The detail behind the grades

A plan to pay a living wage only counts as "working towards one" if its deadline is within 5 years, or if it has real, measurable steps due within 5 years (for example, a set share of factories paying a living wage by a given year). Distant targets with nothing measurable in between don't count.

Schemes that count as working towards a living wage: Fair Wear membership, GOTS certification, the Fair Labor Association and a top Chocolate Scorecard rating.

Small coffee roasters and chocolate makers often buy directly from the farms they name, but can't afford certification or independent audits. Research on specialty coffee shows these buyers usually pay farm owners well above market and living-income prices. So when direct trade is genuinely central to a small business, we grade it On Track and mark it with an asterisk, because the evidence comes from the company itself.

• Coffee and chocolate only. "Single-estate" tea usually means large plantations with hired workers, where low pay is well documented.
• Small and independent: under 250 staff, not owned by a larger group, and named farms or producers at the heart of the business.
• Origins must be disclosed, and less than half can come from origins on the US government's forced or child labour lists for that crop (Brazilian coffee; cocoa from Côte d'Ivoire and Ghana).
• Higher prices for farm owners don't always reach the hired workers who pick the crop. That's why these grades never go above On Track and can't be named "Best available".
• Any confirmed labour problem moves the business straight to a poor grade.

Lists are sorted by grade first: Good, then On Track, then Lacks transparency, then Harm on record. Each grade comes from fixed rules about the Good requirements a company meets and its red flags, and every brand page says why (how grades are decided).

Within a grade, companies with stronger evidence, fewer confirmed problems and lower labour risk where they make things come first, and stars move Good companies up. Companies that can't be told apart are listed by name. The number beside each brand is its place in the list you're looking at.

Brands are compared within their own category, such as footwear or appliances, not against everything else. When no brand in a category is Good yet, we highlight the strongest one as "Best available", so you still have a better choice. It's marked with a rosette (), not a star: stars are only for Good companies that go above and beyond.

• Only On Track brands can qualify: no serious red flags (no forced or child labour, abuse or failing ratings), and a proven living wage or real progress towards one.
• The category needs at least 3 rated brands, and estimated grades don't count.
• Only brands selling new goods are compared. Refurbishers, op shops and resellers don't make anything new, so they easily grade well and would otherwise stop us ever naming the best maker (a refurbished phone seller can't tell you which new phone to buy). Where a category has rated second-hand or refurbished options, we show them separately, first among a brand's alternatives and as one row in product search, because buying second-hand is usually the better choice for workers. Every second-hand option we've rated
• It's recalculated automatically whenever a brand is added or re-graded.

If every brand in a category has a poor grade, we don't name one.A category where everyone falls short has no "best".

Confirmed only

Problems count towards a grade only when confirmed by a court, regulator, government list, completed investigation, the company itself, or consistent reporting by several credible outlets. Unconfirmed reports are shown separately and don't affect the grade.

Old problems

An issue that ended more than 15 years ago is removed from a company's record, and so is one that ended 10 to 15 years ago if the company made amends to the people affected. Every other issue is shown with its date and whether the company has made amends.

Problems put right

A problem the company has fully put right for the people affected, such as paying back the wages owed or compensating the workers, stays on its page but no longer counts against its grade. That means a past problem never keeps a company from Good once it's put right, as long as the company passes every other test today. Partly put right isn't enough: an action plan, a part-payment or leaving the supplier still counts until the people affected are made whole. A link through a supplier is different: when a company was only linked to forced labour, for example through Xinjiang cotton or a supplier that received transferred Uyghur workers, and it ended the link by leaving the supplier or region, the link stops counting too. A policy, a pledge or an investigation isn't enough.

Problems a company finds itself

Problems a company finds through its own monitoring, publishes and is fixing (for example, child labour cases found by a cocoa company's own checks) are shown on its page but don't count against its grade. Looking hard and being open about what you find shouldn't be punished. Problems found first by regulators, courts, investigators or journalists always count.

Hidden or unaudited supply chain

No certification or documentation at all, or 30% or more of production in countries with no guarantee of worker rights (ITUC 5 or 5+) without a full independent audit.

These say how much to trust a grade, separately from the grade itself.

• Verified Backed by independent certification or audits.
• Graded by rule Graded Good by rule for reuse, not on evidence about its own workers.
• Partly verified Based on the company's own reporting, with limited independent checks. Where we don't know which countries it makes in, we use where its industry usually makes its goods.
• Little disclosed We researched the company but found little about where its products are made, its certifications or its suppliers, so the grade assumes where its industry usually makes its goods.
• Self-reported Graded on the company's own direct-trade claims. It hasn't published the prices it pays or any independent checks, so we can't confirm that farm workers are paid fairly.
• Estimated No company-specific evidence found. Assumed from where this industry usually makes its goods.