Solar panel payback period: how to calculate it properly

Payback is the year in which what you have saved catches up with what you spent. Simple to define, easy to get wrong, and the figure most likely to have been quietly flattered on a quote.

The basic sum takes a minute. The corrections take longer and are what separate a number you can rely on from a sales figure.

The basic sum, with real numbers in it

1. Estimate annual generation in kilowatt hours.
2. Estimate the share you use yourself rather than export.
3. Avoided import equals generation, times that share, times your import rate.
4. Export income equals the rest of the generation, times your export rate.
5. Annual saving is the two added together.
6. Payback equals installed cost divided by annual saving.

Worked through for a 5kW system. The government’s installation cost data puts the 2025/26 median for 4 to 10kW systems at £1,697 per kilowatt, so call it £8,485 installed. At our model’s UK average of 850 kilowatt hours per kilowatt, a south facing 5kW array generates about 4,250 kilowatt hours a year. For a household using half of that directly, with import at the capped rate of 26.11p and export on Octopus Outgoing at 12p:

• Avoided import: 2,125 kWh at 26.11p is about £555
• Export income: 2,125 kWh at 12p is about £255
• Annual saving: roughly £810, so £8,485 divided by £810 is about 10.5 years

If your own version of that sum gives a figure comfortably inside the equipment’s expected life, the case is probably sound. If it lands close to the edge, the corrections below will decide it.

Correction one: the saving does not stay flat

The basic sum assumes year twenty looks like year one. Three reasons it will not:

Panels degrade. Output falls slowly but steadily, a fraction of a percent a year, and manufacturers warrant a remaining percentage after twenty five years precisely because the decline is expected. Over a payback window it shaves months rather than years, but it belongs in the sum.

The inverter will need replacing. Inverter design life is shorter than panel design life, so a mid-life replacement is the normal case, not the unlucky one. A projection covering two decades with no inverter replacement in it is understating cost, and if your basic-sum payback already sits near the likely replacement year, the replacement is part of the payback story, not a footnote after it.

A battery, if fitted, loses capacity. Warranties are written as a remaining percentage after a stated number of years or cycles, which tells you what the manufacturer expects to happen. A battery bought today will not shift as many units in year twelve as in year one, and our model degrades it explicitly rather than pretending otherwise.

Handle these explicitly rather than by shaving a percent off at the end.

Correction two: self consumption is an assumption, not a fact

This is the single most influential number in the whole calculation, and it is usually neither measured nor stated.

Rerun the worked example above at different self consumption shares and watch the payback move, with nothing physical about the installation changing:

• Three quarters self consumed: about £960 a year, payback under 9 years.
• Half self consumed: about £810 a year, payback around 10.5 years.
• One quarter self consumed: about £660 a year, payback nearly 13 years.

Four years of swing from one silent assumption. That is why two calculators disagree about the same system, and why the honest answer to “is it worth it” is conditional.

So: find out what the quote assumed. If it does not say, ask. Then satisfy yourself the assumption describes your household, using your own half hourly smart meter data if you have it. If it does not, redo the sum with a figure that does.

Correction three: keep it in today’s money

You will see twenty five year savings totals built by assuming electricity prices rise by some percentage every year and compounding it. The arithmetic of compounding is why this flatters so effectively: an assumed 5% annual rise, compounded, roughly doubles the claimed saving by year fifteen, and the twenty five year total inflates accordingly.

Treat those totals with suspicion. A forecast of energy prices two decades out is not knowledge, and compounding an assumed rate turns a guess into a large and impressive looking number that cannot be spent.

Figures in today’s money are less exciting and far more useful, because they can be compared against what else you might do with the same capital. Every figure our calculator produces is in today’s money for exactly this reason: when its answer and a quote’s answer differ, the gap is usually their inflation assumption, which you can ask them to remove.

What payback does not capture

Payback is a useful summary and a poor complete answer. It ignores:

• What happens after payback, which is the entire return.
• The shape of the curve, and how sensitive it is to one assumption.
• Any value the system adds to the property, which is real but uncertain.
• The cost of the capital, if borrowed, which can consume most of the saving at a high enough rate.

A year by year projection tells you more than a single year number, because it shows you where the sensitivities are rather than hiding them in one figure.

Sanity checks on any quoted figure

• Does the projection include an inverter replacement?
• What self consumption rate was assumed, and why that one?
• Is the total in today’s money, or has an energy price rise been compounded?
• What degradation rate is applied to the panels?
• If a battery is included, does its capacity fall over the period?

A quote that answers all five is being straight with you. One that answers none has given you a marketing number.

Run it yourself

Every input above is yours: your consumption, your tariff, your roof, your capital. Put them into the ROI calculator, which states its assumptions and lets you change them one at a time, and the payback figure stops being a claim to accept and becomes a result you can interrogate. Change the self consumption share first, because that is the assumption doing the most work.