Solar · savings and payback

When does a solar system pay for itself?

A 4.0 kWp system in London, costing £5,600 installed. Generation comes from the same physics engine as every other page here; the money comes from an engine pinned by 67 hand-computed reference cases.

What do kWp and kWh mean?

kWh — kilowatt-hour: A unit of energy, and the thing your electricity bill charges you for. A 1,000-watt heater running for one hour uses one kWh. A typical home uses somewhere between 2,000 and 10,000 kWh a year depending on where it is and how it heats.
kWp — kilowatt-peak: A unit of capacity — how big the solar array is, not how much it makes. It is the output the panels would produce under standard test conditions: bright, cold and perfectly aimed. Real roofs rarely see those conditions, which is why a 4 kWp array does not generate 4 kW for most of the day. One modern panel is roughly 0.4 kWp, so 4 kWp is about ten panels.
Putting them together: kWp is the size of the system; kWh is what it produces over time. The ratio between them — kWh generated per kWp installed, per year — is the honest way to compare locations, because it strips out how big the system happens to be. It runs from roughly 700 in cloudy high latitudes to over 1,800 in sunny deserts.

Location

Where the system is

Searches use OpenStreetMap. Nothing is stored.

Or pick an example location

Or enter coordinates directly

Latitude

Longitude

Searching or using your location fills these in, so you can always see exactly which point the figures are for.

Climate comes from PVGIS, the European Commission's reference dataset — the same source this model is validated against, for examples and searches alike.

Your system

System size

Kilowatt-peak — how big the array is. Most homes fit 3 to 10 kWp; one panel is about 0.4.

Installed cost £

What you pay in total, fitted. Use your own quote — prices vary hugely by country.

Your tariff

You pay per unit p/kWh

The unit rate on your bill. Each unit you use yourself saves you this much.

You are paid per unit p/kWh

What your supplier pays for a unit you export. Often far less than you pay to buy.

Used at home %

Share of generation you use as it is made. Higher if someone is home in the day.

The long run

Energy prices rise %

Per year, nominal. Nobody knows this; it moves the answer more than the physics.

Discount rate %

What the money could earn elsewhere. Set to 0 to ignore opportunity cost.

System life

Panels are usually warranted for 25 years and often outlast it.

Inverter replacement £

Budgeted at year 13. Set to 0 to see how much omitting it flatters the payback.

8.9 yearsto break even 8.0 years to 11.5 years across 80% of modelled scenarios

£568bill saving in year one

£18,092nominal savings over 25 years, before the purchase price

11.5p/kWhdiscounted levelised cost per unit

Only 35% of what this system generates is worth the full 27p/kWh you pay for grid electricity.

The rest is exported, and exported energy earns far less. That single gap moves the payback year more than panel choice, tilt or efficiency ever will — which is why a calculator that values every generated kilowatt-hour at your import rate will always tell you a happier story than the meter does.

The cash position, year by year

Starts at −£5,600 — the money spent — and climbs as savings accumulate. The step down around year 13 is an inverter replacement, budgeted at £900. Leaving that out is the most common way a payback figure is made to look better than it is.

How much does this depend on what was assumed?

The figures above are one run through one set of assumptions. This is the same calculation run 200 times with the six inputs that matter most drawn from their own distributions — how sunny the year is, how fast the panels degrade, how energy and export prices move, how much of the output is used at home, and what the system actually cost.

6.8 years12.7 years

Across 200 modelled scenarios, this system pays back as early as 8.0 years and as late as 11.5 years, with the middle at 9.6 years. Eighty per cent of the modelled outcomes fall between those two figures.

The base case is not the middle of the range. It comes out at 8.9 years while the middle of the modelled scenarios is 9.6 years — later. That gap is not a rounding artefact: the cost distribution has a longer tail upwards than downwards, and an array declared unshaded can turn out to be shaded but cannot turn out to be better than unshaded. Both push the middle away from the optimistic end.

Every modelled scenario paid back within 25 years. That is a statement about the spread of these assumptions, not a guarantee about a roof.

Each end named by what it means rather than by its percentile — a low payback is a good outcome and a low net present value is a bad one, and calling both of them "P10" inverts one.
Measure	Base case	Favourable end	Middle	Unfavourable end
Payback	8.9 years	8.0 years	9.6 years	11.5 years
Net present value	£3,991	£4,689	£2,836	£1,285

What this range is, and what it is not

It is a sensitivity interval, not a forecast. It says that if the assumed distributions are right, 80% of sampled outcomes land between those figures. It does not say those distributions match reality, because they have never been checked against observed installations — nobody has fed this model a few hundred real systems and asked how often the outcome fell where it said. Until that happens the honest phrasing is "80% of modelled scenarios" — which is why nothing on this page calls any outcome probable, expected or typical. Those are claims about roofs, and this is a statement about a model.

The range is wider than it would be for a shaded-declared system. This page does not ask about shading, so the model treats the array as unshaded and then widens the inputs to cover the possibility that it is not — a one-sided allowance, because an array declared unshaded can only turn out worse.
Self-consumption is a single share you typed. That is the weakest input in the model, and its spread is held at ±10 percentage points regardless of how confident the figure above looks.
Import and export prices are modelled as moving together, weakly. Under an export guarantee the export rate is set by each supplier rather than tied to the retail price, so the two are only loosely coupled here. The payback range barely moves if that coupling is changed; the unfavourable end of the net present value does.
Policy is not sampled. A scheme either continues or it does not, and averaging a binary produces an outcome that cannot happen. Where a rule expires inside the horizon the model runs named branches instead — and this page, which takes your own tariff rather than a scheme, has none to run.

Uncertainty model 0.1.0 · 200 samples, Latin hypercube with a Gaussian copula · seed 3322677785, derived from the scenario rather than a clock, so the same inputs always produce the same range.

Where the generation goes

4060 kWh a year, which is 1015 kWh per kWp installed — the figure to compare locations by, because it does not depend on how big the system is. Where generation peaks above what the household can absorb, the export rate starts deciding the outcome.

Show generation as a table

Monthly generation
Month	kWh
Jan	109
Feb	164
Mar	327
Apr	479
May	519
Jun	533
Jul	553
Aug	480
Sep	401
Oct	253
Nov	151
Dec	90
Total	4060

Every assumption behind these figures

Change any of these and the answer changes. They are stated so you can.
Assumption	Value	Why it matters
System cost	£5,600	£1,400 per kWp — change it above to match your quote
Import rate	27p/kWh	What you avoid paying for each unit you use yourself
Export rate	7p/kWh	What your supplier pays you. Rates vary enormously between countries and suppliers
Self-consumption	35%	Without a battery or a demand profile, this is an estimate, not a measurement
Discount rate	5% nominal	What the money could earn elsewhere
Energy price rise	3% a year nominal	Nominal throughout — mixing real and nominal overstates the return
Panel degradation	0.5% a year	Output falls slowly over the system's life
Inverter replacement	£900 at year 13	Included, because inverters do not last 25 years
VAT	0%	Not modelled. Sales tax and VAT treatment of domestic solar differs by country

What these four figures mean

Each headline figure, said precisely.
Figure	Exactly what it is
Break even	The year cumulative savings first exceed what you paid. Undiscounted, so it ignores what the money could have earned elsewhere.
Year-one saving	Bill reduction in the first twelve months: energy you use yourself valued at your import rate, plus energy you export valued at your export rate.
Lifetime savings	The sum of every year's saving, in nominal money, after lifecycle costs such as the inverter replacement — but before the purchase price. It is not profit. Subtract the £5,600 you paid to get the net gain.
Levelised cost	Total discounted cost divided by total discounted generation — what each unit this system produces effectively costs you. Compare it against your import rate: below it, the system is cheaper than the grid.

How much to trust this

The generation figure is the reliable part: it comes from a model measured against PVGIS, and the deviation is published. The money is far less certain, and not because the arithmetic is shaky — it is exact, and pinned by reference cases — but because it rests on guesses about the next twenty-five years. Energy prices, export rates and how much of your own generation you happen to use will each move this answer more than any modelling error could.

That uncertainty is no longer left as advice to be careful about — it is quantified above, by running the same calculation 200 times with those guesses drawn from their own distributions. The honest reading of the break-even figure is the range rather than the single number, and the range is wider than the precision of "8.9 years" suggests.

The figures that deserve most confidence are still the relative ones: that self-consumption matters more than panel efficiency, and that an inverter replacement is worth budgeting for. Those survive every scenario in the range, because they follow from the shape of a solar day rather than from any particular guess about prices.