Independence

Our relationship with Ella's Soapbox

Roll Call is built and owned by Ella's Soapbox, a handmade soap business — we’d rather tell you here than have you find out later. What that means: this app was made by someone who reads INCI lists for a living and got tired of fear-scores. What it doesn’t mean: any advantage for our own products. An Ella's Soapbox label gets the same explanations, the same evidence grades, and the same ordering as any other maker’s — Roll Call cannot tell the difference, and neither will you.

We never rate, rank, or recommend — including our own

The hard rule, for the record: no product scores, no rankings, no “cleaner alternative” suggestions — and no exceptions for products we make. Roll Call will never tell you what to buy or what to avoid. It tells you what’s in the thing in your hand and what the evidence actually says, and the same ingredient gets the same explanation whoever makes the product.

Makers pay for tooling, never for a verdict

Makers can claim their product on Roll Call to add a display name, a short blurb, and a link. They can also pay for tooling: a QR and booth kit, and label pre-flights that catch parsing problems before a label run gets printed. That is a paid relationship, and we’re disclosing it in plain sight. The “On the Roll” badge means exactly two things: the maker claimed this product, and its label parses cleanly. It is not an approval, not an endorsement, and not a safety verdict. A claimed product’s ingredient explanations, evidence grades, and ordering are byte-for-byte identical to an unclaimed one — paying changes the frame around the sheet, never the sheet.

Where our data comes from

When we look a barcode up, the places we look are public, non-commercial databases, and we’ll name them. Personal-care products are matched against Open Beauty Facts, a free, community-maintained product database. Supplements are matched against the Dietary Supplement Label Database (DSLD), published by the National Institutes of Health’s Office of Dietary Supplements — a US-government database of supplement labels, not a brand-funded one. Neither charges us and neither pays us, and we send them only a barcode number, never anything about you. The evidence grades are ours; these databases supply the label, never a verdict. DSLD is public information from the U.S. government, used with attribution to the NIH Office of Dietary Supplements; Open Beauty Facts is used under its ODbL license. Both are credited in the footer of every page, and neither of them endorses us.

Who ran the studies, and who paid for them

Every document we cite carries its kind on the card — Peer-reviewed, Regulatory, Industry review, or Database — because who produced a finding is part of the finding.

• Industry review. The Cosmetic Ingredient Review is funded by the Personal Care Products Council, the cosmetics industry’s trade association. CIR states that its expert panel and review process are “independent from the Council and the cosmetics industry”. We label these Industry review and let you weigh that for yourself.
• Regulatory. The EU Scientific Committee on Consumer Safety, EFSA, the FDA, the Code of Federal Regulations. Publicly funded, and answerable to a government rather than to a customer.
• Database and government reference. Much of the supplement side is the NIH — the Office of Dietary Supplements, NCCIH, MedlinePlus. We would rather name it than blur it into “research shows”.
• Peer-reviewed. Journal studies and systematic reviews. Funding varies from paper to paper, which is the next point.

Those four kinds fall into two groups that need different answers. The regulatory documents and the reference databases — 380 of the 1664 sources this site currently cites — are published by bodies whose name is itself the answer to who did the work, and we print that name on every citation. The other 1284 sources are studies, counting the industry reviews alongside the peer-reviewed papers, and there “who paid for this?” is a real and separate question.

We asked PubMed and Crossref about the 655 that carry an identifier we could look up. The other 629 we could not ask about at all — no identifier in the citation, or no entry under the one we had — and for those we know nothing beyond what the citation itself says. The work we cite runs from 1943 to 2026 — measured on the 1364 cited sources that record a year at all, not on all 1664.

• 53 are recorded as standing as published, and 3 carry a published correction, an expression of concern, or a retraction. That leaves 599 papers Crossref knows about but PubMed has no record of. Crossref only reports a retraction if the publisher filed one, so we cannot say either way, and we would rather say that than count them as clean.
• 372 name a funder in those records — 218 funded by public bodies, 18 by charitable foundations, 136 by industry. Where a study was paid for by both a public body and a company, we count it as industry, because that is the part you would want to know. A further 155 carry a funding statement that names nobody clearly. Those are not counted above: a statement we read and could not follow is not a funder we know.
• 560 carry a conflict-of-interest statement at all; in 137 of those the authors declare a conflict, and in 423 they declare none. A further 2 we have read and found to carry no such statement at all. Those are not counted above: a paper that says nothing about its conflicts has not disclosed them, and that is a different fact from one nobody has checked yet.
• 24 have a free full text you can open and read yourself, without paying anyone.

Those numbers are smaller than we would like, and we would rather show you a small true number than a reassuring one. A study missing from the funder count is almost always one where the publisher never filed that information — not one we established was independently funded. We have not yet read the funding statements inside the papers themselves. When we do, the count will go up and this paragraph will change.

Which leaves the rule that holds the rest of it together: wherever this site says unknown, it means we have not been able to check. It never means we checked and found nothing. Those are different facts, and keeping them apart is the difference between telling you what we know and telling you what you want to hear.

We are not asking you to trust any of these bodies, government ones included. We are telling you which one said it, so you can decide what that is worth.

Membership never buys different facts

Membership changes limits, never answers: unlimited scans and an unlimited watch list. The explanations, the evidence grades, which ingredients answer the roll — identical for a member and a stranger. No brand can pay for a verdict; makers pay only for tooling on their own listing, as above.