |
| Flat | One cost-per-kW number, applied to your peak. | Simple: every kW of peak costs the same. |
| Tiered | The per-kW price changes once you cross a threshold. | Crossing a tier can cost more than the extra kW suggests. |
| Daily demand | Each day has its own mini demand charge based on that day's peak (often experimental rates). | A single bad day can't be averaged away — every day counts. |
| Residential creep | Demand components appearing on residential tariffs, often with storage-linked options. | Increasingly not just a C&I concern. |
Demand ratchets: a peak you keep paying for
A ratchet is a preset reservation. It says: no matter what you actually do this month, you will never be billed on less than a set amount of demand — often tied to a peak you set months ago.
Say a tariff sets a minimum billing demand of 100 kW and adds that in certain months your billable demand can never fall below 70% of your highest summer peak. Set a 400 kW peak in August, and even if quiet November only ever sees 150 kW, you are not billed on 150 — you are billed on 280 kW.
You cannot casually set a wild summer peak and forget it — you'll pay for that one decision in quiet months long after the event. Some ratchets are fixed, some are variable and tied to prior periods. The more complex the design, the less chance a normal operator has of reading the tariff and using it strategically. If a tariff sheet has ever made you feel dumb, you're not dumb. The system is that convoluted.
What this means inside your walls
of the total bill is demand charges for many large users. Not a rounding error — a big chunk of cash that moves with behavior you may not even be aware of.
A single bad window — motors starting together, an HVAC control going haywire, a control-system failure — can set the month's peak, freak incident or not.
That is why bills jump even when total kWh barely move — and where the classic misconception lives: "We used less this month, why is the bill still so high?" If you don't look at demand and kWh together, through the lens of load factor, you never get a satisfying answer.
Step one is real visibility — interval-level data, not just the total on the bill. You need to know, by time of day, which equipment creates your peaks and when. Without it you're guessing, and the only thing worse than doing nothing is confidently doing the wrong thing.
But visibility alone doesn't cut the bill — staring at graphs isn't a strategy. You need a set of principles and plays, each judged against one question: what does this do to my load factor and my all-in cost per kWh on my exact rate?
Nudge a couple of large loads apart so they don't stack in the same window. In a facility that has never watched this, a 20%+ cut in demand charges is not crazy. But if the only way to stagger is a third shift whose supervisor and support costs destroy the savings, it's a bad trade. Energy savings live inside the rest of your operating model.
Systems that monitor load in real time and discharge during peaks to hold the grid-seen stack down. Can cut demand charges and smooth your profile — but these are capital projects, and payback hinges entirely on how demand and energy are priced on your rate.
The utility or an aggregator pays you to cut load during grid-stress events — a second revenue stream, with disruption as the trade-off. Pairing demand response with prime-rated backup generation is often better than just shutting things off: you keep operations live, shave peaks, and turn a stranded backup asset into one with multiple payback channels.
Variable frequency drives on motors, better ramp profiles, better controls — reduce how hard you hit the system on starts and transitions. Always ask the same question: what does it do to load factor and all-in cost per kWh on your utility rate?
When monitoring enables the behavior changes and small process tweaks that trim demand, the net is real money — every year. This is exactly the problem TEG's Virtual Energy Manager was built to solve. Monitoring-as-a-service is not theoretical for us.
Four things a vendor must do — or they're guessing
The vendor landscape around demand is rough. Unless someone can speak to your specific utility rate and show you, accurately, what your all-in cost per kWh will be after their project, they're guessing — and you're paying. Be wary of any vendor who can't do all four of these with confidence.
If they can't explain load factor and show how the project changes it, that's a problem. Load factor is the whole game for C&I cost control.
All-in cost per kWh = total dollars on the bill (demand + energy) ÷ total kWh. If they can't give you that number with penny-level precision after the project — not a chance.
Not a generic spreadsheet — a model that reflects how your utility's billing system actually works today, including riders and ratchets.
Same industry, similar shift pattern, on the same rate or a close cousin — not "we did something like this once, in a totally different context."
If they can't clear that bar, all the glossy decks in the world don't matter.
When demand plays win — and when they're a bad idea
- You can improve load factor — hold demand as low as is realistic while running as many useful kWh through the smaller van as possible.
- You run high-utilization, multi-shift operations with steady equipment — and don't do something dumb that sets an unnecessary peak.
- You have real monitoring and can take low-disruption actions (small sequencing, modest control tweaks, smart storage or generation) and see the result on the bill.
- Someone sells you on "saving demand" but cannot explain load factor, model your specific rate, or show your future all-in cost per kWh.
- Staggering loads forces a cost (like an extra shift) that destroys the savings.
- You ignore ratchets, set a huge peak in an unusual period, and are then surprised to keep paying for it in quiet months.
You don't manage demand and energy separately.
You manage the relationship between them.
That relationship is load factor.
Every utility rate is a playbook, not a fixed number. Run the plays poorly and you'll earn a high all-in cost on a "good" rate. Run them well and you'll earn a low one on a mediocre rate. Your edge isn't memorizing tariff tables — it's asking the right questions and refusing to say yes to projects you can't see clearly, in numbers, on your bill.
The specific plays operators are running with monitoring, storage, and generation to move their all-in cost per kWh in the right direction — and how to weigh them against your broader operations.
| Demand (kW) | The size of the pipe — your highest average draw in a 15- or 30-min window. Sets the demand charge. |
| Energy (kWh) | What actually flowed — power over time. Sets the energy charge. |
| Load factor | kWh used ÷ (peak kW × hours). High = you fill the van; low = a big van barely used. |
| All-in cost / kWh | Total bill (demand + energy + fixed) ÷ total kWh. The number that actually matters. |
| Coincident stack | Loads running at the same moment, adding up. The highest stack sets your demand. |
| Ratchet | A floor on billable demand, often a % of a prior peak — you keep paying for past peaks. |
| Time-of-use demand | Demand charged only inside a defined peak window (e.g. 4–9 pm). |
| Demand response | Getting paid to cut load during grid-stress events — a second revenue stream. |
Tactical Energy Group · The TEG Podcast · Decoding Demand Charges, Part 1