A battery does not make more electricity

Solar arrives in the middle of the day. Most households use the least electricity in the middle of the day. Whatever you do not use as it arrives is exported to the grid — usually for a good deal less than you pay to buy it back at teatime.

So the question is not how much the roof makes. It is what share of it you catch. That share has a name in the method — self-consumption — and it is a percentage of generation, not a number of kilowatt-hours.

This matters more than almost anything else, and it is the input people expect least. Two identical roofs on two identical streets, using the same units a year, catch very different shares of their own solar depending on whether anyone is home at noon.

Your roof and your bill, in three different households — before any battery:

Notice the method never asks how many people live there. Annual consumption is treated as the proxy for that; occupancy is only about when the house is awake. You have left this unanswered, so the middle table is being used — which is what the method itself says to do.

A big roof on a small bill exports most of what it makes. The same roof on a large bill covers less of it proportionally but wastes almost nothing. Both facts come out of the same table.

You have not given a figure, so the method is using 3,500 kWh — the British average, and explicitly a stand-in rather than a measurement of this house. A real number off twelve months of bills changes the answer more than most people expect.

Bigger always catches more. It never catches proportionally more, and the curve flattens hard — a second battery does not do what the first one did.

The table stops at 15.1 kWh of usable capacity. Past that the method has no column and gives no answer — which is a real limit worth knowing before someone quotes you three stacked units.

Usable is not the number on the box. Where a datasheet gives only a nominal capacity, the method multiplies it by the depth of discharge — and where that is not stated either, it must be 90% for lithium and 50% for lead acid.

This is the part that gets sold past. A battery held full for a power cut is not a battery storing your solar. Capacity charged overnight on a cheap tariff is capacity that was not empty at noon. The same kilowatt-hour cannot do two jobs.

The standard is direct about it: decide the proportion of the usable capacity for each, and the two cannot add up to more than the battery. Drag the split and watch the two numbers move against each other.

Note what the right-hand number is not: money. The method puts no tariff on those kilowatt-hours anywhere — it caps the assumed cycling at twice a day and stops. Anyone converting them into an annual saving has left the standard behind, and should tell you which tariff they used.

The figures above exclude electric vehicle charging, heat pumps, immersion diverters and electric space or water heating — all of them, entirely. A house with any of those uses its solar rather differently, and the method says so rather than guessing.

There is a ceiling on the whole thing: self-consumption may never be quoted above 95% of what the roof generates. It is a hard limit in the method, on the grounds that there is always some midsummer afternoon when the car is out, the tank is hot and the export meter spins anyway.

Nor is any of this a prediction for your house. In the standard's own words, it is the average across a sample of homes with similar occupancy, similar consumption and a similar system — for the first year, before anything degrades and before anyone changes their habits because they can suddenly see their own generation on an app.

See this as a plain calculator, with the sources →