Commercial solar payback is quoted in a range – typically somewhere between three and seven years depending on the market – and that range is so wide because it is an output, not a property of the equipment. Two facilities with identical systems can land years apart.
This guide is about building the model so the number you present survives contact with reality. It is written for the people who have to defend a business case to a finance committee or a project owner.
Self-consumption ratio drives more than generation does
The most common error in a commercial model is starting from how much the system will generate. The number that matters is how much of that generation displaces a purchased kilowatt-hour, and at what rate.
Commercial facilities with daytime operations – warehouses, cold storage, manufacturing, retail – typically self-consume 70 to 90 per cent of what they generate. Residential systems, by contrast, often self-consume 30 to 40 per cent because occupancy and generation are out of phase. That difference is the single largest reason commercial payback is faster than residential.
Exported surplus is almost always worth materially less than avoided consumption – in many markets a fraction of the retail rate. So model the export portion at the export rate, not the retail rate. This one correction removes most of the optimism in a first-draft model.
Incentives change the numerator, not the physics
In the US market, the base commercial credit sits at 30 per cent, with adders available for domestic content, energy community location and other categories that can lift the effective credit substantially. Accelerated depreciation and, in some markets, first-year allowances do similar work.
Two cautions. First, confirm eligibility before the incentives enter the model – the sourcing rules we cover in our FEOC guide now gate access to the credit itself. Second, model the project without incentives as well. If the business case only works with the credit, that is a material risk to disclose to the decision-maker, not a detail to bury.
Model degradation, not year-one output
A system does not produce its year-one number for twenty-five years. Modules lose a fraction of a percent of output annually – n-type product typically warrants around 0.4 per cent per year – and a performance warranty that still guarantees the mid-eighties per cent of original output at year twenty-five is the market norm.
Applying a straight-line annual degradation factor to each year’s savings rather than holding year-one savings flat typically moves the payback date by a meaningful margin. It is a two-minute change to a spreadsheet and it is the difference between a model and a brochure.
Include the costs that get left out
These are the line items most often missing from a first-draft model:
- Inverter replacement – inverters generally do not last the full life of the array. Budget one replacement, typically in the second decade, and note whether the warranty covers labour or only the unit.
- Operations and maintenance – annual inspection, cleaning where soiling is real, monitoring subscription, vegetation and roof upkeep.
- Insurance and tax implications – premium changes and, in some jurisdictions, property tax treatment.
- Roof or structure work – if the roof needs remediation before or during the installation, it belongs in the project cost.
Electricity price escalation cuts both ways
Most quoted payback figures assume flat electricity prices. That is conservative in one direction and misleading in another: if retail rates rise modestly year on year, the savings stream grows and payback shortens – typically by several months to a year under a moderate escalation assumption.
The honest approach is to present a base case at flat prices and a sensitivity case at a modest escalation rate, and state both. Do not present the escalated case as the headline without labelling it.
Storage changes the model in two directions
Adding storage increases the self-consumption ratio – often meaningfully, and in some configurations towards the mid-nineties – but it adds capital cost and its own degradation and replacement assumptions. Whether it improves the payback depends on the tariff structure.
- Where the bill is dominated by energy charges, storage pays by shifting surplus into evening consumption.
- Where the bill is dominated by demand charges, storage pays by shaving peaks – and in that case the relevant specification is power (kW), not energy (kWh). We cover that distinction in the BESS specification guide.
Run the sensitivities that actually change decisions
Reference table
Swipe horizontally to view all columns →
| Variable | Direction | Typical effect on payback |
|---|---|---|
| Self-consumption ratio | 90% vs 50% | Largest single driver |
| Incentive eligibility | Full vs none | Can nearly double the period |
| Electricity escalation | Flat vs moderate | Shortens by months to over a year |
| Export rate | Retail vs actual export value | Removes common over-optimism |
| Inverter replacement | Included vs omitted | Small but real, lands mid-life |
| Degradation | Modelled vs ignored | Moves the back end of the curve |
| Financing structure | Cash vs loan vs PPA | Changes cash timing, not physics |
Payback and ROI answer “is this worth doing”; they do not tell a board what the energy costs per kilowatt-hour. Buyers usually end up needing both numbers, so the LCOE version is set out separately in how to calculate LCOE for commercial solar and storage.
Putting the numbers together
The sequence that produces a defensible figure is: load profile, then self-consumption, then incentives confirmed against eligibility, then degradation and lifecycle costs, then sensitivities. Our commercial sizing guide walks through the first step; the commercial and industrial solar page shows typical configurations. If you want a model built against your actual tariff and interval data, send it through the contact page.
BUYER CHECKLIST
Buyer considerations for this scope
Which inputs actually change the answer?
The load profile, the tariff and any demand charge, the incentive and tax treatment, remaining roof life, degradation, and the O&M assumption. Of these, the load profile is what separates a real model from a rough estimate.
Why is there no generic payback figure for this?
Because it depends on when the building uses power and how the utility charges for it. A single published payback period would be wrong for almost every site, and quoting one is the fastest way to lose credibility with a commercial buyer.
What assumption is being made about the roof?
Remaining roof life, and whether a replacement should precede the installation, is usually the largest unpriced risk in a commercial rooftop project. It belongs in the model as an explicit line rather than a footnote.
PRACTICAL ANSWERS
Questions buyers ask about this
Which input moves the payback most?
The self-consumption ratio, because exported energy is usually credited at a lower rate than energy you avoid buying. Generation total matters less than how much of it displaces a purchase.
Should degradation be modelled?
Yes. Modelling year-one output as if it repeats for 25 years overstates revenue. A linear degradation assumption in the 0.4 to 0.55 percent per year range is what the performance warranty supports.
What costs usually get left out?
Insurance, monitoring subscriptions, inverter replacement inside the analysis period, and maintenance of the roof or the ground around the array. Leaving those out is why a model beats a real invoice later.
Do incentives change the model?
They change the numerator, not the physics. Apply them as a separate line so you can see the payback with and without them, because policy risk should not be hidden inside the generation assumption.
EXPLORE FURTHER