AI Methodology

We evaluate product quality and deal quality as two separate measurements. Product ratings use reviewed evidence from official manufacturer sources; deal verdicts use retailer prices and tracked price history.

Every stored specification keeps its source URL and retrieval evidence. Missing or ambiguous values are left unscored rather than inferred.

Each product receives an eRating: a category-specific score out of 100 calculated from fixed, weighted product criteria. Each offer receives a separate deal verdict, so a good product cannot become highly rated merely because it is discounted.

Verified attributes: only evidence reviewed from official manufacturer sources.

Category metrics: fixed measures relevant to that product type, such as display, performance, camera, battery, and build for smartphones.

Metric weights: a published methodology determines how much each area contributes.

Deal evidence: current price, reference price, tracked history, and recency are assessed separately from the product score.

This separation lets you see whether a product is good, whether today's offer is good, or both.

No guesswork. No inflated discounts. Just data-driven decisions you can trust.