The single most common provenance pattern in commerce is the badge. A round green pictogram, a recycled-content claim, a "made in" tag — placed on the product card, linked to a separate page that nobody opens. The pattern is almost universal, and it is almost universally ineffective. Customers ignore the badge. Regulators, increasingly, are starting not to.
The reason the badge fails is straightforward: it is signposting, not information. It tells the customer that the merchant has a position; it does not tell the customer anything about the position. The actual question the customer is trying to answer — where does this thing come from, who made it, what is it made of, is the claim being made about it true — is unanswered, or answered on a page that requires a deliberate detour from the buying flow.
The current state
Most provenance UX is a virtue-signalling badge linked to a press-release page. The page, when opened, contains glossy photography, a paragraph of mission-statement prose, and a vague reference to "our suppliers." There are no specifics. There are no evidence links. The information density is approximately that of a magazine advertisement, and the customer reads it accordingly — which is to say, almost not at all. The conversion impact of these pages is reliably indistinguishable from zero. The trust impact, when the page is read carefully, is mildly negative.
This is not, in our view, a failing of provenance as a concept. It is a failing of provenance as practice. The information being conveyed is poorly structured, badly placed, and not evidenced. The fact that customers ignore it is a perfectly rational response to it; the fact that some merchants have given up on provenance UX altogether is an over-reaction, but a comprehensible one.
Provenance as a working graph
Trove Provenance treats provenance as data, not as signage. Every SKU is linked to a materials graph (what is it made of, where do the materials come from, what is the certification status), a makers graph (who made it, where, at what scale, in what relationship to the merchant), and an origin graph (where the raw material originates, where assembly happens, where finishing happens). The data is structured. It is queryable. It is auditable. Every claim has an evidence link, and every claim has a status — verified, unverified, refused.
That last word — refused — is the disciplined one. There are claims merchants would like to publish that we cannot evidence. Where the evidence is missing, the claim is not surfaced. The merchant sees the gap on her dashboard and either fills it (by sourcing the evidence) or accepts the silence. Some merchants find this uncomfortable in the short term. The ones who persist with the discipline end up, twelve months in, with a provenance dataset that is genuinely defensible — and a customer base that has noticed.
Surfaced naturally
Provenance appears in the product description where it is relevant. The origin of the oak in a chair. The specific dye-house behind a textile. The atelier responsible for the assembly. The country of origin where it actually means something — not the regulatory minimum, the working answer. For categories where provenance is a category-wide concern (sustainable fashion, food, wood objects, handmade textiles), a short category-level summary appears at the top of the page. The summary is short, specific, and silent on anything we cannot evidence.
What we do not do is badge. There is no round pictogram on the product card. There is no separate "our values" page hovering off the main navigation, waiting to be ignored. The information is encountered in the description, at the moment it is relevant, in the same prose register as the rest of the page. The customer does not feel lectured because the customer is not being lectured.
The badge is signposting, not information. The customer ignores it. The discipline is to put the information into the description, at the moment it matters, in the same voice as the rest of the page — and to be silent everywhere it doesn't.
Customer impact
Across the three pilot stores in which we measured the change in detail, surfacing provenance naturally moved conversion up by four to six per cent on the relevant categories. That is the easier number to publish. Return rates dropped in parallel — by two to five per cent on the same categories — because customers buying with full information return less often. The return-rate drop is the operationally significant number; it is what shifts the unit economics of a small boutique by a degree that genuinely matters.
What did not move was anything related to engagement. Click-through rates were unchanged. Time on product pages went up slightly, but not enough to count. The customers who cared about provenance read the information and bought; the customers who did not care moved through the page at their normal pace and were not, by anything we did, made to feel they had been preached at.
What we won't do
Greenwash. We will not publish provenance claims we cannot evidence. We will not aggregate poorly-evidenced supplier statements and present the aggregate as proven fact. We will not run a "premium tier" where claim evidence is treated more loosely. The provenance discipline is one discipline; the audit log is one log; no client account can override a "refused" claim into a public assertion. Some merchants find these positions difficult; we lose those merchants quickly, which we have come to understand is part of what makes Trove the kind of firm it is. The merchants who stay end up with stronger brands than the ones who looked elsewhere — not, in our experience, by a small margin.