← Do Your Own Math

What we changed — and why

Every recommendation we have withdrawn, corrected or qualified, with the reason and the source. Including the ones that cost us money.

Any site can promise to be honest. The only way to show it is to publish the times you were wrong, on a page you keep rather than a correction you bury.

So this is that page. Each entry says what we used to claim, what made us look again, what we found at the source, and what it did to our own income. When a change cost us money we say how much, because a correction that costs nothing proves nothing.

3 August 2026Recommendation withdrawn

We recommended a whole-house filter for lead. It is certified by nobody.

For documented lead in drinking water, our water calculator led with a $400 whole-house system. We had described it the way the retail listing does — as a lead-reducing system.

Checking the certifiers directly, no listing exists for it with NSF, IAPMO or the WQA. iSpring does hold genuine NSF certification, but on its reverse-osmosis units, which are different products. Home Depot's own specification table for this model reads "No Certifications or Listings". And the manufacturer's "tested by an independent third party" line covers stages one and two — not stage three, the lead filter itself, which is the only reason to buy that model.

We also found the reverse-osmosis unit we offered as the alternative holds a certification covering exactly one thing: total dissolved solids. Its lead and PFAS figures are laboratory tests, not certified claims.

The calculator now routes documented lead and PFAS to a $35 filter pitcher, which turned out to be the only product on the page with a listed certification naming those substances — lead 98.8%, PFOA and PFOS 99.8%. For arsenic and nitrate it now says none of our products is certified for either and sends people to the NSF database instead.

What it cost us: we moved buyers from a $400 product to a $35 one, and told a third group to buy nothing from us at all. Partly offset by the pitcher needing twelve filters a year.
Sources: NSF certified-products database · Home Depot specification table · the Waterdrop performance data sheet. The full audit.
3 August 2026Claim corrected

We called an air purifier "independently AHAM-verified". It is not in the register.

Our default air purifier pick carried the phrase "independently AHAM-verified". We could not find that model, or either of the other two we recommend, in the AHAM Verifide directory. Of the three, the one carrying the claim had the least independent backing.

The phrase is gone. In its place, each page now states where every number comes from: the largest unit's 391 CFM is confirmed as its smoke CADR in the ENERGY STAR registry, and the two smaller figures rest on the manufacturer's own documentation, which does not even say which of the three CADR types it is publishing.

The numbers themselves stood up — nothing was contradicted. And the check produced something we did not expect: our arithmetic turns out to be within half a percent of AHAM's own room-size formula, and the threshold above which one unit is no longer enough now rests on a government registry rather than a manufacturer's spec sheet.

What it cost us: nothing directly. The recommendations stand; only the evidence behind them is now labelled correctly.
Sources: AHAM Verifide directory and its air-change white paper · ENERGY STAR certified room air cleaners dataset · Levoit product documentation.
3 August 2026Monetisation removed

Our free widget carried shopping links it should never have had.

We built an embeddable version of the air calculator for other sites to put on their own pages. It contained three affiliate links. Amazon's terms only permit their links on sites an associate has registered — so every site that embedded our widget would have been publishing a breach on its own page, with our tag on it.

All three links are gone. The button under the result now opens our own calculator. The widget carries no shopping links, no cookies and no analytics of any kind, and the terms page explains why rather than hiding it.

What it cost us: nothing yet — nobody had embedded it. What it removed was the risk of losing the affiliate account entirely.
3 August 2026Gap disclosed

Our best-selling bidet pick documents no backflow protection.

Plumbing codes treat bidets as a specific case: the water supply must be protected against backflow, by an air gap or an appropriate backflow preventer. It is the one genuine safety question in the category.

Our main pick — a $56 attachment — documents no backflow protection at all. No vacuum breaker, no check valve, nothing in the manual or on the product page. Its maker does not appear in IAPMO's listee index, and what looks like a certification in the manual is in fact a partial list of test sections. Our budget option is the same, citing a 2009 edition of the standard that has since been superseded twice.

This is not evidence that either is unsafe, and no recall exists for them. It means the protection has to come from the plumbing, and nobody selling them mentions it. That now appears above the buy button rather than in a footnote. On the business page, where a facilities manager may be fitting twenty of them in an inspected building, it appears before the payback arithmetic instead of after.

The audit also moved something the other way. The $410 electric seat we recommend for care at home turned out to be the best-documented product anywhere on this site: listed by model name in IAPMO's electrical directory, with both a vacuum breaker and a check valve in its manual, and a spray temperature capped below the code limit. Its own warnings about scalding and about never using the seat as a support are now quoted directly, because they are better than ours were.

What it cost us: a warning above the buy button on our highest-earning page, and a warning on the business page telling commercial buyers they may need a more expensive product than the one we costed.
Sources: IAPMO EGS file E10023 · IAPMO R&T listee index · manufacturer manuals · 2024 International Plumbing Code §408. The full audit.
3 August 2026Published with nothing to sell

We looked at smart toilets that claim to spot illness. None holds an FDA clearance.

We checked five toilet-based health monitors against the manufacturers' own documents and the FDA's databases. One manufacturer's genuine FDA clearances turned out to be for a heart monitor, not the urine device people assume they cover. Another company's urine test strips were recalled by the FDA in 2026, with the cause recorded as "No Marketing Application".

The resulting page carries no affiliate link and no buy button, and says that if you want to know whether you have a disease, none of these is the answer. We are not in any of these affiliate programmes and did not apply.

What it cost us: nothing, and that is the point — there was never anything to earn here. The page.