What the solar ROI calculator actually models

Most solar calculators ask for your postcode and roof direction, then hand back a payback figure with no working shown. That number is doing a lot of quiet guessing on your behalf, and the guesses are where the disagreement between two calculators usually lives.

This page describes what ours takes into account, with the actual assumptions stated, so you can judge how much weight the answer deserves. The model is a deterministic calculation: the same inputs always produce the same projection, and every step below is arithmetic you could reproduce.

What goes in

The model works from the inputs you give it:

• System size, in kilowatts peak, and the direction the panels face
• Battery capacity, if you are considering storage
• Annual electricity usage, in kilowatt hours
• Home occupancy during working hours, which drives how much generation you use directly rather than exporting
• EV charging, which changes both the size and the shape of your demand
• Whether the system is financed, since borrowing changes the return

How generation is estimated

The starting point is 850 kilowatt hours per year for each kilowatt of panels, which is a deliberately middle-of-the-road UK figure: the sunniest parts of the southwest sit meaningfully above it and northern Scotland below it. Direction then scales it. South facing panels keep the full figure, southeast and southwest lose five percent, east and west about fifteen, and a due north roof drops to around two thirds of the south facing output.

Panel output also fades slightly with age, so the projection degrades generation by a small fraction of a percent each year rather than holding it flat for twenty five years. These are conservative middles, not promises: a steep south roof in Cornwall will beat them and a shaded north roof in Aberdeen will miss them, which is why the result is a comparison tool rather than a forecast.

The split that decides the money

Every generated kilowatt hour is worth one of two amounts, and the gap between them is the single most important thing the model tracks.

Electricity you use as it is generated replaces an import. On the current price cap that import costs 26.11p per kWh, per Ofgem’s July to September 2026 rates. Electricity you export earns an export tariff instead. Export rates are set by suppliers and Ofgem requires only that they are above zero; a widely held example, Octopus Outgoing, pays 12p per kWh. At those rates a self consumed unit is worth more than twice an exported one, so the self consumption split moves the payback more than most people expect.

The model estimates that split from your system size, usage and battery using a published self consumption model rather than a house style guess. For a typical household it lands just under half without a battery and around seventy percent with one, and it shifts with your occupancy answer: a home worker uses more of the midday generation directly than a house that is empty until six.

Import and export rates are modelled separately throughout, because they are different numbers and they do not move together. Time-of-use tariffs are modelled with their own peak and off-peak structure rather than flattened into an average.

What the battery actually does in the model

A battery earns its keep twice: it stores midday generation for the evening, converting export-value electricity into import-value electricity, and on a time-of-use tariff it can also charge cheaply overnight and discharge at peak rates.

The model deliberately does not treat the battery as a lossless, immortal box, because that is how optimistic projections are made:

• Usable capacity is less than nameplate capacity, because batteries reserve a margin rather than cycling to empty.
• Round trip efficiency is below ninety percent, so every stored unit comes back slightly smaller than it went in.
• Degradation reduces effective capacity year on year, the way real batteries fade, rather than holding year fifteen equal to year one.
• Replacement is accounted for within the twenty five year projection, which is why the savings curve visibly dips partway through. A projection whose curve never dips has quietly assumed a battery outlives its warranty by a decade.

Whether that trade is worth it for your usage is exactly the kind of question the calculator is built to answer, because both sides of the comparison carry the same assumptions.

Everything is in today’s money

Two choices separate this projection from the ones on sales quotes, and both are deliberate.

First, future savings are not inflated by an assumed energy price rise. A saving in year twenty is stated in the same pounds as year one. Installer projections routinely compound an assumed annual price increase, which is how a twenty five year total is made to look spectacular; we decline to guess future prices, because nobody can.

Second, no discount rate is applied either. Choosing one is the most opinionated number in any projection, and small changes swing the payback year by years. Declining to choose is the assumption free position, and it keeps the output a sentence a homeowner can act on: what you would save each year, in today’s pounds.

The practical consequence: if a quote’s projected total looks far better than ours for the same system, the difference is almost always their assumed energy inflation, not their hardware. Ask what price rise the projection assumes, then ask to see it at zero. The reasoning behind what a system actually costs and what belongs in the quote is kept to the same standard: sourced figures, stated assumptions.

What it does not do

Being clear about the limits is part of the answer:

• It does not read your actual half-hourly consumption data. It works from annual usage and an occupancy pattern, which is an approximation of a thing that varies enormously between households.
• It does not know your roof. Shading, pitch, and anything that puts a chimney between your panels and the sun are not modelled.
• It does not quote for installation. Real prices vary by installer, region and what your loft looks like, so it uses typical component costs; put your actual quote against its output rather than treating the modelled cost as a price.
• It does not predict future energy prices, because nobody can.

How to read the result

Treat the payback year as a comparison tool rather than a promise. It is most useful for questions of the form “is a battery worth it for me”, or “does going from four kilowatts to six change the picture”, where both sides of the comparison carry the same assumptions and the assumptions therefore partly cancel out.

It is least useful as a single number to take to a lender.

If you want to see how sensitive the answer is, change one input at a time and watch which ones actually move the curve. On most households, occupancy during the day moves it more than roof direction does: the difference between using half your generation and a third of it outweighs the difference between a south and a southeast roof. That is also the order in which to spend effort improving the real system: shifting usage into daylight hours is free, and reorienting a roof is not.

Try the calculator and see what your own numbers do.