Ask three solar developers how long a commercial array takes to pay for itself and you will get three different numbers, all of them defensible. The spread is not dishonesty. It comes from the fact that payback on a solar installation is not one calculation but a stack of them, and reasonable people make different assumptions at each layer. Understanding where those layers sit is the difference between signing a proposal you believe and signing one you merely hope about.
What payback actually measures
Simple payback is the crudest useful figure in the business. Take the net cost of the system after incentives, divide it by the value of the electricity it produces in a year, and you have the number of years until the investment returns itself. It is popular because a facility manager can repeat it in a board meeting without notes.
Its weakness is that it ignores everything after the break even point and everything about the cost of money. A system with a seven year payback and a twenty five year life is a very different proposition from one with a seven year payback and a fifteen year warranty, yet simple payback treats them identically. For that reason most finance teams look at payback first and then move to internal rate of return or net present value before committing capital.
The production side of the equation
Everything depends on how many kilowatt hours the array will actually generate, and that estimate has more moving parts than most buyers expect. Panel nameplate capacity is measured under standard test conditions that almost never occur on a real roof. Actual output depends on irradiance at the site, the orientation and tilt of the modules, shading from parapets, vents and neighbouring structures, module temperature, inverter efficiency, wiring losses and soiling.
Developers handle this with a performance ratio, sometimes called a derate factor, applied to the theoretical maximum. Two proposals for the same roof can differ by a tenth or more in annual production purely because one used a more conservative derate. When comparing bids, ask what performance ratio was assumed and what shading analysis supports it. Modelling tools built on public irradiance datasets, including those published by the National Renewable Energy Laboratory, give a reasonable baseline, but they do not know about the chiller your facilities team is planning to put on the roof next spring.
Production also declines over time. Crystalline silicon modules lose a fraction of a percent of output per year, and manufacturers warrant a minimum retained capacity at year twenty five. A payback model that assumes flat production for two decades is optimistic by a small but compounding margin.
The value side, which is where the real variance lives
A kilowatt hour produced on your roof is worth whatever you would otherwise have paid for it, and commercial tariffs are rarely simple. Many combine an energy charge per kilowatt hour with a demand charge based on the highest fifteen minute or thirty minute peak in the billing period. Solar reliably reduces the energy charge. Whether it reduces the demand charge depends entirely on whether your peak happens to fall in the middle of a sunny afternoon.
A warehouse whose peak is set by refrigeration compressors cycling at four in the morning will see almost no demand savings. A retail building whose peak is set by air conditioning at two in the afternoon may see substantial savings. The same array, the same production, very different economics.
Net metering and export rules
What happens to electricity you generate but do not use at that moment matters enormously. Full retail net metering credits exports at the same rate you pay, which makes oversizing harmless. Net billing or avoided cost export credits pay considerably less, which means every exported kilowatt hour is worth a fraction of a consumed one. Under those rules, the right system is often smaller than the roof allows, sized to match daytime load rather than to fill the available area.
Escalation assumptions
Most proposals assume utility rates rise each year, and the assumed escalation rate has a large effect on the twenty five year total. A model using a high escalator will show spectacular lifetime savings. Ask to see the figure, ask what it is based on, and ask to see the same model run with no escalation at all. If the project only works when rates climb steeply, you are buying a bet on energy markets alongside the hardware.
Costs that proposals sometimes understate
The installed price covers modules, inverters, racking, wiring, labour, permitting and interconnection. Several real costs sit outside it. Inverters have a shorter service life than modules and usually need replacing once during the system life. Monitoring platforms may carry a subscription. Insurance premiums can change. If the roof is within a few years of replacement, the sensible sequence is to reroof first, because removing and reinstalling an array later is a meaningful expense that nobody budgets at the outset.
Operations and maintenance is modest for a fixed rooftop system but not zero. Periodic inspection, occasional cleaning in dusty or agricultural settings, vegetation control for ground mounts and the cost of someone noticing when a string stops producing all belong in the model.
Putting it together honestly
A credible payback figure comes with its assumptions attached. It states the assumed annual production and the performance ratio behind it. It separates energy savings from demand savings and says which tariff components it touches. It names the export rules it assumes. It shows the result with and without rate escalation. It includes inverter replacement. And it says plainly which incentives it relies on and whether they are secured or merely expected.
When a proposal does that, the number it produces may be less impressive than a competitor’s, and it will be far more likely to survive contact with your actual utility bills. The purpose of the calculation is not to justify the project. It is to find out whether the project justifies itself.
Frequently asked questions
Is a shorter payback always better?
Not necessarily. A short payback achieved by assuming aggressive rate escalation or generous export credits is less valuable than a slightly longer one built on conservative assumptions. Compare the assumptions before comparing the results.
Why do two bids for the same roof produce different production estimates?
Usually because of different derate factors, different shading analysis, or different assumptions about module orientation and row spacing. Ask each bidder for the performance ratio they used.
Does solar reduce demand charges?
Sometimes. It reduces them only if your billing peak occurs during hours of strong solar production. Reviewing interval data from your meter is the only reliable way to know.
Should the array be sized to fill the whole roof?
Only when exports are credited at or near retail rates. Where export compensation is low, sizing the system closer to daytime consumption usually produces better returns than filling the available area.
What happens to output as the system ages?
Production declines gradually, by a small fraction of a percent each year for crystalline silicon modules. Manufacturers typically warrant a specified retained output at twenty five years, and a sound financial model reflects that slope rather than assuming constant generation.