Stepwise.

Tools · Household bills

Why did my bill go up?

Search that question and every answer is an article explaining, in general terms, that bills go up when it is cold. The method that answers it properly already exists — it is standard energy accounting, and it is sold as software to building managers. There has never been a household version.

This is one. Enter two bills and it separates the increase into the unit rate, your usage, the standing charge, and the part that exists only because rate and usage moved together — four figures that add up to the whole change exactly. Then it goes further and splits the usage itself into what the weather explains, what a longer billing period explains, and what neither of them does. That last number is the one you actually want, and no bill, no usage graph and no article will give it to you.

Loading…

How it works, so you can check it

The money split is an identity, not an estimate. A bill is rate × usage, so a change in it expands into three terms: the rate change at the old usage, the usage change at the old rate, and the product of the two changes. Written out,r₁U₁ − r₀U₀ = (r₁−r₀)U₀ + (U₁−U₀)r₀ + (r₁−r₀)(U₁−U₀). It is the expansion of a product and holds for any numbers you put in. Most presentations bury that third term inside one of the other two, which overstates whichever one it lands in; it is reported separately here because it genuinely belongs to neither.

Billing periods are not months. They run from about 26 to 35 days. The same identity applied to usage — days × usage-per-day — separates the part of a rise caused purely by the calendar from the part caused by using more each day. A six-day-longer period is over 20% more energy at identical behaviour, and nothing your utility shows you accounts for it.

The weather model has a name. Fitting usage = baseload × days + slope × degree days across your bills separates weather-independent consumption — refrigeration, lighting, standby, hot water — from the energy the cold costs you. This is the PRISM method, the Princeton Scorekeeping Method, formalised in ASHRAE's Inverse Modeling Toolkit, and it is what “weather-normalised consumption” means wherever the phrase is used professionally.

The model tells you when it has failed. A fit that puts weather-independent consumption below zero is claiming your home uses negative energy in mild weather. A fit with a negative slope is claiming it uses less as it gets colder. Neither is possible, so either result is proof that something changed between those bills that weather and calendar cannot explain — and the tool says so instead of printing the impossible number. That refusal is the most useful output it has.

Questions people actually ask

Why did my electricity bill double?
Almost never for one reason. A doubled bill is usually a rate rise, a cold snap, and a longer billing period arriving together, and each of those is modest on its own. This tool separates them arithmetically: it computes what the rate change alone would have cost at your old usage, what the usage change alone would have cost at your old rate, and what part exists only because both moved at once. The four figures it produces add up to the whole increase exactly, so nothing is hidden in a rounding.
Is anything I enter uploaded?
No. Everything here is computed by this page in your browser. There is no server that receives it, and nothing is stored unless you explicitly save it.
What are heating degree days and where do I find them?
A heating degree day measures how cold it was and for how long: each day contributes the number of degrees the average outdoor temperature fell below a base temperature, and the days are summed across the billing period. The base is conventionally 65°F in the United States and 15.5°C in the United Kingdom. Many utilities print the figure for your billing period directly on the bill — Nicor Gas, We Energies and National Grid all do — because it is the honest answer to "was it me or was it the weather". If yours does not, free degree-day data is published for most weather stations.
My billing period was longer than usual. How much does that actually matter?
More than almost anyone expects, because billing periods are not months. They commonly run anywhere from 26 to 35 days, and a 34-day period next to a 28-day one is 21% more energy at completely identical behaviour. It is frequently the single largest component of a bill that looks shocking, and it is the one component no usage graph on any utility website shows you.
What does the "neither of those" figure mean? Is my meter broken?
It is the energy that the weather and the length of the period together do not account for — the residual after the model has explained what it can. It is a prompt to look, not a verdict. The ordinary explanations are a new appliance, someone at home more than usual, a heating system running longer than it should, or a previous bill that was estimated too low. A faulty meter is possible and is far down the list of likely causes.
Why does it want more than two bills?
Because with exactly two, the weather model has no degrees of freedom. Two equations and two unknowns give an exact fit whether or not the model is true, in the same way a straight line always passes perfectly through two points — so a perfect-looking result proves nothing. The tool says this rather than reporting a flattering R² of 1. From three bills onward the fit can be wrong, which is what makes it worth having, and a year of bills makes it genuinely informative.
One of my bills was estimated rather than read. Does that change things?
Considerably, and it is the most common innocent explanation for a shocking bill. An estimate that came in too low is repaid by the next actual reading, which produces a large increase that has nothing to do with how much energy anyone used. Tick the estimated box and the tool flags it rather than attributing the rise to your behaviour.
Does this work for gas and water too?
Yes. The money decomposition — fixed charge, unit rate, usage, and the part caused by both moving at once — is arithmetic and applies to any metered utility. The weather model is meaningful for anything used for heating or cooling, which makes it strongly relevant for gas and often for electricity, and largely irrelevant for water. Where degree days are not supplied the tool reports the usage change whole rather than inventing a split.

Where it stops

  • · It computes; it does not advise. It will not tell you to switch supplier, insulate anything, or replace an appliance.
  • · The effective rate is the consumption charge divided by usage, so on a tiered or time-of-use tariff a change in the mix of your usage appears in the rate term rather than in its own.
  • · A tariff that changed partway through a period produces a blended rate, which this treats as a single rate for that period.
  • · Degree days computed from a monthly average temperature understate the real figure when temperatures straddle the base. Where your utility prints degree days, those are better.
  • · It cannot see anything your bills do not record. A meter fault, a shared supply and an unrecorded occupancy change all land in the same unexplained residual.

More tools like this

Expense tracker

Log spending in seconds — monthly totals by category, a budget bar and an honest pace projection.

Renovation budget planner

Room by room: planned vs quoted vs paid, contingency built in, and variance showing where the project grew.

Net worth tracker

Assets minus debts, snapshotted over time — the trend, the category mix, and no bank linking ever.