EDITORIAL POLICY

How we keep scoring separate from selling.

Least Processed scores products by ingredient and processing data alone. The score formula doesn't know who pays us, and people who pay us don't get to read or change the score formula. This page explains the mechanics so you can verify for yourself.

What our score looks at

For every vertical the score is a single 0–100 number on the same five-band scale (A 85+, B 70+, C 50+, D 30+, F <30). Different inputs feed it per vertical:

The full per-vertical formula is in scripts/build_catalogue_multi_vertical.py. We re-run it any time the underlying Open Food Facts / Open Beauty Facts / Open Pet Food Facts / Open Products Facts data changes. The score doesn't read any sponsorship data, by construction — sponsorship data isn't in the same code path.

Where sponsored placements appear

Some pages carry sponsored retailer placements. Every sponsored placement is disclosed inline with the placement itself, never in a footnote, in plain language: "SPONSORED — DOESN'T INFLUENCE SCORE."

A sponsored placement buys visibility in three places:

It does not buy a higher score, a higher rank in the index, or removal of competitor brands. The same brand that pays for a sponsored slot still has every product they sell scored on the public methodology, alongside every other brand. If their score is poor, their score is poor — placement chip or no.

What we won't take money for

Compliance

This policy is written to satisfy the UK CAP Code §11 (advertorial), the ASA non-broadcast guidance on identifying ads, and the US FTC §5 plus 16 CFR Part 255 (Endorsement Guides). Disclosures are positioned and worded to be clear, prominent, unambiguous, and in close proximity to every sponsored claim.

Spotted something that looks like it breaks this policy? Email alexdenne@gmail.com. We'll investigate, fix, and (if there's a wider lesson) update this page publicly.