It is BestBuysUK's own data-led editorial assessment. We do not convert Amazon stars, retailer ratings or review counts into the BestBuysUK Score.
1. We identify the exact product first
Before a score is useful, we need confidence that the information describes one exact model or variant. Our workflow checks the product name, brand, model, identifying codes and description together.
An Amazon ASIN can be used internally as an identifier for an imported product, but an official manufacturer website is not expected to publish Amazon's ASIN. The important question is whether the actual BestBuysUK product identity can be verified.
2. We assess the available product information
Our process brings together the available product description, specifications, image, price, availability and identifying details. Missing information is not treated as a reason to invent a claim.
We do not use Amazon customer reviews, Amazon star ratings or Amazon review counts as evidence for our editorial score, and our research workflow is designed not to scrape Amazon product-page copy as its research source.
3. We block unsupported or mismatched claims
If the available information does not support a specification, the AI must leave it out. If a generated feature belongs to another product type — such as a cooking function on a dishwasher or tumble dryer — the content is corrected or withheld.
Examples of material conflicts include capacity, dimensions, power, propulsion, accessories, functions, supported modes, area suitability and other facts likely to affect a buying decision.
4. Six core scoring dimensions
Each product is assessed across six core dimensions. These provide a consistent framework across BestBuysUK while the evidence used to judge each dimension remains relevant to the product category.
How well does the product's verified capability match the job it is designed to do?
Does the overall proposition justify its market position, capabilities and compromises?
How straightforward is ownership, operation, setup, handling, cleaning or maintenance?
Are the genuinely useful functions competitive and relevant rather than simply numerous?
What do the design, materials, construction, warranty/support and product positioning indicate?
How efficiently does the product use relevant resources such as energy, battery, water, space or time?
Performance 30% + Value for Money 20% + Ease of Use 15% + Features 15% + Build Quality 15% + Efficiency 5% = BestBuysUK Score /100.
5. Category-specific interpretation
The six headline dimensions remain consistent, but the underlying evidence changes with the category. A feature that matters greatly for one product can be irrelevant to another.
Manufacturer lawn suitability, power source, drive type, cutting width and height range, collection, mulching, rear roller, storage and maintenance.
Capacity, energy performance, spin, useful programmes, noise, dimensions, installation, smart functions and warranty/support.
Extraction capability, intended room size, laundry function, tank/drainage, noise, humidistat control, energy use and portability.
Usable capacity, basket/zone design, temperature control, programmes, countertop footprint, cleaning and energy use.
Power format, runtime/reach, floorhead suitability, pet-hair capability, filtration, weight, capacity, tools and storage.
Machine type, drink choice, milk system, grinder controls, water capacity, cleaning, footprint and ongoing consumable costs.
6. Manufacturer suitability is not replaced by arbitrary assumptions
Where a manufacturer provides explicit suitability guidance, we preserve it. For example, if a lawn-mower manufacturer describes a model as intended for “smaller lawns”, BestBuysUK does not relabel it “large lawns” simply because a separate square-metre figure seems large to us.
If only a numerical area is published and the manufacturer does not use small/medium/large wording, we can show that figure without inventing a size classification.
7. Best Overall, Best Value and “Best for…” labels
Best Overall is the strongest all-round ranked product in the current comparison. A Best Value label is intended to recognise a strong combination of overall quality and value rather than merely the lowest price.
Other “Best for…” labels provide buyer context. They should be supported by verified traits and a sufficiently strong overall product assessment. A product should not receive a use-case label simply because a keyword happens to appear in marketing copy or in a sentence explaining what the product cannot do.
8. Structured filters are evidence-led
Category filters such as Petrol, Cordless, Self-Propelled, Rear Roller, Dual Zone or Pet Hair are intended to use structured product facts. Negative phrases, comparison text and “avoid if” wording are not treated as positive evidence that a product possesses a feature.
9. AI assists the process; it does not become the evidence
AI helps BestBuysUK classify products, compare supported facts and write original explanations. It is used alongside structured checks, and unsupported or uncertain claims are not presented as established facts.
AI output can also be rechecked. If authoritative evidence contradicts existing BestBuysUK copy, our workflow is designed to correct the specification, verdict, pros/cons and related buyer guidance rather than preserve an earlier mistake.
10. We do not claim hands-on testing unless it happened
Most BestBuysUK assessments are data-led. We do not use phrases that imply a product was physically handled, measured or tested by us unless that genuinely occurred and the page clearly says so.
11. Product score versus evidence confidence
A high product score and strong evidence are different concepts. A product may appear impressive but still require caution if important identity or specification evidence is uncertain. Identity and information quality are therefore considered separately from the score.
12. Ranking independence
Products are not awarded extra BestBuysUK Score points because a retailer pays a higher affiliate commission. Affiliate relationships are a way of funding the site, not a scoring criterion.
13. Rankings can change
Product ranges, availability and evidence change. A model can move up or down when its assessment is updated, when stronger alternatives are added, when a specification changes, or when a material factual error is corrected.
The date shown on a ranking page tells you when that page was generated or reviewed; it should not be read as a guarantee that every external retailer fact remains unchanged at the moment you click through.
14. Corrections
If you believe we have identified a model incorrectly or published a factual specification that conflicts with authoritative evidence, please send us the details. Useful correction reports include the exact product name and a link to the manufacturer's relevant page or manual.