How We Evaluate
"Trust us" isn't a method. This page is ours — written out in full, so you can hold us to it. It covers how we assess a health claim, how we assess a product, and the specific things we will not do to earn a commission.
Two kinds of content, two methods
We publish two broad types of content, and they're held to different standards because they answer different questions:
- Health & nutrition guides — "how does insulin resistance work," "what actually helps perimenopause." These live or die on the quality of the evidence behind every claim.
- Health-tech buyer's guides — "which smart ring," "Fitbit vs Garmin." These live or die on whether our evaluation reflects how the device actually performs, not how its marketing reads.
The thread running through both: a claim is only as good as the source that's independent of whoever profits from it.
How we weigh evidence
Not all "studies show" are equal. When sources disagree, we weight them in this order — strongest first — and we tell you where on this ladder a claim sits rather than rounding everything up to "proven":
| Tier | Source type | How we treat it |
|---|---|---|
| 1 | Systematic reviews & meta-analyses | Strongest. Multiple trials pooled. Our default anchor for any health claim. |
| 2 | Randomised controlled trials (RCTs) | Strong, especially larger and pre-registered ones. We note sample size and who funded them. |
| 3 | Cohort & observational studies | Useful for associations, but we say "linked to," never "causes." |
| 4 | Mechanistic / animal / in-vitro | Explains why something might work. We never present it as proof that it does in humans. |
| 5 | Expert consensus & guidelines | Context and safety framing. Cited, but not a substitute for primary data. |
| 6 | Manufacturer / seller claims | Can confirm a fact about a product (dose, dimension, sensor). Can never vouch for a benefit. |
Before any effect claim goes live, it has to pass one test: "Who profits if the reader believes this?" If the answer is "the seller," the claim isn't sourced yet — we either find a tier 1–4 independent source or we soften the statement to exactly what the evidence supports.
How we evaluate a health claim
- Start from primary research, not other blogs. We build each guide from peer-reviewed studies (PubMed, NIH/NIDDK, major journals) and link them so you can read the evidence yourself.
- Verify the citation actually says what we claim. Every study is checked to confirm it exists, the numbers match, and the link resolves. A citation that doesn't support the sentence it's attached to is a defect, not a decoration.
- Mechanism before recommendation. We explain the root cause and the biological mechanism first, so a recommendation feels earned rather than asserted.
- Honest about uncertainty. When the evidence is early, mixed, or animal-only, we say so. "We don't know yet" is a legitimate answer and we use it.
- No price claims. We never publish a static product price — they go stale and mislead. We compare on dose, form, and specs instead.
How we evaluate a product or device
This is where most review sites quietly overstate themselves, so we'll be precise about what we are.
WiseGoodness is research-led, not a hands-on testing lab. We do not currently claim to have personally worn, stress-tested, or lab-measured every device we write about — and we will never pretend we did. What we offer instead is rigorous synthesis of the evidence that already exists, assembled more carefully and more transparently than most "we tested it" pages actually manage.
What that means we actually do:
- Aggregate independent accuracy studies. Wearable heart-rate, HRV, and sleep-tracking accuracy have been measured in peer-reviewed validation research (e.g. independent academic comparisons against ECG and polysomnography). We lead with those numbers, not the manufacturer's.
- Separate facts from benefits. A spec sheet can tell us a sensor's sampling rate or a battery's rated life (fact). It cannot tell us the device "improves recovery" (benefit) — that needs an independent source.
- Read the owner data at scale. Patterns across large volumes of verified owner reviews (what breaks, what the app gets wrong, what people quietly return) are evidence the spec sheet hides.
- Cross-check across independent reviewers who do run hands-on labs, and cite them — rather than laundering their work as our own.
- Score against fixed criteria (below), the same way for every device, so a comparison is consistent rather than vibes.
What we will not do:
- Claim first-hand testing we didn't perform.
- Accept payment to feature, rank, or score a product more favourably.
- Let an affiliate commission change a verdict — the link follows the recommendation, never the other way round.
- Publish a price that will be wrong next week.
Our device evaluation criteria
Every wearable and health device is assessed on the same fixed dimensions, weighted toward the things that actually change a buying decision:
| Criterion | What we look at |
|---|---|
| Accuracy | How the sensors perform in independent validation studies vs gold-standard references (ECG, polysomnography) — not the brand's own figures. |
| What it measures | The metrics that matter for the stated use (HRV, sleep stages, SpO₂, GPS) and whether they're genuinely actionable. |
| App & data | Whether the app turns data into a decision, and what happens to that data — exportability, subscription lock-in, privacy. |
| Battery & wearability | Real-world battery life and whether the device is comfortable enough to actually wear 24/7 (the only way it collects useful data). |
| Total cost of ownership | Hardware plus any subscription, over a realistic ownership window — qualitatively (budget / mid / premium), never a stale dollar figure. |
| Who it's wrong for | Every recommendation names the person who should not buy it. A review with no "not ideal for" is marketing. |
What would change our recommendation
A recommendation isn't a monument. We revisit it when a newer validation study contradicts an accuracy claim, when a firmware or subscription change alters the value calculation, when owner data reveals a reliability problem, or when a better option ships. When a verdict changes, we change the page and update the review date — we don't quietly leave the old call standing.
Independence, affiliates & corrections
How we fund this without it bending our recommendations, how we fact-check, our use of AI, and how to flag an error are all set out in our Editorial Policy. The short version: affiliate commissions keep the lights on, they never buy a verdict, and if we get something wrong we'd rather you tell us than let it stand.
Who stands behind this
Guides are written by Michael Thomassen and reviewed for accuracy by Nicolas Aubineau. Real names, accountable bylines. More about the team and what we cover is on our About page. This page is educational and is not medical advice — always speak with a qualified professional before changing your diet, supplements, medication, or health routine.