Kitchen carried the most overstated Amazon.ca discounts. Office carried the fewest.

The pattern holds across categories, at different rates.

Every category below cleared a 30-product floor for its comparable cohort — below that, a single product’s classification can swing the rate by several points, which is not a sample worth publishing a percentage for. Toys and Baby did not clear it (Toys at 21, Baby at 19) and are not broken out individually here, though their products remain in the catalogue-wide count.

A higher rate is not proof any specific listing is misleading.

This describes a category’s tendency across many products, not any one listing. A genuine, temporary reference price is indistinguishable from a stale one by this method alone — see how to read a reference-price discount for what to check on a specific product. The catalogue-wide version of this same test — the September reference-price study — found reference prices overstating the discount by 10+ points in 26% of its comparable cohort; the category rates above range from 10% to 33% around that figure.

“Overstates” is a defined comparison, not an accusation: it means the reference price implies a larger discount than Dipp’s own recorded typical price would, by 10 percentage points or more, for that product. It does not establish that any seller broke a rule, mispriced anything deceptively, or that the resulting discount is invalid — a reference price can legitimately differ from Dipp’s own recorded window for reasons that have nothing to do with deception, including a genuine price change Dipp has not yet observed, a different comparison period, or a real, temporary promotional price the product actually sold at elsewhere.