How to Choose the Right Electricity Rate Schedule

Utilities publish more than one tariff for the same service territory. The rates in each are approved by the state commission and are not negotiable — but which of them you are billed under is, within the eligibility rules, your choice.

Very few commercial accounts have ever revisited that choice. The schedule was set when the account was opened, often by whoever installed the service, and the business has since changed its shift pattern, added a production line, closed a wing or installed rooftop solar. The tariff has not moved with it.

Why the wrong schedule is so common

Three reasons, all structural.

The default was chosen by circumstance. Somebody picked a general service schedule when the meter was fitted, on the basis of connected load and nothing else.

The site crossed a threshold without noticing. Eligibility is usually set by peak demand, sometimes by voltage level or load factor. A site whose demand grew past a threshold may now qualify for a large-power schedule with a very different structure. Nothing prompts a review when the threshold is crossed.

Nobody owns the question. Operations owns the load, finance owns the bill, and the tariff sits between them. It belongs to whoever is willing to spend a day on it.

What actually determines the answer

Not the headline rates. The shape of your load, expressed as load factor and as the distribution of consumption across the day.

A high load factor site draws close to its peak most of the time. It uses the capacity it is billed for, so a demand-heavy schedule with a low energy rate suits it — the demand charge is spread across a large number of kilowatt-hours.

A low load factor site draws its peak briefly and idles the rest of the time. The same demand charge is spread across far fewer kilowatt-hours, so it lands heavily per unit consumed. That site is usually better off on a schedule that puts more weight on energy, if it qualifies for one.

The relationship is worked through in demand-heavy versus energy-heavy tariffs. The important point here is that it can be computed rather than debated, from two numbers already on the bill.

There is a second dimension beyond load factor, and it matters just as much: when the consumption occurs, not merely how evenly. Two sites with identical load factors can face very different bills if one runs its shift through the expensive afternoon window and the other runs overnight. Load factor tells you whether a demand-heavy structure suits you; the distribution across the day tells you whether a time-varying structure does. A complete assessment needs both, and both come out of the same interval file.

A third consideration is stability. A load factor that is steady all year is a much safer basis for a twelve-month commitment than one that swings between 65 percent in winter and 30 percent in August. Where the seasonal swing is large, the schedule that wins on the annual total may be badly wrong for a quarter of the year, and it is worth knowing that before signing rather than in the third bad month.

The method

The only reliable approach is to price your own consumption under each candidate schedule, interval by interval, using the tariff's own rules.

Step four is the one people skip and the one that makes the exercise trustworthy. A model that cannot reproduce a bill you have already received is not going to predict one you have not.

A comparison, worked

The instructive part of that result is that LP-1 has the lower energy rate — the number a casual comparison would lead with — and is still the more expensive schedule for this load. Comparing rates rather than modeling loads produces exactly the wrong answer, and it produces it confidently.

This is not an accident of the example. Rate design generally works this way: a schedule that recovers more through demand can afford to advertise a lower energy rate, and it is aimed at customers whose flat load makes that trade favorable. Presented side by side without a load behind them, the demand-heavy schedule looks like the better deal to anyone who reads the first number. Which customers it is actually better for is a question the tariff cannot answer, because the answer lives in the customer's own data.

The same logic explains why a utility will rarely tell you which schedule to take. It is not obstruction; the utility does not know your production plans, your expansion, or what you are willing to change operationally. What it is obliged to do in most jurisdictions is make the alternatives available and inform you they exist. The analysis is yours.

Reconciling before comparing

The single step that separates a trustworthy comparison from an expensive guess is reproducing your existing bills.

Build the model of your current schedule first, run your interval data through it, and compare the output with the last twelve statements. Expect small differences from rounding and from riders that update mid-period; do not accept a large one. A model that is 6 percent out on a bill you already have will be at least 6 percent out on one you do not, and the whole point of the exercise is a decision that survives the first bill on the new schedule.

Discrepancies at this stage are informative rather than annoying. They are usually a provision you have not found: a ratchet floor you did not apply, a power factor adjustment, a minimum bill clause, a rider on a determinant you assumed was untouched. Every one of those is something you would otherwise have omitted from the alternative schedules too, in which case the comparison would have been wrong on both sides and looked entirely convincing.

The terms that decide it

Four provisions do most of the work in any comparison, and all four differ between schedules.

The demand determinant. Facility demand, on-peak demand, or both. A schedule billing only on-peak demand is far more attractive to a site with timing flexibility: facility, on-peak and billing demand.

The ratchet. Its presence, percentage and look-back period. A ratchet materially changes the cost of a spiky load and is frequently the largest single difference between two otherwise similar schedules: ratchet clauses.

The power factor clause. Penalty, kVA billing, or nothing. A site with poor power factor can face a substantial difference between two schedules for reasons unrelated to its load shape: power factor penalties explained.

The riders. Different schedules carry different rider sets, and a schedule with a lower base demand rate can be more expensive once its transmission rider is added: riders and surcharges.

Omit any of the four and the comparison is not a comparison.

Time-varying options

Beyond the conventional schedules sit time-of-use and real-time pricing, which price energy by period or by hour rather than at a flat rate. They reward flexibility heavily and penalize inflexibility just as heavily.

They are worth modeling on the same data as everything else, and they need one extra test: how the site would have fared during the most expensive periods of the year, not just on average. A schedule with an excellent annual total and a disastrous fortnight in August is not obviously the right choice. See time-of-use rates for commercial accounts and real-time pricing.

Order of operations

There is a sequencing question worth settling early: change the tariff first, or change the load first?

Change the tariff first when the load has no meaningful flexibility. There is nothing to optimize, so the only lever available is which price structure applies to what you already do.

Change the load first when substantial shape improvements are available cheaply. Sequencing and scheduling can change the load factor enough to alter which schedule is optimal, and switching twice may not be possible inside the minimum stay. See how to reduce peak demand charges.

Model both together when a capital measure is planned. A battery or thermal store changes the profile, and the tariff that suits the post-project profile may not be the one that suits today's. The right sequence is to model the post-project load under every candidate schedule, then choose both together.

There is a version of this that goes wrong often enough to name. A site elects a schedule on the strength of a demand reduction project, the project slips two quarters, and the minimum stay runs its course on a load the schedule was never suited to. Where a switch is contingent on a project, the election should follow the commissioning rather than the approval.

Before electing

Switching has rules of its own: notice periods, minimum stay obligations, and in some cases a restriction on returning. A switch made on a good model and a bad reading of the switching clause can lock a site onto the wrong schedule for a year: switching rate schedules.

And the whole exercise depends on reading the tariff correctly in the first place, which is a skill in itself: how to read a utility tariff book.