The Fair Score rates how fair a robot is to own: can you repair it, will it keep working, does it stay yours, and what does it really cost over the years. It does not measure how well the robot cleans or mows. Performance is shown separately, from cited third-party tests, so the two never blur together.
First market: Spain (region ES), with EU-wide terms recorded as EU. Prices are in EUR from Spanish retailers or the brand's Spanish/EU store; warranties are compared with Spain's 3-year statutory guarantee; a part counts as available only if it can be bought for delivery to Spain. Other markets get their own assessment later.
Principles
Evidence or zero. Every score needs a cited source checked on a stated date. Not disclosed scores 0 and is shown as "Not disclosed", never guessed. Manufacturers can raise their score by publishing the information.
Money never touches the score. Affiliate status, commission rates and brand relationships are not inputs. Scorers don't see offer or commission data. Brands cannot pay for review, placement or rescoring. Products without an affiliate program are listed and recommended on the same terms.
Reproducible. Each criterion has fixed levels. Two scorers with the same evidence should land on the same number; a second person reviews every assessment before it is published.
Versioned. Scores record the Standard version they were made under. Changes to criteria or weights get a new version and a public changelog; existing scores are recomputed, not silently mixed.
Rechecked. Every published assessment is re-verified at least every 6 months, and whenever a firmware, warranty or parts policy change is reported.
If a criterion doesn't apply (e.g. P2 for a robot with no camera), it is marked N/A and the pillar is rescaled to its full weight from the criteria that do apply. R1 and C2 are N/A for mains-powered robots with no battery (e.g. corded pool robots).
Under each level: which of the 20 robots in our first Spanish batch landed there. Draft scores, in review.
Repair & parts 30 points
R1 Battery replacement
8 points
8User-replaceable with common tools, and an official replacement battery is sold4 robots
Whether the brand still sells parts, for delivery to the scored region, for its models in the same category launched at least 5 years ago. (Spanish law already requires parts for 10 years after a product stops being made; this tests whether that happens in practice.)
5At least half of key modules orderable for a model in the category launched 5+ years ago, including the battery5 robots
Wear modules (brushes, filters, blades, cleaning pads) never lift R3 above 1 on their own. The comparison model is the brand's model in the same category from the most recent launch year at least 5 years before the check date; where there are several, the one sold in Spain in the largest volume or, failing that, the flagship. Parts count when orderable for delivery to the scored region from the brand or an authorised dealer; a general marketplace selling on its own account doesn't count.
R4 Repair information
5 points
5Official service manual or step-by-step repair guides public for key modulesnone yet
3Partial (e.g. videos for some parts only)9 robots
Commercial warranty in the scored region, measured against that region's statutory guarantee (3 years in Spain since 2022; 2 years in most EU countries)
6Device covered at least 2 years beyond the statutory guarantee (5+ years in Spain), battery includednone yet
3Device covered at least 1 year beyond it (4+ years in Spain), or battery capacity covered for the whole statutory period2 robots
Official consumables at the manufacturer's recommended intervals, as a percentage of the device's launch price. Where the manufacturer gives a range of intervals, the midpoint is used; kits that duplicate individually listed items are not counted twice
8At most 5% of launch price per year (robot vacuums: 10%)2 robots
Core consumables: robot vacuums: main brush, side brush and filter, plus dust bags and mop pads where the robot uses them; lawn mowers: blades; pool robots: filters, plus brushes where fitted; window robots: cleaning pads. All must have an official price and the manufacturer's replacement interval, or C1 scores 0 ("Not disclosed"). Otherwise the missing items would make the robot look cheaper than it is. Detergents and other usage-based supplies don't count.
Robot vacuums have their own thresholds: their mop pads, rollers and bags make consumables a much larger share of the price (18–47% a year in the first Spanish batch where every core item is priced, against 1–9% for lawn mowers). Pool and window thresholds will be reviewed once there is enough data.
C2 Battery replacement price
6 points
Official battery price as a percentage of the device's launch price
6Official battery at most 10% of launch price2 robots
A gate failure doesn't change the number, but the product cannot be a Fair Pick, and the failure is shown on its page. A gate fails only on evidence that the problem exists. When the criterion scores 0 because the information is not disclosed, the points are lost but the gate does not fail; the page says "Not disclosed" instead of making a claim we can't back up.
G1Needs the maker’s cloud. Core functions depend on the manufacturer’s servers.3 robots
Source type: manufacturer documentation, manufacturer statement (in writing), our hands-on test, independent teardown, independent lab test, or owner reports (only as a supporting source, never the only one)
For every source: the URL, an archived copy, the date it was checked, and the exact excerpt relied on
Still being decided
Should older models' track record (S2, R3) be scored per brand or per product line?
How should a model that is sold but discontinued be shown (score frozen, "parts until" date shown)?
Pool and window C1 thresholds, once at least 10 products per category are scored.
Changelog
v0.3 (2026-10-10, draft)
C1 requires every core consumable for the category to have an official price and interval. Previously C1 was computed from whichever items were priced, so gaps made a robot look cheaper (6 of the first 20 Spanish assessments were affected).
v0.2 (2026-10-08, draft), after the first 20 Spanish assessments:
R3 now tests whether parts are orderable for models launched at least 5 years ago, replacing "published parts commitment": in Spain, a 10-year parts obligation is already law.
C1 has its own thresholds for robot vacuums (10 / 20 / 30%).
The bottom tier is renamed from "Not recommended" to "Doesn't meet the Fair Standard". It states a fact about our criteria, not a verdict on the product.
Criteria that almost no product meets yet (R5 independent repair, I4 end-of-life commitment) keep their full weight as targets.
R3 clarified after the first research pass: wear parts alone never lift it above 1, the top level needs the battery, and the comparison model is fixed by rule rather than chosen.
New evidence source type press (trade press and news), so press articles aren't labelled as independent tests.
v0.1 calibration (2026-10-08, draft):
Gates fail only on evidence.
S3 is measured against the region's statutory guarantee (3 years in Spain).