How can you measure inflation for a specific product category when official statistics are too broad?
Official price indices often group products together (for example, US CPI reports candy and chewing gum as one item). Category-specific inflation can be estimated from online prices by tracking identical products over time (matched-SKU), normalizing to a price per unit to capture shrinkflation, and stating the retailer coverage and weighting clearly.
Last reviewed
The method
- 1Define the category precisely using a standardized category tree, not retailer menus.
- 2Build a matched panel of the same SKUs observed in both periods.
- 3Convert shelf prices to price per unit (per ounce, per 100g, per liter).
- 4Aggregate with a stated weighting: unweighted, or weighted by retailer share.
- 5Report the difference between average-price change and matched-SKU change as mix shift.
How to cite it
Scope the claim in the sentence itself: which retailers, online listed prices, matched-SKU basis, price per unit. For example: 'Across Data Purl-covered US retailers, the matched-SKU price per ounce of the category rose X% between two dates.' That scoping is what makes the number defensible.
Metrics used
- Matched-SKU inflation (same-SKU price change): The price change of identical products observed in both periods, which isolates true price moves from changes in product mix.
- Shrinkflation: A price increase delivered by reducing pack size while keeping the shelf price the same or similar.
- Mix shift: The part of an average-price change explained by a change in which products are on sale, rather than by price changes on the same products.
How each measure is built: methodology.
See the data for your coverage
Start free in the analytics portal, or talk to the founders about the full dataset delivered to Snowflake or S3.