How I test things — Guy Flexor

The short version

Primary sources first. Guidelines, trials, and the paper itself — not a press release paraphrasing a press release.
“Researched” and “tested” are not the same word. If I have not used the product, the piece will not pretend I have.
Physiology is graded before the article is written, not after a draft needs a conclusion.
I am not a laboratory. No fake bench protocols, no cosplay of clinical trials, no invented sample sizes.
Money is disclosed. Commission and free units are stated where they apply. See the disclosure policy.

What the words mean

Researched means the claims were checked against primary sources: the study, the guideline, the regulator page, the manufacturer’s own labelling — not a blog that cites another blog. Evidence types stay labelled: established evidence, early research, expert opinion, what people report, and what a seller claims are five different things. None of them borrows another’s weight.

Tested, on this site, means I personally used the product under ordinary conditions a reader might recognise — at home, in the gym, on a walk — and I am reporting what that was like. It does not mean a blinded trial. It does not mean n=30 with a control group. If the piece says I used it, I used it. If it does not say that, assume I did not.

Compared means the alternatives were weighed on stated criteria (price, seller reliability, what the evidence actually supports), not that every SKU on the market was ordered and race-tested in a garage.

How a claim gets through

Body claims get checked before drafting. Mechanism and outcome stay apart. “Feels looser for an hour” is not the same claim as “fixes your back.”
Where the honest answer is “nobody knows yet,” the article says that instead of manufacturing a neat takeaway.
Seller claims are treated as advertising until a primary source backs them. Packaging copy is not evidence.
Corrections stay on the page with the date and what changed. Quiet edits that erase a mistake do not count.

Limits — said out loud

I am not a doctor, physiotherapist, or coach. Nothing here is medical advice. A personal trial is one middle-aged man’s experience, not a study. Units vary. Sellers change. A product that was fine in March can be a different listing in September.

I do not run a lab. I will not dress up anecdote as assay data. If a piece needed a mass spectrometer and I do not have one, the piece will say what was actually done — or it will not pretend a test happened.

When I have not used something personally, the writing stays in research and criteria. Inventing “I tried it” is the failure this site is least able to survive.

Conflict of interest

Affiliate commission and free samples are material connections. They are disclosed on the piece and summarised in the disclosure policy. The standing question before a recommendation publishes: would Guy still publish this if there were no commission attached? If the answer is no, it does not ship.

Guy Flexor is not the merchant of record. Bad fulfilment by a seller still damages trust here, which is why anonymous no-name appliances with nobody to call tend to die in screening rather than get a glowing write-up and a link.

If I got something wrong

Mail help@guyflexor.com with correction in the subject. Name the article and the sentence. A source helps; the absence of one does not stop you writing in.

Related: Disclosure policy, About, Contact.