Where these numbers come from
Every figure these tools rely on, who publishes it, and what happens when it cannot be reached. If something here is out of date, that is a bug worth reporting.
Solar
- Production: NREL PVWatts v8. The National Renewable Energy Laboratory's model, queried for a 1 kW array at your location and scaled to the system size. Defaults are 20 degrees tilt, azimuth 180 (due south), standard silicon modules on a fixed open rack, all overridable.
- Electricity prices: EIA. The Energy Information Administration's retail price data, both for current rates and for the growth assumption of 2.9% a year. The EIA's finding that retail prices tracked inflation over the past decade is the reason that figure is not the 3.5% the industry conventionally uses.
- Upkeep: NREL Annual Technology Baseline. Residential operations and maintenance benchmarked at $30 per kW per year, covering insurance, cleaning, monitoring and the inverter replacement reserve.
- Degradation: 0.5% a year, the standard manufacturer warranty figure rather than a measured one.
- Federal tax credit: zero. IRC Section 25D was terminated by the One Big Beautiful Bill Act (P.L. 119-21), signed 4 July 2025, for expenditures after 31 December 2025. Verified 2026-08-20 and worth re-checking in a later tax year.
- Install price: yours. There is deliberately no fabricated fifty-state price table. If you have a real quote, its price per watt beats any average.
Money
The financial calculators mostly need no external data - amortisation is arithmetic, and the inputs are yours. Where there are fixed figures they are statutory rather than estimated:
- PMI thresholds. Cancellation may be requested at 80% loan-to-value; termination is automatic at 78%, on the original amortisation schedule. These are legal thresholds, not conventions.
- Rates, tax and insurance are inputs, not lookups. A national average mortgage rate would be a worse predictor of your payment than the rate on the offer in front of you, and property tax is set locally enough that a state figure would be false precision.
AI visibility
AnswerRank measures your own site rather than consulting a dataset. It fetches the raw HTML the way a non-rendering crawler would, compares it against the rendered version, and reads your robots.txt for the named AI crawlers. The scoring weights are a judgement rather than a measurement, and they are published so they can be argued with.
When a source cannot be reached
The rule is the same across all three tools: refusing beats guessing. If PVWatts cannot be reached, the tool says so rather than substituting a regional average and presenting the result with the same confidence. A plausible wrong answer is worse than no answer when the decision involves five figures, because you cannot tell it is wrong.
What is an assumption, and what is data
These are not the same thing and are labelled differently throughout. Production from PVWatts is a modelled output from measured weather. Electricity price growth is an assumption informed by history, and history is not a promise. Degradation is a warranty figure. Upkeep is a benchmark. Export value at zero is a deliberate floor, not a prediction.
Where a number is an assumption, it is labelled as one and you can change it. Where it comes from a published source, that source is named with the date it was checked. The point of the distinction is that you should be able to disagree with a specific input rather than with the conclusion.
Corrections
Two of the figures above are corrections to earlier versions of this site. Electricity price growth was 3.5% and was wrong in a way that flattered solar. Upkeep was zero, which is what most calculators still charge. Both changed results materially, and both were changed because the original was indefensible once checked rather than because anybody complained.
If you find another, tell me. I would rather be told than be right, and a correction with a source attached is the most useful message this site can receive.