How We Test & Review

1. Where our data comes from

• Manufacturer specifications are the primary source for every scored number — claimed weights, gear ranges, tyre clearances, rolling-resistance figures — cross-checked against authorised retailers.
• If a figure isn’t published, we don’t invent it.Unknown specs are stored as unknown and excluded from scoring (the remaining weights renormalise). The data note under each product’s spec table tells you exactly what’s missing.
• Marketing ratings are labelled as such.A manufacturer’s claimed frame weight is a standardised best-case figure — smallest size, often without hardware — not what your complete build will read on a scale. We score it because it’s the only like-for-like number that exists — and we say so wherever it appears.
• Prices are street/MSRP figures as of our last research pass (June 2026) and inevitably drift. Always verify before buying.

2. How rankings are produced

Rankings are not editorial opinion. They are the output of RideScore™, our quantitative scoring model, whose complete formula, weights and changelog are public on the methodology page. Two consequences worth spelling out: a product we love can rank mid-pack if its specs don’t earn it, and we cannot quietly nudge a score — any change to the model is versioned and logged.

3. What we don’t do yet — honestly

We do not currently publish first-party lab measurements: roller-measured rolling resistance, frame and wheel stiffness on a jig, or wind-tunnel drag numbers. Building that test bench is on our roadmap, and when those numbers exist they will enter the model as their own versioned update. Until then, we will not dress up spec analysis as field testing — sites that do are part of why this site exists.

4. Independence & how we’re funded

• Bikefinest currently runs no affiliate links, ads, or sponsorships — nobody pays us to rank or recommend anything (see how we’re funded).
• No brand can pay for a placement, a score, or a position on any list.
• Commission rates differ between retailers and brands; the scoring model has no input for that and never will.

5. Updates & corrections

• Product data is re-verified on a rolling basis; pages carry the date of the last research pass.
• Found an error — a wrong spec, a stale price, a bad assumption? Email editors@bikefinest.com. Verified corrections are applied to the data (which recomputes every affected score, comparison and list automatically) and material errors are noted on the affected page.

6. AI & automation, disclosed

The comparisons and ranked lists on this site are generated programmatically from the scored dataset — that’s the point: every possible matchup, computed identically, with no cherry-picking. Editorial text is human-reviewed before publication. The scoring engine itself contains no AI: it is a deterministic formula you can check by hand against the published weights.