How Scoring Works

A rating, or an honest "not yet"

When we've assessed a brand, you get a plain-language rating. When we haven't, you get told so. You never get a number.

Behind every brand is a question: how much do we actually know? That answer gets stated instead of blurred into one confident-looking figure. This site used to publish a 0–100 score built from corporate disclosure paperwork; it read like a verdict it couldn't support, and it rewarded big companies for filing reports rather than for doing better. It's gone.

What the labels mean

Every brand lands on one simple ladder, from Below Average up to Excellent.

Why a paperwork-only brand tops out at "Good"

A giant company can publish mountains of paperwork. That shouldn't let it out-rank a genuine mission brand — so a data-only score stops at "Good."

Large corporations have whole teams devoted to disclosure. If we let the raw number climb without limit, the top of the ladder would fill up with the biggest companies simply because they produce the most reports — not because they're doing the most good.

So we draw a line. "Great" and "Excellent" can't be reached with paperwork alone. They're reserved for mission-driven brands that back up their commitments with real, independent certifications — the kind that require outside verification, not a press release.

How the assessment is built

One rating is really a summary of dozens of specific checks, grouped into three big areas.

Nothing here is scored on a vibe. Hundreds of individual data points are grouped into sub-categories, which combine into three pillars, which combine into the single overall assessment behind a brand's rating.

How estimated scores are built

For brands without verified data, a fixed set of focused questions is answered using publicly available information — sustainability reports, certifications, news, and filings — and those answers roll up through the exact same three-pillar structure. A research agent does the reading, working only from the brand's own public pages, certifier directories, and credible reporting — and every answer it returns has to carry a direct quote and a link, or it is dropped before it reaches the page. The agent never sets the rating itself: the rules that turn those answers into a rating are fixed, and identical for every brand. The result is always shown as a "Likely" estimate, never as a verified number, so the two are never confused. The full sourcing rules are on Methodology.

What moves a score up or down

Real-world actions adjust the score — certifications lift it, controversies pull it down.

The baseline comes from the data above, then two things nudge it to reflect what a company actually does. They are also the two things we check by hand: a certification only counts once it's confirmed in the certifier's own registry or records, and a controversy only moves a rating once a person has read the source it came from — see Methodology for how that review works.

• Certifications add points. Independent certifications are worth more than promises. A B Corp certification adds the most of any single certification, because it reflects a verified, company-wide standard. Others — like Fair Trade, 1% for the Planet, or a validated climate target — each add a more modest lift.
• Controversies subtract points. The penalty scales with how serious the issue is: a minor concern nudges the score down a little, while a severe or critical problem docks a lot. The most serious issues also ripple beyond their own category — because a major failure in one area usually signals something broader about how a company is run.

Where these show up on a brand's page

On a brand's page you'll see a Certifications & Commitments section and a Reported Controversies section, each noting how it affected the score. Reported controversies are drawn from public sources and clearly flagged as allegations we haven't independently verified — their inclusion doesn't imply wrongdoing. Where a review found that a finding belongs to a parent company or a supplier rather than to this brand, it stays on the page labelled that way and scores nothing; where the source didn't support the claim at all, it's removed.

How we protect smaller brands

A small brand that doesn't publish a giant sustainability report isn't penalized for it. Missing paperwork isn't the same as bad behavior.

Big corporations have the staff to document everything; small and independent brands often don't, even when their actual practices are strong. If we scored purely on how much a company publishes, we'd punish exactly the kind of brands many shoppers want to support.

So when a brand has little public data, we nudge its score toward the middle of the pack rather than dragging it to the bottom. This only ever helps — a score is never pushed down because a brand is quiet.

Confidence and data coverage

You get told how sure we are, not just where a brand landed.

Alongside every rating we track two honesty signals: confidence (how sure we are of the read) and data coverage (how much public information there was to work with). When confidence is low, or coverage is simply too thin, no rating is forced onto the brand — it stays Rating pending.

Where this falls short

Every rating method has limits. Here are ours, stated plainly, so you can weigh a rating knowing what's behind it.

Coverage is uneven, and thin in places

We haven't assessed every brand on the market yet. Whole categories are sparse. Some brands are missing entirely. If a brand you're looking for isn't here, that's a gap in our coverage, not a judgment about the brand.

Coverage is also skewed toward brands with a public record. A company that reports, certifies and gets written about is easier to assess than a small private label that publishes nothing — so the brands you see rated are not a random sample of what's on the shelf.

What a certification does and doesn't prove

Certifications are the strongest evidence we have, which is exactly why it's worth being precise about what they mean. A certification proves that an outside body checked something specific, at a point in time, against its own standard. That's real, and it's why certifications carry weight here.

It is not proof that a company is good. It doesn't cover everything the company does — a certification may apply to one product line, one facility, or one part of a supply chain while the rest goes unexamined. Standards differ in how strict they are and how hard they audit. And certifications lapse: a badge on a brand's own website may be years out of date, which is why claims get checked against the issuing registry rather than taken from the brand.

So read a certification as what it is: verified evidence about a defined thing, not a character reference.

Why some brands stay Unrated

A brand stays unrated when the evidence won't support a call. Usually that's one of four situations: there's too little public information to judge; the available information is too contradictory to resolve; we can't confidently identify the company actually responsible for the product; or we simply haven't got to it yet.

None of those mean the brand is bad, and none mean it's good. Unrated is an admission about our information, not a finding about the company. A brand that would score badly and a brand nobody has written about look identical from the outside, and collapsing that distinction into a rating would be the single easiest way to mislead you.

The rest of the fine print

• Ratings describe companies, not individual products. A well-run company can make a wasteful product, and a poorly-rated one can make a genuinely better item. The rating won't catch that.
• Product tags aren't reviewed. The attributes on a product card — BPA-free, organic, plastic-free and the rest — are read off the listing's own metadata and shown as the seller states them. They never feed a brand's rating, and unlike certifications they aren't checked one by one. More on that.
• Public evidence lags reality. Reporting is published on a delay, and a company that changed course this year may not show it in the record for another one.
• Controversies are reported, not adjudicated. They come from public sources and are flagged as allegations we haven't independently verified.
• Absence of a controversy isn't a clean record. It often just means nobody has reported on that company.
• We get things wrong. When you find one, tell us — corrections from readers are among the better signals we get, and they get acted on.