Editorial policy
Who writes CleaningHQ, what the words on it are allowed to claim, how disagreements between sources are handled, and how to tell us we are wrong.
Everything here is written, not collected
CleaningHQ stores structured facts, provenance and its own explanations. It does not reproduce source text. Where a trade body or a regulator states something, the statement is restated in CleaningHQ’s words with a pointer to where it came from, and the licence position of each source is recorded so the constraint travels with the record.
Nothing on this site is generated per request. Every page is written and reviewed before it is published, and every record carries the date a person last looked at it. The two exceptions are the advisor and search, whose output is assembled per query — and both are kept out of search indexes for exactly that reason.
What the words are allowed to claim
An automated editorial audit runs on every build and fails it. Two categories are checked.
Overclaiming— “miracle”, “guaranteed”, “always works”, “completely safe”. Marketing language that the evidence model has no way to express.
Safety vocabulary— a harder problem, because the words are technical-sounding and therefore believed. “Non-toxic”, “chemical-free”, “food-safe”, “pet-safe”, “hospital-grade” and “kills 99.9%” are prohibited outright. “Disinfect” and “sterilise” are permitted only where the text also does the work of bounding them, because they describe processes with contact-time requirements rather than properties a product has.
There is one framing the audit checks sentence by sentence: natural does not mean safe. Vinegar etches marble because it is an acid, and being derived from fermentation does not soften the reaction. The UK regulator lists that belief among the myths it exists to correct, and we cite it for that.
When sources disagree
Two credible sources disagreeing is not a problem to resolve before publishing. It is frequently the most useful thing we know, and it gets a record rather than a resolution.
There are 9 recorded. Notably, three of them are internal to a single source— a trade body prohibiting abrasives and recommending an abrasive powder, a stone association calling granite acid-resistant and declining to recommend acids on it. In each case the source was consistent and we were reading it carelessly, which is why the field that does the work is the one naming what each statement is actually about.
Where a disagreement is genuinely unresolved and safety-critical, it blocks the actionable advice. The disagreement still publishes; the clearance does not.
Popularity is not evidence
Widely repeated techniques are assessed rather than repeated, and they land wherever the evidence puts them — including on unresolved, which is an honest report of an investigation that did not reach an answer. A popular hack is never promoted because it is popular, and a traditional one is never dismissed for lacking a randomised trial nobody had reason to run. How hacks are classified.
Independence
No commercial arrangement influences a safety state, an evidence level or a recommendation. Where commercial links eventually appear they will be disclosed, they will be editorially subordinate, and they will not appear on hazard pages at all. Nothing on this site is currently monetised. Commercial disclosure.
Corrections
If something here is wrong, we want to know, and a correction that changes a safety statement is a change to the underlying record rather than an edit to a sentence — so it propagates everywhere that record is used.
Email hello@hekno.co.uk. If you can point at the source that contradicts us, that is the fastest route to a change: this product is built to be argued with using evidence.
Every record shows the date a person last reviewed it. Where a claim has been narrowed, widened or withdrawn after a source was read, the source register records what changed and why.