How the Trust Score works.

The score looks at Health, Planet, Ethics, and Longevity. The criteria below are read from the method tables for each category in the pilot.

Consumer electronics, Fashion, Furniture, Personal care are not in the pilot yet.

The EU product registry, EPREL, supplies model facts such as energy, water, noise, and guarantee information.

EU law supplies facts that apply to a whole product group. These include spare parts periods and end of life routes.

Company reports and public registers supply company facts. A company fact applies to every product of that company.

Owners' reviews and problem reports are shown on product pages. They are not part of the Trust Score.

Each fact moves its dimension up or down from a midpoint of 50. A fact confirmed by an independent body counts most. A brand's own claim without a document counts for nothing. Our own research counts only with a source link.

A dimension's score, from 0 to 100, is the weighted average of its facts.

See the worked example for a real product and its sources.

A dimension gets a score only when enough facts are held. That means one required fact or two optional facts. One optional fact is enough when the dimension has no required fact for that category.

An overall score needs 3 of the four dimensions. Otherwise we show "Not enough data yet". We never show 0 or a guess.

A negative finding, such as a safety alert or a controversy, is shown when found. It never makes a score on its own.

Some checks count as checked without moving a score. Examples include "not listed in the B Corp directory" and "publishes a sustainability report".

Each category gives the four dimensions a set share of the overall score. When a dimension has no score, the others are reweighted to add up to 100.

High means at least 70% of the facts that decide that dimension are held. Medium means at least 45%. Low means less.

Verified is High where an independent body confirmed a fact. Beside the overall score, the tier is the lowest of the dimensions that make up that score.

The count shows how many of the facts that apply to this product we hold.

The newest check of a fact decides. "Last checked on" is the date of the latest check for that product.

"Not confirmed" means a fact we recorded earlier could not be confirmed from its source. It no longer counts.

When two registry records for the same model disagree, we count neither until the difference is resolved.

No brand can pay for a place in a list or to change a score. What a brand says in its marketing never moves a score. A brand can send us documents. We review them and publish what we decided, accepted or not. No one can type a score by hand. Every score is recalculated from the facts.

Reviews and problem reports come from members who added the product to their Ownership log. We don't check receipts. They are not part of the Trust Score.

A description is shared with the brand, or shown on the product page, only if its owner turns that on. It is never shared with a name or contact details. Problem counts appear only once three or more owners report.

Today no product has a Health score. Every overall score is made from Planet, Ethics, and Longevity, reweighted.

The catalogue currently has 189 appliances from 6 brands, in these groups: air conditioners (25), dishwashers (19), ovens (25), range hoods (24), refrigerating appliances (27), tumble dryers (25), washer-dryers (25), washing machines (19). 62 have an overall Trust Score.

Anyone can report a wrong fact. The report asks us to review the source. It does not change evidence or a score by itself.

When evidence is corrected, the score is rederived from the current facts. A number cannot be written directly into the score.

Brands can send primary documents for review. We publish what we decided. A score changes only when admissible evidence is stored as a fact and the score is recalculated.