Two suppliers can quote the same commercial system and produce levelised cost of energy figures that differ by more than the difference between the quotes. That is not because one of them is lying. It is because LCOE is sensitive to inputs that are rarely stated — the discount rate above all — and because the inputs that make the number look best are also the ones that are hardest to defend.
This guide is for the person who has to compare those quotes. It sets out which inputs belong in the calculation, which ones people quietly leave out, and why we will not hand you a single LCOE figure without your numbers.
What LCOE tells you, and what it does not
Levelised cost of energy is the total cost of owning and operating a system over its life, divided by the total energy it delivers over that life, with both discounted to present value. The result is a cost per kilowatt-hour that can be compared against a utility rate or against another generating technology.
That comparison is the whole value of the metric, and it comes with a boundary: LCOE answers what does this energy cost to produce. It does not answer what is this energy worth to you. On a commercial or industrial site those are different questions, because the value of a kilowatt-hour depends on when it is generated and on whether it reduces a demand charge. LCOE contains neither.
The formula, in inputs you can actually fill in
In its standard form:
LCOE = Σ (Ct ÷ (1 + r)t) ÷ Σ (Et ÷ (1 + r)t)
where Ct is the cost in year t, Et the energy delivered in year t, r the discount rate and t runs from 0 to the end of the analysis period.
Nothing in that formula is controversial. Every disagreement about LCOE is a disagreement about what goes into Ct, what goes into Et, and what r should be.
The numerator: every cost that belongs in it
The costs that belong in Ct:
- Year 0 capital cost — equipment, installation, interconnection, and the engineering and permitting that go with them.
- Recurring operating cost — monitoring, maintenance, insurance, and any land or roof lease. These recur every year and most of them escalate.
- Replacement cost in the year it falls — inverters do not last as long as modules. A string inverter replacement around the middle of the analysis period is a real cash flow, and putting it in year 0 or omitting it entirely both flatter the result.
- Residual value as a negative cost in the final year — if the analysis period is shorter than the asset life, the remaining value belongs in the numerator with a negative sign.
The costs that are frequently left out, and that consistently move the answer in the optimistic direction: inverter replacement, escalation of O&M, insurance, and any roof or site work that the electrical scope assumes someone else is paying for. Our breakdown of what drives solar equipment pricing covers where these items sit in a commercial quote.
The denominator: lifetime energy, not year-one output
The mistake here is simpler and more common: taking year-one production and multiplying by the number of years. Modules lose output every year, and over a 25-year analysis period the cumulative difference is large enough to change the ranking between two quotes.
Two further points on the denominator:
- State which production estimate you used. A P50 estimate and a P90 estimate are both legitimate, but they are not the same number and they must not be mixed between the quotes you are comparing. P90 is the appropriate basis when the energy is being relied on to meet a savings commitment.
- Degradation should come from the module datasheet, not from a rule of thumb. The guaranteed figure in the performance warranty is the defensible input, and it differs between cell technologies — the differences are set out in our module technology comparison.
Soiling, shading and availability losses belong in the production estimate as well. A system that is modelled at ideal output and then operated without a washing regime will underperform its own LCOE.
The discount rate: the input that moves the answer most
Of all the inputs, the discount rate has the largest effect on the result, and it is the one most often chosen to produce the answer someone wanted.
The defensible choice is your cost of capital — your weighted average cost of capital if you are a corporate buyer, or your required rate of return on a comparable capital investment if you are not. It is a finance decision, not an engineering one, and it should come from whoever owns the capital budget.
What matters more than the exact number is that it is stated. An LCOE figure presented without its discount rate cannot be checked, and two figures calculated at different discount rates cannot be compared. Ask for it in writing.
Nominal or real: the mistake that doubles the error
The most common technical error in commercial LCOE calculations is mixing a real discount rate with nominal cash flows, or the reverse. Both are valid; mixing them is not.
- Real terms — discount rate and costs both expressed in today’s money, with no inflation. Use this if your cost escalation assumptions are uncertain, and keep everything in today’s dollars.
- Nominal terms — discount rate includes inflation, and costs escalate. Use this if your O&M contracts have stated escalators and your electricity rate assumptions include escalation.
If the comparison you are making is against a utility rate that you expect to escalate, the cleanest approach is to calculate in nominal terms and escalate both sides, or to compare LCOE against a levelised utility rate calculated the same way. Comparing a real-terms LCOE against today’s retail rate understates the saving; comparing a nominal LCOE against today’s rate overstates it.
Five ways LCOE misleads on commercial projects
- It ignores when the energy arrives. Under a time-of-use tariff, a kilowatt-hour generated at 2 pm and one generated at 7 pm have different values. A system with a lower LCOE can deliver less value than one with a higher LCOE if its output falls in cheaper hours. West-facing arrays and tracking are judged on this, not on LCOE alone.
- It does not capture demand charge reduction. For many C&I sites the demand charge is a substantial share of the bill, and shaving a peak is worth far more per kilowatt-hour than the energy itself. Storage that does this well looks mediocre on LCOE and excellent on the bill.
- It says nothing about grid export terms. Whether surplus can be exported, and at what rate, changes the value of every kilowatt-hour the site does not consume directly. Our storage sizing guide treats self-consumption and export separately for exactly this reason.
- It usually excludes tax and incentives. Depreciation, tax treatment and available incentives are project-specific and change over time. They belong in a cash-flow model with your tax advisor, not in a generic LCOE. Confirm what applies to your project and your jurisdiction with them rather than with an equipment supplier.
- It hides the risk allocation. Two quotes at the same LCOE can differ in who carries performance risk, warranty claim handling and equipment replacement. The cost of those differences shows up later, not in the levelised figure.
LCOE, payback and IRR: which answers which question
These three are often used interchangeably. They are not substitutes.
- Simple payback answers how long until the capital is recovered. It ignores everything after that date and ignores the time value of money. It is easy to explain to a board and easy to mislead with.
- IRR answers what return the whole cash flow delivers, over the whole period. It is the most complete single number for an investment decision and the least useful for comparing technologies of different sizes.
- LCOE answers what the energy costs per kilowatt-hour, and is the only one of the three that lets you compare a 500 kW rooftop system against a 2 MW ground-mount or against a utility tariff.
Use LCOE to compare options, IRR to approve the project, and payback to explain it. Our commercial solar payback model walks through the second and third of those, and it is built on the same load data this calculation needs.
Inputs to fix before you compare quotes
Get these settled first and the quotes become comparable. Leave them open and no LCOE comparison is meaningful.
Reference table
Swipe horizontally to view all columns →
| Input | Where it comes from | What goes wrong if it is left open |
|---|---|---|
| Analysis period | Your asset life assumption or term of occupancy | Different periods produce numbers that look comparable and are not |
| Discount rate | Your cost of capital | Largest single effect on the result; cannot be checked if unstated |
| Nominal or real basis | Set once, applied to every input | Mixing the two is the most common technical error |
| Production basis | P50 or P90 from the model, stated | Mixing bases between quotes invalidates the comparison |
| Degradation rate | Module performance warranty | Overstates lifetime energy and understates LCOE |
| Inverter replacement year | Equipment service life | Omitting it flatters every option equally — until it is spent |
| O&M and escalation | Your maintenance provider | Recurring cost compounding over 25 years is not a rounding error |
| Roof or site scope | The structural and site survey | Costs assumed into someone else’s scope still get paid |
One item deserves separate mention because it changes the analysis period rather than a line item: if the roof will not last as long as the array, the system’s useful life is capped by the roof, and a 25-year LCOE overstates the case. Check the remaining roof life before the period is fixed — the sequence is covered in our guide to commercial roof life and solar warranties.
Why we will not quote you a LCOE figure
It would be easy to publish a number. It would also be wrong for your project, because LCOE is not a property of the equipment — it is a property of your capex, your discount rate, your load profile and your tariff.
What we will do instead: supply the photovoltaic modules, inverters, mounting structures, storage and distribution equipment for the configuration you have settled on, confirm stock availability, and agree the model-specific performance data, documents and delivery schedule before the order is placed. If your roof and configuration are still open, our commercial and industrial solar pages cover how the two are decided together.
Send the load profile, the tariff and the discount rate your finance team uses. The LCOE comes back calculated on your inputs, with every assumption written down so that the next quote can be checked against it on the same basis.
BUYER CHECKLIST
Buyer considerations for this scope
Ask for the discount rate in writing
A LCOE figure presented without its discount rate cannot be checked, and two figures calculated at different rates cannot be compared. It is the input with the largest effect on the result and the one most often chosen to produce the answer someone wanted.
Fix nominal or real before anything else
Both bases are valid; mixing a real discount rate with escalating costs is not, and it is the most common technical error in commercial LCOE work. Set the basis once and apply it to every input on both sides of the comparison.
Cap the analysis period by the roof, not by the array
If the roof will not last as long as the modules, the system's useful life ends when the roof does. Check remaining roof life before the period is fixed, or a 25-year calculation will overstate the case entirely.
PRACTICAL ANSWERS
Questions buyers ask about this
What discount rate should I use for a commercial solar LCOE?
Your cost of capital – weighted average cost of capital if you are a corporate buyer, or the required return on a comparable capital investment if you are not. It is a finance decision rather than an engineering one, and the important part is that it is stated in writing, because a LCOE figure without its discount rate cannot be checked or compared.
Should I use P50 or P90 production in the calculation?
Either is legitimate, but you must use the same basis across every quote you are comparing. P90 is the appropriate basis when the output is being relied on to meet a savings commitment. Mixing a P50-based figure against a P90-based one invalidates the comparison, and this is one of the most common ways two quotes are made to look closer than they are.
Does LCOE capture the value of demand charge reduction?
No. LCOE measures what the energy costs to produce, not what it is worth at the time it is produced. On commercial tariffs a kilowatt-hour that shaves a peak can be worth substantially more than one generated in a low-rate period, and that value does not appear in the levelised figure. Storage in particular often looks mediocre on LCOE and strong on the bill.
How long should the analysis period be?
It should match your asset life assumption or your expected term of occupancy, and it should be capped by the remaining life of the roof rather than by the array – if the roof will not last as long as the modules, the system's useful life ends when the roof does. Check the roof first, then fix the period, then compare.
Why is LCOE usually quoted without tax or incentives?
Because depreciation, tax treatment and available incentives are specific to your project and jurisdiction and change over time. Confirm what applies with your tax advisor rather than relying on an equipment supplier, and keep those items in a cash-flow model rather than folding them into a generic LCOE.
EXPLORE FURTHER