One of the questions I brought to my apparel trust review was whether a larger business would consistently show stronger evidence of trust.
The result was less tidy. Different brands showed different strengths. A larger site could have more developed returns and warranty information while a smaller business made the people behind its work more visible. Useful evidence could also sit on an About page without following the shopper to the product page.
That last observation became more interesting to me than the overall scores.
Publishing evidence somewhere on a website and making it useful at the moment of a decision are different tasks. A certification, product detail or policy can exist and still leave the person considering a purchase with unanswered questions.
My review did not measure whether moving that information would improve conversion or AI citations. It did give me a concrete way to inspect where evidence appeared, where it was only partly visible, and what I would test next.
Conceptual illustration of evidence placed near a purchase decision. The seal is decorative and does not represent a real certification or endorsement.
What I reviewed
For Does Size Buy Trust?, I reviewed A Bronze Age, tentree and Herschel Supply Co. on October 24, 2026.
For each brand, I sampled five page types: the homepage, About page, one product page, returns or warranty information, and contact information. That produced 15 sampled pages.
I used 16 author-defined criteria grouped under experience, expertise, authoritativeness and trust. Each criterion received one point if present, half a point if partly present and zero if not found in the sampled material. Automated reads of visible text and structured data were reviewed, with collapsed content noted as partly present in the method.
Those numbers describe my checklist. They are not scores issued by Google, and they do not measure everything a shopper or search system might consider.
Google’s people-first content guidance discusses E-E-A-T and explains that it is not a specific ranking factor. I use the framework here to organize evidence questions, not to reverse-engineer a secret trust formula.
The study was independent public-page research. I did not work with these brands, access their internal systems or test the effect of changing their pages.
Why I would not lead with the league table
The checklist totals were 4.5 out of 16 for A Bronze Age, 9.5 for tentree and eight for Herschel. The report grouped the businesses by relative size for the comparison.
Those scores can help summarize the method, but they are a poor basis for claiming that one business is more trustworthy overall. The framework gives equal weight to criteria that may have very different relevance to a particular purchase.
It also depends on sampling. One product page cannot represent an entire catalogue. A criterion marked absent may exist elsewhere. The way content is rendered or collapsed can affect what the review detects.
Most importantly, three businesses cannot establish that size causes trust infrastructure, better search performance or stronger customer confidence. The original report’s more assertive headings should be read as exploratory interpretations, not causal findings.
I would use the table to find questions for investigation. I would not use it to tell a client that adding four checklist items will deliver four points of search authority.
Different pages carried different kinds of evidence
In the sampled material, the smallest brand made its founder and production story visible. The larger sites showed other strengths, including reviews or more developed policy information.
I would interpret those as different forms of evidence, not interchangeable badges of quality.
A founder story can explain who is responsible for the business and how it works. It does not establish the fit, durability or suitability of every product. A returns portal can explain a process without demonstrating product expertise. A review can describe one customer’s experience without proving a company-wide claim.
That distinction matters when teams ask for “more E-E-A-T.” The phrase can hide several different problems. An unclear policy needs a policy explanation. An unsupported material claim needs evidence about that material. A vague company identity needs verified identity information.
The right intervention depends on the uncertainty, rather than on which checklist category is easiest to decorate.
The product-page gap was a more useful finding
The report recorded prominent certification and impact information on tentree’s sampled About page. The sampled product page did not display the same certification evidence in the material reviewed.
That is a dated observation about those pages. It does not establish that certification was absent throughout the buying journey or that the page is unchanged today.
What it gave me was a practical audit question: if a shopper arrives directly on this product page, can they find the evidence needed to assess the claim without reconstructing the company’s story across several pages?
The answer does not have to be a large block copied from the About page. A short, scoped explanation and a clear route to supporting material may be more useful. The relationship between the claim and the evidence needs to remain obvious.
I would also verify that the evidence applies to the claim being made. Company-level certification and product-level certification are not automatically the same thing. A general impact figure does not establish the attributes of a specific item.
Moving an ambiguous badge closer to the buy button would not solve that problem. It could make the ambiguity more persuasive.
Evidence placement is a relationship problem
I find it useful to describe the page as a set of relationships: this product has these attributes; this policy applies under these conditions; this claim is supported by this source.
That approach connects trust work with entity and knowledge-graph thinking without requiring an elaborate graph database. The important question is whether the relationship is accurate and understandable.
For example, an organization may hold a certification, a product may use a particular material, and a programme may measure a specific impact. Those are different subjects. Combining them into one broad “sustainable product” statement can lose the distinction the evidence depends on.
A clear page preserves the entity, the claim and its scope. Structured data can express supported relationships, but it cannot repair a claim whose meaning is unclear in the visible content.
My article on why schema cannot rescue a confusing page covers that wider issue. This review gave me a set of actual page observations to bring to the discussion.
Collapsed information deserves inspection, not automatic condemnation
Some details in the review were placed inside tabs or other collapsed sections. The checklist recorded that as partial presence.
That is a methodological choice, not proof that search engines or AI systems cannot access the information. An accordion can be a useful interface. Its behaviour depends on how it is implemented and how the content reaches the page.
For a real audit, I would check the rendered experience, keyboard access, labels and whether the relevant content is available in the page output. I would also ask whether a shopper can predict where to find it.
The conclusion should follow that inspection. “Behind a tab” is not sufficient evidence of a crawlability problem, and “present in the source” is not sufficient evidence that the interface is easy to use.
This is a good example of why a score needs its method attached. Without it, partial presence can look like an objective defect rather than a signal to investigate.
How I would audit proof at the decision point
I would start with a small set of claims that can change a purchase decision. For apparel, those might concern dimensions, material composition, care, return conditions or a specific certification statement.
For each claim, I would identify the evidence source, the person who can verify it, its scope and the page where the shopper encounters it. Then I would follow the journey from discovery to product selection and policy review.
At every stage, I would ask what the person still needs to know. Does a short description introduce a promise that the product page does not explain? Does a policy apply to the selected item? Is a measurement attached to the right variant?
I would keep missing evidence separate from missing navigation. If the source exists but is hard to find, the work may be a clearer label or contextual link. If the source does not establish the claim, the work is verification or revision.
That distinction prevents a content project from becoming a campaign to make unsupported claims more visible.
Reviews can contribute evidence without becoming a script
The sampled product pages used reviews differently. Some included ratings and comments; the smallest brand’s sampled product page did not show reviews in the reviewed material.
I would not infer from that observation that the business has no satisfied customers or no customer feedback elsewhere. It describes what appeared in the sample.
For implementation, I would look for genuine accounts that help a future shopper understand use, fit or limitations. A useful review can describe the conditions of one experience without being promoted into a universal product claim.
I would also avoid asking customers to repeat the attributes a brand wants AI systems to associate with it. The point is to collect honest feedback, preserve its context and identify recurring issues—not manufacture a vocabulary of approval.
The review can then inform both content and operations. Repeated confusion about a size label may require a clearer guide. A complaint about return eligibility may point to a policy explanation or a process problem.
How I would apply this to hotels and tours
This is a proposed application, not a travel result from the apparel review.
A tour listing may describe flexibility while the conditions for changing an itinerary sit elsewhere. A hotel page may advertise family suitability without making the relevant room occupancy clear. A marketplace might explain its supplier model on an About page while leaving the booking page ambiguous about who confirms arrangements.
I would bring the evidence needed for the decision into the relevant page, with enough context to avoid overstating it. That might mean a precise physical-demand description, an explicit cancellation boundary or a link to the applicable room terms.
A generic trust banner cannot answer those questions. Neither can a long company story. The evidence has to match the uncertainty the traveller is trying to resolve.
Where a fact changes frequently, I would assign an operational owner and a review trigger. Better placement does not help if the information is stale.
Education adds a second decision context
My international-education work at IMPACT gives me another context for asking where evidence belongs. That experience is separate from this apparel review; I did not run the same 16-criterion study on school websites.
A parent comparing schools may need to distinguish the curriculum offered, the age or grade it applies to, admissions requirements and the available route to an enquiry or visit. A reassuring statement about support does not establish how a particular child's needs will be assessed.
I would place verified curriculum and pathway information near the relevant comparison, and the applicable entry conditions near the admissions decision. If an external authorization supports a claim, I would make its scope clear rather than imply it covers every programme or age group.
An admissions statement also needs an owner and a date. Entry-year availability, assessment steps and deadlines can change. A general marketing page should not be treated as a guarantee of a place.
This is the same evidence-placement question from a different professional setting: what does the person need to verify at this stage, and which source can actually establish it? It does not require copying the entire institutional story onto every page.
My IMPACT case study describes the executed international SEO work. The ISKL campaign concept is a separate proposed journey with landing-page mockups, not evidence of an implemented admissions campaign or additional results.
What I would measure after making a change
This study did not test conversion improvements or citation gains. If I changed a page based on the review, I would design a separate evaluation.
For the user experience, I might test whether people can find and correctly explain the relevant condition, and record the task and success criteria in advance. For the content, I would verify that the rendered claim matches its source and that the scope is consistent across surfaces.
If AI representation were part of the objective, I would use repeated prompts and check the resulting claims as well as links. A new citation is not enough if the answer drops the qualification I added.
I would report those outcomes separately. Better findability, fewer source contradictions and a changed citation pattern are different results, and each deserves an appropriate interpretation.
What the review changed in my approach
The most useful output was not a verdict on which brand had the most trust. It was a more specific way to ask where evidence belongs.
I would still review company identity, expertise, policies and external support. But I would connect each observation to a decision, a claim and a supporting source. That makes the resulting task easier to assign and easier to verify.
Proof can be present and still fail to answer the question in front of the shopper. Before commissioning another generic trust section, I would examine that gap.
Read Does Size Buy Trust? for the original sample and checklist. For maintaining the facts behind these claims, see my brand-fact governance article. The companion practical resources are available in the workbook library.
