How we test.
And what we don’t claim.
Updated 34 criteriaBy Screenless Bands
Most tracker reviews imply more testing than they performed. We would rather publish a narrower claim we can defend, so this page states exactly where our evidence comes from, how the score is weighted, and which parts of a review are not our own measurements.
Our evidence standard
Screenless Bands is an evidence-synthesis publication. Each review is built from three clearly separated tiers, and every review page publishes the full source list so you can check the tier yourself.
- Official documentation — manufacturer specifications, manuals and support articles. Used for specifications, subscription terms and feature availability. Never presented as measured performance.
- Repeatable professional testing — independent reviewers who publish their protocol, reference device and conditions. Used for accuracy, battery and real-world behaviour, always attributed by name.
- Owner reports — forum and community discussion. Used only to identify what deserves testing. Support threads overrepresent failures, so we never convert them into a defect rate.
What we do not do. We do not present third-party observations as our own laboratory results, and we do not have controlled validation for every claim on this site. Where testing conflicts, our reviews say so instead of picking the flattering result.
How the score is built
Every band is scored against 34 criteria in eight groups. The band score and the app score are calculated separately, because a good sensor attached to a bad app is a different product from the reverse — and in this category the app is usually the weaker half.
| Group | Criteria | Weight | What it covers |
|---|---|---|---|
| Comfort and fit | 4 | 12% | Wrist and upper-arm fit, clasp security, skin irritation over 24/7 wear, and whether the sensor keeps contact on smaller wrists. |
| Battery and charging | 4 | 12% | Claimed versus reported runtime, charge time, whether the device must be removed to charge, and what a power-saving mode actually disables. |
| Health signals | 5 | 15% | Continuous heart rate, HRV, SpO₂ where offered, skin temperature, respiration and how each metric is explained rather than merely displayed. |
| Workouts and accuracy | 5 | 18% | Steady-state and interval heart rate against references, automatic detection reliability, manual start workflow and connected-GPS behaviour. |
| Companion app | 6 | 18% | Sync reliability, dashboard clarity, workout editing, coaching quality, onboarding for new users and how much the app explains its own conclusions. |
| Integrations and data ownership | 4 | 11% | Apple Health and Health Connect behaviour, third-party services, export paths and field-level completeness rather than a logo list. |
| Privacy and subscription terms | 3 | 8% | What a paid tier gates, what happens to access on cancellation, and whether warranty depends on continued payment. |
| Cost of ownership | 3 | 6% | Upfront price, required recurring fees over one, three and five years, and essential accessories. |
Two rules constrain the result. A band cannot score above 9.0 if a core advertised feature is unreliable in independent testing, and a required subscription is scored as a permanent cost rather than a one-off — which is why a technically excellent subscription band does not automatically outrank a cheaper one you own outright.
The hands-on protocol
Where we do test directly, this is the protocol, and it is the same one we publish in our wrist versus bicep guide so readers can repeat it.
Wear it continuously
Comfort, skin irritation, clasp security, charging interruptions and real battery drain are recorded across sleep, work and exercise — not in a single test day.
Compare against references
Workout heart rate against a chest strap, distance against a GPS reference, steps against controlled manual counts. Every protocol is repeated, and both wrist and upper-arm placement are recorded where the band supports it.
Audit the app as a product
Sync reliability, dashboards, recommendations, workout editing, integrations, exports, privacy controls and precisely which data disappears without payment.
Record the conditions
Firmware version, phone platform and sensor position accompany every result. Early reports on new bands are unusually mixed, and a result without those three details cannot be compared with anyone else’s.
Money, and where it cannot reach
This site earns through the Amazon Associates programme. As an Amazon Associate, Screenless Bands earns from qualifying purchases. That relationship is disclosed on every page that carries a commercial link, and it is constrained in three specific ways:
- No paid placement. No manufacturer can buy a position, a badge or a score, and none has been offered one.
- Rankings are set before links are added. Scores come from the criteria above; affiliate links are attached afterwards to whatever the ranking produced.
- Commission rate is never an input. A band that pays more is not promoted over one that pays less or nothing at all.
Amazon links point to product search rather than a fixed listing, so availability can be checked by region. Affiliate tracking activates only once an Associates tag is configured.
Corrections and updates
Firmware and subscription terms change faster than reviews do. Three things follow from that:
- Every review and guide carries a visible last-updated date.
- Where a finding depends on a firmware version or phone platform, we say so rather than presenting it as permanent.
- Factual errors are corrected in place, and material corrections are noted rather than quietly overwritten.
If something here is wrong, the source list on each review is the fastest way to show it — point at the tier we got wrong.