A payback period is a single number produced by six inputs, five of which have nothing to do with the equipment.
Commercial review also depends on how people organise decisions and handoffs; for a separate workplace-management perspective, see this reference.
Reviewed August 9, 2026.
The six inputs
One. Annual production. The yield figure, at some P-value that should be stated, declining with degradation. The only input about the plant.
For broader context on solar procurement, equipment markets and current industry practice, consult IEEE.
Two. The value of a kilowatt-hour you consume yourself. Your retail tariff, including the parts of the bill that scale with consumption and the parts that do not.
Three. The value of a kilowatt-hour you export. Usually much lower, sometimes near zero, occasionally negative in periods of oversupply.
Four. The split between the two. What proportion you actually use on site, hour by hour.
Five. The electricity price path over the projection. The widest honest range of any input and the one most often given as a single confident figure.
And six. The total installed cost, including everything outside the quote.
What each does to the answer
Production is the least sensitive. A ten per cent error in yield moves payback by roughly ten per cent — meaningful and proportionate.
The self-consumption split is far more sensitive, because it multiplies against a price difference that can be three or four to one. Getting the split wrong by twenty points can move payback by years.
And the price path dominates everything at long horizons. Assuming steady real increases against assuming flat prices produces two different investments from identical hardware.
So a proposal that models yield carefully and states the price path as one number has spent its effort on the input that matters least.
The question that exposes it
"What does payback look like if electricity prices stay flat in real terms?"
Any competent supplier can produce it in minutes. The gap between that answer and the headline figure is the whole of what the price assumption is doing, and seeing both is more informative than arguing about which is right.
Then ask the same at P90 rather than P50, and the two together bracket the realistic range.
What payback leaves out
The cost of capital. Money spent now against savings arriving over fifteen years is not a simple division, and a payback period ignores discounting entirely.
Inverter replacement, which lands mid-life and is a capital event.
Ongoing costs — insurance, monitoring, maintenance, cleaning.
Residual value and decommissioning at the end.
And the alternative use of the money, which is the actual comparison an investor makes.
Why the figure persists anyway
Being fair.
It is comprehensible. "Seven years" communicates instantly in a way that a discounted cash flow does not, and comprehensibility has real value in a decision involving non-specialists.
It is roughly directionally right for a simple case with stable consumption.
And it is what buyers ask for, which is why every proposal contains one.
Use it as a screen, not as the decision. The deeper objection is that payback is the wrong question entirely, and that is a separate argument.
The short version
- Six inputs: production, the value of self-consumed energy, the value of exported energy, the split between them, the price path, and total installed cost
- Only the first is about the plant, and it is the least sensitive of the six
- The self-consumption split multiplies against a price difference of three or four to one, so a twenty-point error moves payback by years
- The price path dominates at long horizons and is most often stated as a single confident number
- Ask what payback looks like with flat real prices, and again at P90 — the two bracket the realistic range
- Payback ignores the cost of capital, inverter replacement, ongoing costs, residual value, and the alternative use of the money