Good Health Reviews covers the health launches people are talking about, mostly supplements and health programs sold through ClickBank. This page explains how we pick what to cover, what we check, how we weigh the research, and what we don’t do.
The short version: we read the sales page and the label, look up the research on each key ingredient, work out the price per day, read the refund terms, and look for real complaints. Then we write it up in plain English, with our sources linked. We’re not doctors, and we don’t lab-test products.
How we pick products
We follow what’s new and what’s being promoted the most. Our main source is the ClickBank marketplace, where sellers list their products and affiliates like us promote them.
ClickBank shows a “gravity” number for each product. It roughly reflects how many different affiliates have recently earned a commission from it. A high gravity usually means a product is everywhere: in ads, in emails, and in a flood of look-alike “reviews.” That’s when buyers have the most questions, so that’s when we step in.
To be upfront: we also look at how much a product earns for affiliates, and we mostly cover products with an affiliate program that pays us a commission. Being covered here isn’t a recommendation. It means people are talking about the product.
We skip products whose main pitch is to treat or reverse a disease, such as programs that promise to fix diabetes, kidney disease, or high blood pressure.
What we check for every product
- The claims. We note what the sales page promises, using a transcript or written version of the sales video when we can find one. If we couldn’t review the video, the review says so. When we repeat a claim, we attribute it (“the sales page says…”) instead of stating it as fact.
- The label. We list every ingredient and dose that’s disclosed. If doses are hidden inside a “proprietary blend,” we say so, because it means nobody outside the company can compare the amounts to the research.
- The research on each key ingredient. We search PubMed for human studies, read them, and summarize what was found: who was studied, at what dose, for how long, and what changed. Every study we cite links to its PubMed page.
- Label dose vs. studied dose. Is the amount on the label close to the amount used in the studies? Sometimes it’s lower, and sometimes it isn’t disclosed at all. Either way, we point it out.
- Price per day. We work out the cost per daily serving for each package, including shipping when it’s charged, from the official order page on the date shown in the review.
- Guarantee and billing. How long the money-back guarantee lasts, what the terms say about returns, whether there’s auto-ship, and any add-ons or upsells the sales or order page shows. Offers that only appear after payment can’t be seen without buying, so we don’t describe them.
- Complaints and feedback. We look for real buyer feedback on Trustpilot, the BBB, Reddit, and forums, and summarize the patterns we actually find, with links. The seller’s own testimonials don’t count.
- Safety. Who should skip the product or talk to a doctor first, and known interactions with common medications, like blood thinners, diabetes medications, or thyroid medication.
For digital products, like audio programs, e-books, or online courses, we look at what you actually get, who created it, the format, and the refund terms instead of an ingredient label.
How we weigh the research
Not all studies are equal. A single study in mice tells you much less than a review that pools many human trials. Here’s the order we use, from strongest to weakest:
- Meta-analyses and systematic reviews: studies that combine the results of many trials.
- Randomized controlled trials (RCTs) in people: one group gets the ingredient, another gets a placebo, and chance decides who gets which.
- Observational studies: researchers track what people already do and look for patterns. Useful, but they can’t prove cause and effect.
- Animal and lab (in-vitro) studies: early clues only. Results in mice or test tubes often don’t carry over to people.
Sales-page claims, testimonials, and “doctor-formulated” badges aren’t on this list. They tell you what the seller says, not what the research shows.
In our ingredient tables, each ingredient gets one of four evidence labels:
- Strong: consistent results from meta-analyses or several well-run human RCTs.
- Moderate: some positive human RCTs, but they’re small, short, or mixed.
- Limited: a few small human studies, observational data, or research on a different outcome or group of people.
- Insufficient: only animal or lab studies, or no relevant research we could find.
How we write it up
We’re not here to trash products, and we’re not here to hype them either. Each review explains what’s new about the product and why it’s getting attention, what’s inside, what the research says, the real cost, the guarantee, who it may suit, and what to check before buying.
Every review lists real downsides next to the upsides. If a dose is undisclosed, the research is thin, or the refund process has strings attached, we say so plainly.
You won’t see one big overall star rating at the top of our reviews. A single number can hide the details that matter for your situation, so we show you what we found in each area and let you weigh it.
What we don’t do
- We don’t lab-test products. We can’t confirm what’s in a bottle beyond what the label says.
- We don’t claim to have used a product unless the page clearly says so and shows our own photos of it.
- We don’t give medical advice or tell you what to take. That’s a conversation for you and your doctor.
- We don’t publish testimonials, made-up results, customer counts, or star ratings we can’t verify.
- We don’t repeat a seller’s claims as facts.
- We don’t take payment from sellers to cover a product or to change what a page says.
How AI helps
We use AI tools to speed up the research and write first drafts: pulling details from the sales page, finding and summarizing studies, and organizing it all into a readable page.
Our software then checks that every study we cite really exists on PubMed and flags phrases we never use, like common disease-claim wording (“cures,” “treats,” or “reverses” a disease). A pattern check can miss unusual wording. Finally, the page is published automatically only if it passes every check; if it fails one, it stays unpublished. Mauro Costa sets these rules and corrects any error that’s found. More in our Editorial & AI Policy.
Updates and corrections
Each review shows the date we last checked prices. Prices, bundles, and guarantees change often, so we re-check our most-read reviews about every three months, and sooner when a reader or seller tells us something changed.
When we make a meaningful change, like new prices, a changed formula, or a correction, we change the page’s “last updated” date. Typo and broken-link fixes don’t change it, and we never change a date just to make a page look fresh.
If we get something wrong, we fix it, and for anything bigger than a typo we add a note saying what changed. See our Corrections page.
How we make money
We earn a commission when you buy through our paid links, mostly through ClickBank. It doesn’t add anything to the price you pay. Paid links are labeled, and every review has a disclosure near the top, before the first paid link.
The details are on our Affiliate Disclosure page.
Last updated: November 2026.