Generation is the easy half of the sum and it is the half every calculator gets right. The half that decides whether a scheme is worth doing is what share of that generation a school uses at the moment it is made, and on a school that share is structurally lower than on a warehouse. This calculator puts the assumption on the screen instead of hiding it inside the answer.
A school buys electricity on a different rhythm from a warehouse. Maintained schools in England must be open to pupils for at least 190 days a year under the Education (School Day and School Year) (England) Regulations 1999, and academies broadly follow the same shape. That leaves the buildings shut for the other 175 days, and those days are not spread evenly. Six of the weeks fall across late July and August, which is the highest yielding stretch of the year.
We can put a figure on that. Taking the EU PVGIS v5.2 monthly generation profile for each of the 397 towns we hold one for, and weighting it by a 191 day school year, only 49 percent of a year's generation lands on a day a term time only school is open at all. The figure runs from 49 to 50 percent across the whole country, so this is not a southern problem or a northern one. It is the shape of the school year. The summer break alone accounts for about 16 percent of the year's output, and the buildings are empty for all of it.
The second effect sits inside the day. A school's load falls away from about three o'clock, while the array carries on generating until dusk. In June that tail is long. So even on an open day, a roof sized array pushes more into the grid than a building working two shifts would. Neither effect is a reason not to do it. Both are reasons to size the array against the load rather than against the roof, and to be honest about the split before anyone signs anything.
The calculator reads the modelled yield for your own town rather than a national average, because across the 397 towns we cover that figure runs from 704 to 1028 kWh per kWp a year, a spread of about 46 percent. Enter the roof area, or the array size if a designer has already given you one, and the annual electricity figure from your supplier's statement. Then tell it how the buildings are used out of term, which is the input that moves the answer most.
One conversion is worth stating plainly, because a lot of calculators get it wrong in the optimistic direction. A kWp is rated at 1,000 watts per square metre of sunlight, and a commercial module converts roughly a fifth of what lands on it, so a kWp needs about five square metres of roof when panels sit flush on a pitched or shallow slope. On a flat roof the rows have to be tilted and spaced so they do not shade each other, which pushes it to around seven and a half. A 1960s teaching block therefore carries a good deal less array than its footprint suggests, and any tool that tells you otherwise is quietly inflating every figure downstream of it.
Self-consumption is the share of generation a site uses at the moment it is made, and on a rooftop scheme it is what the whole appraisal turns on. Electricity used on site displaces electricity you would otherwise buy at your delivered day rate. Electricity exported earns a rate set by your supplier that is materially lower. Two schools with identical roofs and identical arrays can sit a long way apart on payback purely because one of them runs a pool and the other locks up on the third Friday in July.
The default here is built from two steps and one stated assumption, all of them visible in the calculator rather than folded into a single constant:
The model is deliberately conservative in one direction. It counts generation on closed days as entirely exported, when in practice servers, fridges, frost protection and security lighting keep a small load running through the holidays. So the measured figure normally comes out a little above the modelled one. We would rather the surprise at survey were in that direction.
The Smart Export Guarantee obliges larger licensed electricity suppliers to offer at least one tariff that pays for exported electricity. It replaced the export element of the Feed in Tariff for new installations from January 2020. Ofgem administers the scheme and reports on it, and this is the part that catches people out: Ofgem does not set the rate. Each supplier sets its own price, structure and terms, and the offers vary widely, which is why no export figure is printed anywhere on this site. You put your own in, or you leave the field blank and read the kWh.
Two practical points follow for a school. The first is that an export rate sits well below what the site pays to import, so a kWh exported is worth a fraction of a kWh displaced. That is the arithmetic that makes availability matter more than headline generation. The second is that eligibility depends on the installation being certified under the Microgeneration Certification Scheme or an equivalent scheme for the size of the array, which is one of several reasons the installation is carried out by an MCS-certified partner rather than by us. If the school buys its electricity through a public sector framework, the export arrangement may need to be agreed alongside it, and that is worth raising early.
Battery storage moves generation across time rather than creating more of it. That is a real gain and a limited one, and it is the first of four levers that matter more than the choice of module.
The single largest lever, and the cheapest. Every kWp added past the point where midday output meets the daytime load earns an export rate instead of a saving. The sizing strip in the calculator shows the taper live: cut the array and the used on site percentage climbs.
A battery carries the after three o'clock generation into the evening load and, sized larger, a weekend into the following week. It cannot carry July into October, so it improves the term time picture rather than the summer one. The Department for Education's own Education Estates Strategy of 16 February 2026 estimates that a typical school could save up to 25,000 pounds a year with solar panels and complementary technologies such as batteries installed. Read the conditional in that sentence, because it is doing the work: it is an estimate with the battery attached, not a figure panels reach alone.
Lettings, a summer play scheme, a community pool, a nursery, a kitchen producing for another site: anything that puts a load in the building in August changes the economics more than any equipment decision. If the trust is already considering it, model it before you size the array.
An array on a sports hall serves the sports hall's meter first. Sites with several supply points, or a leisure block on a separate account, often find the constraint is the wiring between buildings rather than the roof. The survey maps that before a layout is drawn.
Devolved formula capital is an annual allocation a school receives to spend on its own capital priorities. The Department for Education publishes the formula it is calculated from, which means you can work out your own figure rather than waiting to be told it. For 2026 to 2027 it is four thousand pounds per school plus eleven pounds twenty five for every weighted pupil, where primary pupils count once and secondary pupils count one and a half times, multiplied by 1.08 for voluntary aided schools only.
The panel below is that formula and nothing else. It is arithmetic from a published source rather than a model, so unlike the array calculator above it carries no assumptions of ours at all. It opens on the Department's own worked example of 50 primary and 500 secondary pupils. Switch the type selector to voluntary aided and it returns 14,040 pounds, which is the figure printed in the guidance, so you can check our arithmetic against the source before you trust anything else on this page.
Three points of detail matter for planning. The allocation is paid as a single payment in June. It can be spent across three financial years, the first being the year it is paid, so it does not have to be committed in the year it arrives. And academies within a trust can pool their allocations with each other's agreement, which is what turns a set of small per school figures into a number large enough to do something structural with.
One thing we will not tell you is whether this can be spent on solar. The condition funding guidance does not name solar, photovoltaics, renewables or decarbonisation anywhere in its devolved formula capital provisions, so it neither permits nor prohibits it by name. What it does say is that the allocation is for the school's own capital priorities, which the school itself sets. That is a question for your responsible body and your auditors, not for us, and anyone telling you the answer is a straightforward yes is going beyond what the guidance says. Our funding page sets out the capital and grant routes in full, including the ones that do name this kind of work.
An academy trust holds several sites under one set of accounts, and that changes the question. A single school asks whether its own roof works. A trust asks which of its roofs works first, and the answer is rarely the largest one. It is usually the site with the highest daytime baseload against the roof it has, which in practice means the site with a pool, a commercial kitchen, a data room or an attached nursery.
Run the calculator once per site and compare the used on site percentages rather than the generation figures. A 400 kWp array holding 30 percent and a 150 kWp array holding 45 percent are very different propositions once the export rate is applied, and the smaller one often reaches a governors' meeting in better shape. Our roof survey page sets out what gets measured, and the procurement page covers how this is bought under public procurement rules.
Next: what drives the cost, how it gets paid for, the grant and capital routes, or the modelled yield for every town we cover across 48 counties. The mean across them is 886 kWh per kWp a year.
Send the postcode, the rough roof area and last year's electricity figure, and we will come back with what the buildings can carry and what the split is likely to be. No charge and no obligation.