What UK shoppers actually check before buying (and what they don’t)

If you ask a UK shopper how they decide whether an online store is trustworthy, you get a reasonable-sounding list. They check the reviews. They look for a padlock on the address bar. They check the returns policy. They might mention Companies House. A confident minority will say they Google the brand name.

None of that describes what most of them actually do. And the gap between the stated version and the observed version is the most important thing a merchant can understand about how trust signals work.

The stated versus revealed preference problem

Behavioural economics has a long-running distinction between what people say they will do and what they actually do. On trust checking, the gap is unusually wide. Surveys consistently report high rates of active trust checking. Eye-tracking, session recordings and usability observation consistently report that active checking is rare, quick, and concentrated in a very small number of places on a page.

This is not shoppers lying. It is shoppers describing an idealised version of themselves. Under any real time pressure, and with any prior familiarity with the merchant, the checking they perform is far lighter than the checking they think they perform. The signals that decide the sale are not the ones they can name in a focus group. They are the ones that catch the eye during a two- to five-second scan.

What shoppers say they check

The consistent survey picture is that UK shoppers report checking, in order of stated frequency, the reviews on the product, the reviews on the merchant more generally, the returns policy, the security of the payment process, whether the site looks professional, and, further down, the actual identity or company details of the seller.

These are all plausible checks. The problem is that “checking” for most of them means glancing at a signal, not actively investigating one. Almost nobody clicks through to read a full returns policy. Almost nobody scrolls to the footer to hunt for a company number. Very few visit Companies House. The stated version implies deliberate verification. The revealed version is a set of quick scans looking for signals that verification could be done, not doing it.

What they actually check

The behavioural literature on ecommerce scanning is dominated by three findings that are consistent across studies.

First, attention on a product page is heavily weighted to the top-left region, following the well-documented F-shaped scanning pattern first described by Nielsen. The star rating and price receive strong initial attention. Elements more than one screen down receive a small fraction of the attention of elements above the fold.

Second, footers are essentially unread during purchase-intent visits. Baymard’s aggregated eye-tracking work suggests footer engagement during a shopping session sits well under 5%. This matters because footers are where most stores put their trust badges, company registration numbers, and “About us” links. Any signal that only lives in the footer is doing almost no work. The signals that earn their place are the ones held in the viewport by design: the header band, a fixed-position seal, or the area around the buy button.

Third, the checks that do happen are proxy checks. Shoppers do not read the returns policy, but they scan for whether one visibly exists. They do not verify the seller’s identity, but they scan for whether an identity signal visibly exists. Presence of a signal is doing most of the trust work, not the content of the signal. This is not lazy checking. It is efficient checking, and it is what human beings do when the information asymmetry is high and the time pressure is real.

Where attention actually lands

The proxy-checking pattern

The single most useful concept for understanding modern trust checking is the proxy. A shopper does not verify each claim on a site. They scan for signals that a serious business would have, and treat the presence of those signals as a good-enough substitute for verifying them individually.

A visible reviews rating with a real count is a proxy for “this site actually sells things”. A visible returns policy link is a proxy for “I could get my money back if I needed to”. A visible trading identity or verification badge is a proxy for “there is a real company behind this”. What matters, for each of these, is that the signal is visible in the two-to-five-second scan window. Not that the shopper reads or clicks or verifies.

This has an uncomfortable implication for merchants. It means the seller-identity work is being done by whether an identity signal is visible, far more than by whether the shopper actually investigates. Sites with a strong identity signal near the buy button will earn trust from shoppers who would not, under any circumstances, actually click through to verify it. Sites without one will lose trust from those same shoppers for the same reason.

Why this makes verification signals so leveraged

If shoppers actually pursued the checks they claim to pursue, the value of a verification signal would be modest. It would just save them a click. But shoppers do not pursue those checks. What they do is scan for the presence of signals that make the pursuit unnecessary.

This is why the leverage of a well-placed verification signal is higher than the shopper survey data would suggest. It is doing the work of a check that will almost never actually be performed. The signal is not there to help the diligent shopper verify. It is there to reassure the far larger group of non-diligent shoppers that the diligent version could, if they wanted to.

What this means for how trust signals should be structured

The behavioural evidence collapses into three practical rules.

1. Move trust signals up. Footer placement is almost always wasted. The header, the product-title band, and the area immediately around the buy button are where attention actually lands.
2. Prioritise scannable presence over readable depth. Long trust copy in a paragraph is rarely read. A visible badge with a clear one-line label does more work than three sentences of explanation.
3. Serve the proxy check. The signal is the reassurance. Design it so a shopper who never intended to verify can see, in two seconds, that verification exists and could be pursued.

Stated versus revealed is the reason merchants can spend a year commissioning better returns copy or improving their About page and see almost no lift in conversion. The shoppers were not reading the returns copy. They were not reading the About page. They were scanning for whether either visibly existed. The lift comes from placing that visibility in the small number of regions where the eye actually goes.