Whatsbetter Score
The Whatsbetter Score is a comprehensive product quality rating on a scale from 40 to 100. It combines four independent signals — customer reviews (50%), brand reputation (20%), bestseller ranking (20%), and technical features (10%) — into a comparable score within the same product category.
What does the Whatsbetter Score measure?
The Whatsbetter Score answers the key question: "How good is this product compared to other products in the same category?"
Unlike simple star ratings, which can be misleading when comparing different categories, the Whatsbetter Score ensures a fair comparison: a highly-rated jigsaw puzzle is never compared to a security camera, only to other products in its own category.
The score is made up of four independent rating pillars, each offering a different perspective on product quality. Combining multiple signals balances out the distortions of any single data source and produces a robust overall rating.
The four rating pillars
Each pillar covers a distinct aspect of product quality. The weightings reflect the reliability and relevance of each data source.
Ratings and reviews from real buyers — the most direct indicator of real-world product quality.
A brand's quality track record within the category, plus public perception on social media.
Market validation through real purchase decisions — the "wisdom of the crowd" as an objective reality check.
Objective analysis of category-specific product attributes such as material quality, feature set, and build quality.
Pillar 1: Customer reviews
Customer reviews form the core of the Whatsbetter Score and have the greatest influence on the final result. They reflect real user experience and are built from multiple signals to give a nuanced picture of quality.
A multi-signal approach
Rather than relying solely on the average star rating, the Whatsbetter Score analyzes several dimensions of customer feedback:
- Overall ratings: The average star rating from platforms like Amazon and Google Shopping, normalized to a common scale. Ratings from other platforms are corrected per category, since they tend to run higher than on Amazon. This signal carries the most weight within the reviews pillar.
- Review sentiment: Beyond the star count, we analyze what customers actually write. Only product-related experiences are counted — comments about delivery, packaging, or seller service are filtered out so external logistics don't distort the score.
- Price perception: How do customers rate value for money? Products that the majority consider overpriced receive a small score reduction.
- Defect reports: Systematic quality issues matter most: when multiple customers independently report the same defect, it lowers the rating. Isolated, one-off complaints are not overweighted.
Cross-platform rating correction
Ratings from platforms other than Amazon (e.g. Google Shopping) aren't taken at face value. For each category, we calculate a platform offset based on overlapping products and correct for systematic rating differences, keeping the comparison fair regardless of the rating source.
Expert review bonus
Verified results from independent tests can further strengthen the reviews pillar: products that score especially well in expert reviews receive a moderate bonus of up to 4.5% within this pillar.
Statistical normalization
Different product categories naturally have different rating levels. The Whatsbetter Score corrects for this distortion in customer reviews, brand reputation, and bestseller ranking, so only products within the same category are compared. Technical feature analysis is the exception: it is not normalized relative to the category but scored on absolute performance.
Pillar 2: Brand reputation
Brand reputation reflects a manufacturer's quality track record within a product category. It consists of two components:
Rating-based brand score
How have all of a brand's products performed within the category? Brands that consistently deliver high quality receive a higher score. Products with more ratings carry more weight — a single niche product influences the brand score less than a bestseller with thousands of reviews.
Social media sentiment analysis
We also analyze public brand perception on social media such as Reddit and Twitter. We account not just for tone (positive vs. negative) but also the quality and relevance of the discussion. Product-related criticism counts more heavily than general complaints about delivery or the website.
Brand position bonus
Leading brands within a category receive an additional bonus based on their overall standing within that product group.
Pillar 3: Bestseller ranking
The bestseller ranking captures the "wisdom of the crowd": which products are actually bought most often? The underlying data source is Amazon Sales Rank.
Amazon Sales Rank is collected for every product in the category.
Mis-categorized products and accessories are removed from the bestseller list.
The rank is converted to a percentile relative to the cleaned category size.
The result feeds into the overall rating as market validation.
Bestseller proxy for non-Amazon products
Not every product has an Amazon Sales Rank. For products sold mainly on other platforms, we calculate a bestseller proxy based on rating activity on that platform. The value is adjusted per category and weighted conservatively, so market relevance stays fair outside Amazon too.
Category cleanup
Amazon's bestseller lists often contain mis-categorized products (e.g. video games showing up under baby monitors). Such outliers are systematically removed so only relevant products influence the rating.
This signal provides an objective reality check: products that have proven themselves in the market get appropriate credit in the score. A category winner among 10,000 competitors receives the highest possible value.
Pillar 4: Technical feature analysis
Every product category has specific quality attributes that can be assessed objectively. Relevant attributes are extracted with AI from spec sheets and product descriptions, then scored:
For each category, relevant product attributes are automatically identified from technical data and descriptions and stored in a database.
The identified attributes are ranked and weighted by their importance to customers and to product quality.
Every product is scored against its attributes to identify the product with the best features in the category.
Absolute rather than relative scoring
Unlike the other pillars, technical features aren't normalized relative to the category. Actual performance is what counts — so simple products are neither artificially inflated nor penalized.
Fairness principle
When ratings for individual attributes are still missing, a neutral starting point is used. Products aren't penalized just because data is missing for some aspects.
Data quality and minimum requirements
To ensure the reliability of the Whatsbetter Score, strict quality requirements apply to the underlying source data:
Bayesian smoothing
Products with very few ratings could randomly land on extreme scores. The Whatsbetter Score applies statistical smoothing that cautiously pulls new products toward the category average. As the data pool grows, the score depends increasingly on real feedback.
Minimum data baseline
A product needs at least one real signal (e.g. customer reviews or bestseller ranking) to receive a Whatsbetter Score. Products with no data foundation at all are excluded from scoring — preventing empty records from skewing the distribution.
Scoring and highlighting
The Whatsbetter Score is calculated on a scale from 40 to 100 points. The weighted results of all four pillars are combined per product and normalized within the category. We then apply S-curve smoothing (a sigmoid transformation) to better distinguish similar products at the top of the ranking.
The higher a product's Whatsbetter Score, the better it performs relative to other products in the same category. The rating is always relative to its category — a score of 85 for headphones isn't directly comparable to a score of 85 for suitcases.
Rating pillars at a glance
Analysis of product ratings, product-focused review sentiment, price perception, and defect reports — including platform correction and an optional expert-review bonus.
A rating-based quality track record for the brand within the category, supplemented by sentiment analysis on social media such as Reddit and Twitter.
Market validation via Amazon Sales Rank and a conservative bestseller proxy for products on other platforms.
AI-assisted extraction and objective scoring of category-specific product attributes based on absolute performance.
Research and scoring are conducted independently and published editorially. The Whatsbetter Score is calculated entirely from data, with no influence from manufacturers or retailers. Positively-rated suppliers may, after the review is complete, license the quality seal for promotional use. Licensing has no effect on either the methodology or the results.