The yearly number is a model
Kilowatt-hours measure energy. The yearly total in a proposal is a forecast, not a measurement of what your future system will produce. The estimate depends on tilt, compass direction, location, equipment, and shading. Change any input and the lifetime savings story moves with it.
Ask how shade was accounted for
Trees, chimneys, neighboring roofs, and winter sun angles can reduce energy production. Ask how the proposal accounts for them. If an independent model estimates less energy, the projected savings may be lower and recovering the purchase cost may take longer.
What to compare
Look at the yearly estimate, the assumed shading loss, tilt, and direction. If two quotes use different roofs in the model — one with shade, one without — they are not quoting the same system even if the panel count matches.
Compare estimates using the same inputs
PVWatts estimates energy output using inputs including system capacity, tilt, azimuth, and losses. Its output is a model result, not a production guarantee. PVWatts calculator and model inputs (U.S. national laboratory calculator; accessed September 18, 2026)
Illustrative example: one 8 kW proposal estimates 12,000 kWh per year and another estimates 10,800 kWh. That is a 1,200 kWh difference, or 10% of the first estimate. Ask what inputs explain it before assuming either is right.
- Confirm the same DC system size and roof planes are modeled.
- Compare shade, losses, tilt, direction, and weather assumptions.
- Ask whether annual energy refers to system output or energy delivered after other losses.
- Keep the forecast separate from any written production guarantee.
Translate capacity into an annual-energy question
A system’s kilowatt rating describes capacity. Its yearly kilowatt-hour forecast describes energy. For an illustrative 8 kW DC system, 12,000 kWh per year corresponds to 1,500 kWh per installed DC kW. An estimate of 10,800 kWh corresponds to 1,350 kWh per DC kW.
This ratio helps expose a difference for discussion; it is not a universal performance target. Weather, roof orientation, shading and the other model inputs affect the result. Ask the seller to explain those inputs rather than accepting or rejecting the forecast from one ratio.
Ask for inputs you can reproduce
Request the roof-plane layout, DC capacity on each plane, tilt, compass direction, shade assumptions, system losses and the model or weather dataset used. A single average roof direction may conceal panels installed on different slopes. A screenshot of the final annual total is less useful than the input report.
| Input | Question to resolve |
|---|---|
| Roof layout | Which planes actually receive panels and how much capacity is on each? |
| Tilt and direction | Are these measured or assumed? Are directions labeled consistently? |
| Shade and losses | What obstructions were considered and where are the losses recorded? |
| Energy output | Is the figure AC energy, and which losses are included? |
| Year and uncertainty | Is this a typical-year model, a first-year forecast or a guarantee? |
Keep production, bill savings and backup capability separate
Electricity generated is not always electricity used at the retail rate. Some production may be exported under a different compensation rule. A battery also has capacity and power limits; a yearly solar-energy total does not show which appliances it can run during an outage.
A modeled shortfall needs investigation, not an automatic conclusion that the seller is wrong. Ask whether the models use comparable inputs and whether missing information prevents a meaningful comparison.
BidProof’s methodology explains the production model and its limits. A review can leave a check uncompleted when the inputs or reference data do not support it. Production methodology and source limits · Get an independent solar proposal review