The next generation of shopping extensions needs to earn consumers’ trust back

For a decade the default advice to anyone buying online was to install a shopping extension. The category earned that advice, then spent it. Coupon injection, last-click affiliate grabs, and a “savings found” animation that did not always match the till have left shoppers — and the creators who promoted the tools — with a reasonable question: whose side is this thing on?

The next generation of shopping extensions will not win by being louder about discounts. It will win by being boringly checkable. That is a product requirement, not a brand line.

The job is the price, not the coupon

A coupon trims the price of the item where you already are. It does nothing about whether that item is the right price in the first place. The spread on identical products across UK retailers is routinely larger than a typical checkout code. An extension that only hunts codes is optimising the last five percent and ignoring the rest.

The useful job is earlier: identify the product on the page, show what the same item costs elsewhere in the same currency, and show whether today’s price is actually low against a history the shopper can see. If nothing is cheaper, say so. If the history is too thin to judge a “deal”, say that too. Silence is a feature.

Five principles that should be visible in the product

1. No coupon injection. Codes are someone else’s product. Stack them at checkout if you want; do not bury them in a comparison tool that is supposed to be about the price.
2. No inbound-cookie override. If a creator or a cashback site brought the shopper, the extension does not steal the last click. That is not a slogan. It is a line of code a journalist can verify in the shipped extension.
3. Rank by price, never by commission. Affiliate funding is fine when it is disclosed and when it cannot move the order of the results.
4. Same-currency comparisons only. A “saving” that appears because the row is in another currency is not a saving.
5. A published evidence bar for calling a discount real. Sampled history cannot prove a price never existed. It can prove a price did. Claims should follow that asymmetry.

Trust is earned in the miss, not the hit

The Honey era trained people to expect theatre: a badge, a percentage, a sense that the tool is always working. The honest product will often show nothing. A used listing at a third of the page price is not a deal. A barcode match to a refurbished unit is not “the same product” in the way a shopper means it. An empty result is better than a confident wrong one — for the shopper, and for anyone who has to defend the category in print.

That is also why shopping extensions should stop leading with “AI”. Every catalogue already uses models somewhere. The consumer proposition is simpler: before you buy, check whether the price is real. The model is the how, not the headline.

What the trade press should demand

Retail Gazette, InternetRetailing, UKTN and the rest do not owe any extension a launch write-up. They do owe their readers a test: can a non-engineer verify the claims in ten minutes? Permissions that match the privacy policy. A public methodology for “was that sale real?”. Affiliate disclosure that is not buried. A named human who will answer on deadline when the CMA next names a retailer.

WEM is a UK price-comparison service I founded. We make money from disclosed affiliate commission; checkout is on the retailer; the extension injects no coupons and does not override an inbound affiliate cookie. Those sentences are true only because the product was built that way. They are also the floor I think the next generation of this category has to stand on — whether or not the logo on the toolbar is ours.