About Enzyme Cleaning Products
Enzyme Cleaning Products is an evidence-led directory of household and trade cleaning products.
Ask which enzyme cleaner is best and you get three answers. A forum thread from 2019, a retail listing with a hundred thousand reviews of unclear provenance, and a magazine roundup of five products refreshed once a year. None of them says which brands hold up on the hard messes, or which claims on the bottle anyone has checked.
We currently publish a page on each of 42 products across 12 brands, with additional catalog rows on brand pages. Facts appear only where the publication record carries supporting evidence.
Who Publishes This Site
The website was designed and is hosted by The Discoverability Company, and it is maintained by contractors.
Pages on this site carry the byline ECP Staff, the editorial team behind Enzyme Cleaning Products. We do not attach an individual byline or credential unless it can be verified.
What we publish instead is the method: where every number comes from, how large its sample was, and the date it was read. Manufacturer statements are labeled as manufacturer claims.
Plural voice is deliberate. The site reads as one specialist because the standard is one voice, and behind that voice is a working team doing the reading, the counting and the re-checking. What would not survive an individual byline is the correction record. A named author's error reads as one person's mistake, while the site's error reads as the site's. Ours are published the same way anyone else's would be.
The method is the credential on offer. Anyone can publish a list of five bottles. The difference between that and this directory is that every figure here can be walked back to a receipt, a sample size and a date. Where a claim is the manufacturer's, it is labeled as one. Where a number is ours, the counting method sits next to it.
Independence has a mechanical meaning here. Placement is not for sale, and no manufacturer sees a page before it publishes. Numbers are read from their sources, dated, and reproduced exactly rather than tuned. Retailer links carry no payment of any kind, so there is no commercial relationship whose weight a page could feel. That is the whole independence claim, and it is checkable line by line.
How We Compare Enzyme Cleaners
We do not run product tests. What we do is read documents: public manufacturer documents, product labels, retailer listings and other documented public sources. We compare the facts those sources support, record where they are silent, and correct the record in public when we get one wrong.
Dates and sample sizes appear only when their source receipt supports them.
In practice that means three things. We read label facts from the manufacturer rather than the reseller. We verify every price against the page that publishes it. We also follow every retailer link ourselves, so the one you click lands on the product you were reading about.
On the questions where nobody has published a number, we say what we searched. A PubMed search for "enzyme cleaner pet urine odor carpet" returns zero records, and so do "enzymatic cleaner odor removal comparison oxidizer" and "urease uric acid carpet odor deodorization". The only relevant peer-reviewed test of cat urine odor products on carpet was published in 1989, and its full text sits behind a paywall we did not read.
Comparison happens inside a category, never across one. A gallon concentrate and a trigger spray are different purchases at the same price. A pet spray is not comparable to a laundry powder even when both carry the same enzyme. Grouping by use is what makes a price or a rating mean something, and it is why the directory is organized the way it is.
Where the record is silent, the silence is published. A figure nobody has printed does not become an estimate, and a gap does not become a guess: it becomes a sentence saying what we searched and what came back. That is slower than rounding off, and it is the only version of this work that survives being checked.
Where the Price and Rating Data Comes From
Three properties make that table a promise rather than a caption. Every figure is tied to the page that published it, so a reader can stand where we stood. Every rating is shown with its count, because a rating without its pool is decoration. And every date is the date the figure was read, not the date the source says it was true. Prices drift and listings change, so a dated figure can be re-checked or retired rather than defended.
Provenance travels with every number. A price names the retailer that published it, a rating names the retailer whose buyers produced it, and a label fact names the manufacturer document it was read from. A figure with no origin does not appear, which is a stronger promise than it sounds in a category built on round numbers. Round numbers without origins do not survive an audit.
| Prices | Retailer listings, checked and dated on the page. |
|---|---|
| Label facts | Read from manufacturer copy. Where a label says fragrance free, that is a fact about the label and we record it as one. |
| Enzyme evidence | Three levels, printed on every product page: stated on the retail listing, confirmed in the manufacturer's own copy, or read from the product line. The third is inference and is labeled as inference. |
Editorial Rules
- Every number carries its source and its sample size. A rating never appears without the count behind it.
- Every negative names the search that produced it. The database, the query and the result count.
- Prices carry the date they were checked.
- Claims are attributed to whoever made them. A manufacturer calling a product pet-safe is recorded as a manufacturer claim, and "pet-safe" has no legal standard.
The four rules share one shape: each one costs us something. Naming a sample size limits how impressive a number can sound. Dating a price admits it will drift. Attributing a claim gives the manufacturer the credit and the blame for it. Printing the search behind a negative shows the reader exactly how little is underneath some of the strongest sentences on the site.
The rules are also what make the corrections policy possible. A site that dates what it reads can correct itself cleanly, because the correction names the same receipt the original claim named. That is the standing offer: a wrong figure, pointed out, gets checked against its own source and either fixed or defended in public, on the page. A dated receipt is the whole mechanism. Material corrections are published on the corrections page, newest first, each naming the affected page, the error, the corrected statement and its source.
Everything else on the site inherits these rules without exception. A number that cannot carry its receipt does not publish, and a page that cannot cite its figure does not ship.