How the Nexus Score works

Our principles

• Independent. Advertising and affiliate relationships are not inputs. The scoring code has no field for them.
• Transparent. Weights, formulas and category benchmarks are published on this page and on every scorecard.
• Reproducible. Scores are computed automatically from the recorded data, so the same data always produces the same score.
• Safety first. No amount of good reviews can outweigh an open recall or an uncertified battery.
• Fair comparisons. Products are only scored against their own category. A folding e-bike is never penalised for not being an e-MTB.
• Honest about uncertainty. Every score shows a confidence level and margin. Thin data is labeled Provisional.

The five pillars

Each pillar produces a score from 0 to 100. The Nexus Score is their weighted average.

1. Expert consensus (30%)

We collect scores from independent reviews and convert each one to a 0 to 100 scale. Each review is weighted by how deeply it tested the product and how recent it is:

• Recency: a review's weight halves every 24 months, so current models are judged on current evidence.
• Measured bonus: reviews that measured performance themselves get a further 15% weight.
• Calibration: outlets that score everything high (or low) are re-based to a common scale once we have 20+ of their scores.
• Our own testing counts as one source and can never exceed 30% of this pillar when other sources exist.

2. Owner satisfaction (25%)

Owner ratings come from retailers, marketplaces and owner communities. Star ratings are mapped so that 1★ = 0 and 5★ = 100.

• Verified purchases count fully, and unverified reviews count at 0.6×.
• No single platform can exceed 40% of the pillar, which limits the effect of review manipulation on any one site.
• Small samples are shrunk toward the category average using a Bayesian average, so ten perfect reviews can't beat a thousand very good ones:

Owner score = (n × R + 150 × C) ÷ (n + 150)

where n is the weighted number of ratings, R the weighted average, and C the category's prior (typical owner satisfaction for that category).

3. Performance & specs (20%)

Each category has its own benchmark specs, such as battery Wh, weight, torque, tested range, brakes and water resistance. Each spec scores 0 at a fixed floor and 100 at a fixed ceiling. We use fixed benchmarks rather than ranking against whatever else happens to be in our database, so one extreme product can never distort everyone else's score.

We only score this pillar when verified specs cover at least 50% of the benchmark weight. We prefer measured values (for example, tested range) over claims. Every category's benchmarks are published on its leaderboard page.

4. Reliability & safety (15%)

The product's score makes up 75% of this pillar. The other 25% is the brand trust score, which combines customer-support satisfaction (35%), years in business (25%), recall history (20%) and warranty and service coverage (20%).

5. Value for money (10%)

Value asks: how much more (or less) quality do you get than the typical product at this price? First we take a product's quality score, the weighted average of the other four pillars. Then we compare it with the quality the typical product delivers at that price, modeled as a curve from 60 at the category's entry price to 85 at its premium price.

Value = 50 + 2.5 × (quality − expected quality at this price)

A product that delivers exactly what its price suggests scores 50. One that delivers 10 points more scores 75. Expensive products can still score well on value if they truly earn their price.

Final score & safety caps

The Nexus Score is the weighted average of the available pillars. If a pillar has no data yet, we never guess: its weight is shared proportionally across the others, and confidence drops. Two safety rules then apply:

• Open safety recall: score capped at 59 until the recall is resolved.
• Battery-powered product without a recognized battery/electrical certification: capped at 79. It can't be rated "Excellent".

Confidence & margin of error

Confidence reflects how much evidence sits behind a score: the number of expert sources, the size of the owner sample, spec coverage and how recent the latest review is.

• High: ±2 points
• Medium: ±4 points
• Low: ±7 points, labeled Provisional and never ranked as a "best" pick

Two products within each other's margin should be treated as roughly equal. Read the pillar details to decide which suits you.

Independence & corrections

• Brands cannot pay for, preview or edit their scores.
• Affiliate commissions and advertising never affect scores, and the engine has no input for them.
• Manufacturers may submit factual corrections (a wrong spec, a resolved recall) to hello@ridingnexus.com. We verify every correction before changing data.
• Scores are reviewed at least quarterly, and immediately when a recall is announced or resolved.

Changelog

• v1.0 (October 2026): launch of the five-pillar methodology, category benchmarks, Bayesian owner scores, safety caps and confidence margins.

Have a suggestion to make the Nexus Score fairer? We'd love to hear it.