Why free calculators flatter you
A calculator that costs nothing still has to be paid for. Once you know how it is paid for, its answers stop being surprising.
The arrangement
The common business model for a free financial tool is lead generation. The calculator is not the product. You are. You arrive wanting to know whether solar is worth it, or what you can borrow, and somewhere between the result and the fine print you hand over an email address or a phone number. That contact detail is then sold to somebody who wants to sell you a system, a policy or a loan, typically for somewhere between a few dollars and a few hundred depending on how close to buying you appear.
Nothing about that is illegal, and much of it is disclosed if you go looking. But it has a consequence that almost nobody states out loud: the tool is paid more when the answer is encouraging. A result that says "this pays for itself in seven years" produces a lead. A result that says "this will never pay for itself where you live" produces a closed tab. The incentive does not have to be acted on cynically to work. It is enough that the version of the calculator with the cheerier assumptions performs better, and so that is the version that survives.
This is selection, not conspiracy. Somebody tries a more conservative electricity-price assumption, conversions fall, and the change is reverted because it "underperformed". Nobody in that meeting believes they are being dishonest. Over a few years of that, an entire category of software drifts in one direction.
Four places the flattery hides
The tilt is rarely a false number. It is usually an omitted one, or an optimistic choice among several defensible ones. Four patterns cover most of it.
1. A cost that is quietly left out
The clearest example is a mortgage payment quoted as principal and interest. That figure is real, it is just not what leaves your account: property tax, insurance and often private mortgage insurance are collected alongside it. The quoted payment can be a third smaller than the actual one, which is the difference between comfortably affordable and not. There is more on this in what your mortgage payment leaves out.
Solar has the same pattern. Panels need insurance, cleaning and monitoring, and the inverter usually needs replacing once during the panels' life. Most solar calculators charge nothing at all for upkeep, which overstates the lifetime result by somewhere between 7% and 40% depending on where you are.
2. An assumption presented as a fact
Long projections are extremely sensitive to their growth rates, and a growth rate is a guess. Solar tools have historically assumed electricity prices rise around 3.5% a year, usually justified with the claim that utility rates rise faster than general inflation.
The EIA's own data does not support that. US residential prices went from a little over 12 cents per kWh in 2013 to 16 cents in 2023, which is 2.9% a year, and in inflation-adjusted terms they rose less than 1% across that entire decade. Electricity has tracked inflation rather than outrunning it. The gap between 3.5% and 2.9% sounds like rounding. Over twenty-five years, in a city with cheap power, it is the difference between an installation that pays back and one that never does.
3. A rule that has changed
Tax law moves faster than software. The Residential Clean Energy Credit under IRC Section 25D, the familiar 30% off a home solar installation, was terminated by the One Big Beautiful Bill Act (P.L. 119-21) for expenditures made after 31 December 2025. A system bought and installed today receives no federal credit.
A calculator still applying 30% understates the cost of a $28,000 system by roughly $8,400 and shortens the quoted payback by about three years. It is not lying. It was correct in 2024. Nobody updated it, and there is no commercial pressure to, because the outdated version converts better.
4. A precise number invented from nothing
The most subtle one. A tool asks for your state and returns a cost per watt to two decimal places, giving the strong impression that somebody measured it. Often no such dataset exists at that granularity, and the figure is an interpolation, a national average wearing a local label, or simply made up. Precision is being used as a proxy for accuracy, and the two are unrelated. A number given as a range, or refused outright, is usually the more honest one.
How to check a tool in about a minute
Four questions, in rough order of how much they reveal:
- Does it ask for your contact details before showing the result? If so, you already know what it is for. A tool that gates the answer is selling access to you, and the answer will be shaped accordingly.
- Can you find the assumptions? Not a vague note that results may vary, but the actual figures: the growth rate, the degradation rate, the tax treatment, and the date each was last checked. If they are not published, they cannot be argued with, which is generally the point.
- Can it return bad news? Try inputs that should produce a discouraging answer. A cheap-electricity location for solar. A loan you plainly cannot afford. If every combination comes back positive, the model has a floor under it that reality does not.
- Does it say what it cannot see? Every model has a boundary. A tool that names its blind spots, the tree shading your roof, your actual tax position, whether your job is secure, is telling you where to apply your own judgement. A tool that names none is implying it has none.
What this site does instead
These tools collect nothing. No account, no email address, no phone number, and nothing you type is stored or sold. That is not a virtue so much as a structural fact: with nothing to sell, there is no reason to flatter you, and no meeting in which a more conservative assumption gets reverted for underperforming.
The practical result is that the answers here are often worse than the ones you will get elsewhere. Solar payback figures are longer, sometimes materially, and for some locations the honest answer is that it never pays back at all. Mortgage payments are larger, because they include the parts that are usually omitted. That is not conservatism for its own sake. It is what the arithmetic produces once the missing costs are put back in and the assumptions are set to what the data actually shows.
Every figure is published on a methodology page with the source and the date it was checked, so you can disagree with a specific number rather than with the conclusion. If one of them is wrong, I would genuinely rather be told than be right, and you can get in touch directly. The full source list is in where these numbers come from.