Anyone who has searched for a supplement review lately knows the feeling. The first page of results is a wall of near-identical “top 10” articles, each with a scoring widget and a button that leads to a checkout page. Further down are forum threads, short-form videos, AI-generated summaries that cite each other, and lab report screenshots that may or may not belong to the product in question. Information is not scarce. The problem is that very little of it was produced by someone whose incentives line up with yours.
If you spend your days evaluating software, hardware, or data pipelines, you already have the skills to cut through this. Supplement information is just another untrusted input. The job is to work out who produced it, what they stand to gain, whether the underlying data exists, and whether you can verify any of it. This guide is a practical workflow for doing exactly that.
Start with the business model behind the page
Before reading a single claim, work out how the page makes money. Most supplement “review” sites are affiliate publishers, earning a commission when a reader clicks through and buys. That does not make them dishonest, but it shapes what they publish. Products with generous commission programmes get reviewed; products that do not pay get ignored. Negative verdicts are rare because a negative verdict earns nothing.
There are a few quick tells. Look for an affiliate disclosure near the top or bottom of the article. Check whether every product in a comparison links to a retailer with a tracking parameter in the URL. Notice whether the “best overall” pick changes month to month in ways that track promotions rather than new evidence. Phrases like “editor’s choice” are marketing labels, not findings.
Contrast this with sites built around independent testing or research review. An Australian example is suppscience.com.au, which commissions its own lab tests on supplement products, compares what the label says against what was measured, reviews the published research, and gathers user experiences, rather than steering readers toward a purchase. The distinction matters because it tells you what kind of error to expect. An affiliate page is likely to overstate benefits and bury drawbacks. A testing site is likely to be narrow in scope and cautious in its conclusions. Knowing which bias you are dealing with lets you correct for it.
Learn to read a certificate of analysis properly
A certificate of analysis, or CoA, is a document from a laboratory stating what was found in a specific sample. Brands increasingly publish them, and reviewers screenshot them as proof of quality. Read carefully, a CoA is one of the most useful documents you can find. Read carelessly, it is just a reassuring-looking PDF. Here is what to check:
- Who issued it. A CoA from a third-party lab carries more weight than one generated by the manufacturer’s in-house quality team, because the in-house team answers to the people selling the product. The lab’s name should be searchable and the document should carry its contact details.
- Which batch it covers. A CoA applies to one lot or batch. If the product you are considering has a different lot number, the document tells you about a different product. Many brands display a single old certificate for years; that is a red flag, not a reassurance.
- What was actually tested. A certificate that only reports identity (“the sample contains the stated ingredient”) says nothing about how much is present, or whether contaminants were checked. Look for assay results showing measured content versus labelled content, and for separate sections covering heavy metals, microbial contamination, and residual solvents where relevant.
- The method and the units. Reputable reports name the analytical technique used and report results in clear units with detection limits. Vague entries like “passes” or “conforms” without a number are weaker than a figure you can compare against the label.
- Whether it has been altered. Mismatched fonts, missing page numbers, inconsistent dates, or a pasted-on logo are all reasons to contact the lab and ask whether the document is genuine.
Finally, remember what a CoA cannot tell you. It describes one sample at one point in time. It does not establish that the product is safe for you, or that it does what the marketing says. Independent testing sites tend to say this explicitly, and their disclaimers are worth reading for that reason: results relate to the tested samples only and do not establish safety or effectiveness.
Click the study links, then read past the abstract
Supplement articles love citations. A paragraph claims an ingredient “has been shown to” improve something, followed by a superscript number or a hyperlink. Most readers stop there, satisfied that evidence exists, which is precisely what the citation is designed to produce.
Follow the link. Many resolve to another blog post, a press release, or the brand’s own site. When one does reach a journal, check three things. First, was the study in humans, or in cells or animals? Findings in a dish or in rodents are the beginning of a research question, not an answer. Second, how large and how long was the trial? Many supplement ingredients have been tested only in small short-term trials, and that is a legitimate finding to report, but it should be reported as such, not dressed up as settled science. Third, does the study actually test the claim the article is making? It is common to see a trial on one population or one outcome cited in support of a quite different claim.
Watch, too, for who funded the research. Industry-funded studies are not automatically wrong, but they deserve the scrutiny you would apply to a vendor-published benchmark. And be alert to citation laundering: article A cites article B, which cites article C, which turns out to be a single uncontrolled observation. Tracing the chain back is tedious, but it is often where the claim falls apart.
Use independent testing sites as a cross-check, not an oracle
A handful of organisations buy supplements off the shelf, send them to laboratories, and publish what comes back. Their value is that they are measuring rather than reviewing. If a product’s measured content falls well short of its label, that is a fact no marketing copy can argue away.
Understand their scope, though. They test a limited number of products and suppliers, so absence from their results means nothing either way. They report on specific samples, and quality can drift between batches. And they measure what a lab can measure: content, purity, contaminants. They cannot tell you whether the ingredient will work for you.
Within those limits, there are several ways to get more out of them. Look for sites that explain their methodology and publish results by product and supplier rather than as a single pass-or-fail badge. Check whether they accept requests for testing, which is a useful signal that the testing is driven by reader questions rather than by which brands are willing to pay for coverage. Supp Science, for example, runs a “Request Testing” form alongside its published results, which gives readers a route to get a product they are curious about onto the list. And read their research summaries alongside their test results, because the two answer different questions: one tells you what is in the bottle, the other tells you whether the ingredient has any evidence behind it in the first place.
Treat user experiences as data with known biases
Forum threads, subreddit posts, and comment sections are the supplement world’s bug tracker. They are noisy and occasionally gamed, but they contain signals you will not find anywhere else: delayed side effects, batch-to-batch inconsistency, or a brand that quietly reformulated.
Read them the way you would read error reports from a large user base. Individual accounts prove very little; one person feeling better could be regression to the mean, placebo, or coincidence. Patterns across many independent accounts are more interesting, especially when they describe something specific and unflattering that nobody has an incentive to invent. Be sceptical of clusters of glowing reviews posted within a short window, reviews that repeat marketing phrases verbatim, and accounts with no other history. Trust long-running threads where people disagree, report dropping a product, and update their views over time.
Some testing and information sites now collect user experiences through structured review programmes rather than free-text comments. Supp Science does this alongside its lab work and research reviews, which lets readers see measured content, published evidence, and reported experience for the same product side by side. That combination is more informative than any one of the three on its own, because each corrects for the blind spots of the others.
Be especially careful around research compounds and prescription medicines
The hardest part of the online supplement landscape in 2026 is that the word “supplement” has stretched to cover things that are not supplements at all. Peptides such as BPC-157 or GHK-Cu, and compounds like retatrutide or semaglutide, show up in the same feeds and the same forums as magnesium and creatine. Testosterone enanthate is a prescription medicine, used under medical supervision for diagnosed deficiency and regulated in Australia by the TGA, yet it is discussed in some corners of the internet as if it were a gym product.
For this category, the vetting bar should be much higher. Treat peptides and similar compounds as research topics with limited human evidence rather than products to try. Be suspicious of any site that pairs enthusiastic how-to content with a shop link; that is the clearest possible conflict of interest. Prefer sources that discuss these compounds as emerging research, are explicit about what is not known, and point readers toward medical advice. If a page offers to help you obtain something that requires a prescription where you live, close it.
A practical checklist you can reuse
None of this requires special expertise, just the habit of asking the same few questions every time. Before acting on any piece of supplement information online, run through the following:
- Who published this, and how do they make money from me reading it?
- Is there a lab report, and does it cover the batch I would actually be buying, from a lab I can verify?
- Do the study links go to real human research that tests the claim being made, and how big and how long were those studies?
- Has an independent testing organisation measured this product, and what did they find?
- What do long-running, unmoderated user discussions say, and do the patterns hold up across many independent accounts?
- Is this actually a supplement, or a research compound or prescription medicine being marketed as one?
If a product survives all six questions, you still do not know that it will work for you. But you will know the claims around it are grounded in something you can check, which is more than most of what you read online can offer. In a space where the loudest voices usually have something to sell, that is the whole advantage.
