Product page errors cost pet brands in returns, one-star reviews, suppressed conversion and lost AI recommendations. Measured across UK pet retail, 96.5% of matched products have an ingredients issue, 99.2% a nutrition-table issue and 100% a feeding-guide issue somewhere they're sold — a meaningful share of every brand's digital shelf is describing something it doesn't sell.
We took real UK pet listings, matched them by EAN — the same physical product, on different retailers' shelves — and compared what each page claims, field by field. The disagreement isn't at the margins. It's on the fields a pet owner actually buys on.
Eagre analysis of 6,000+ EAN-matched products across 14 UK pet retailers, July 2026. An issue means the field is missing from at least one listing, or word-for-word mismatched between listings after formatting normalisation; sister storefronts of the same retail group are excluded.
These aren't different marketing treatments of the same facts. Ingredients, analytical constituents and feeding guides are matters of record — and the record disagrees with itself almost everywhere.
We went looking for one product whose factual content was complete and consistent everywhere it's sold. Of 2,934 products matched across two or more genuinely different UK retailers, we didn't find one. The average product's listings differ on nearly four of five content fields somewhere in the market, and 85% of products have at least one listing missing a factual field outright.
Drift isn't even stopped by a shared owner: two storefronts run by the same retail group disagree on half of their shared listings. And on several major UK pet retail sites, product pages expose no machine-readable barcode at all — there is no programmatic way to verify which product a page is showing, let alone whether its content is right.
Even under deliberately tolerant matching — forgiving formatting and near-matches — roughly one product in three still disagrees with itself on what's actually in the product. The figures above are the word-for-word reading; the problem survives any reasonable definition.
A wrong flavour on a dog food listing isn't a cosmetic slip. Follow one error through, with a fictional but entirely typical example.
Say PawNourish sells a chicken recipe, and one retailer's page lists it as beef — a re-keying slip, nothing more. A shopper whose dog reacts badly to chicken reads "beef" and buys it. Best case, the dog refuses the bowl and the bag comes back as a return the brand ends up paying for one way or another. Worst case, the dog has an allergic reaction — and the owner, reasonably, blames the brand whose name is on the bag, not the retailer whose CMS holds the error.
Then comes the review. "Listed as beef, arrived as chicken — my dog was ill" sits on that product page long after the field is corrected. The fix takes a day; the one-star review outlives it by years, quietly taxing the conversion of every shopper who reads it.
And the error compounds before a human ever sees it. On-site search indexes the wrong flavour, so the product surfaces for the wrong queries and misses the right ones. AI shopping assistants read the same wrong field and confidently steer shoppers with chicken-sensitive dogs straight to it — or steer everyone else away. One re-keyed word, four separate costs: the return, the review, the suppressed conversion, the misdirected recommendation.
The pet-specific numbers above are Eagre's own measurement. The behavioural consequences are documented well beyond pet.
Put the two together: virtually every product's factual content is wrong or missing somewhere it's sold, seven in ten shoppers have returned an item over incorrect content, and the discovery channel growing fastest makes its recommendations from exactly the fields that disagree.
Because the process that creates the errors is still the process. Every retailer re-keys the same product datasheet by hand into its own systems — its own templates, its own field names, its own character limits, its own timing. Every re-key is a fresh chance to drop a constituent, transpose a feeding guide or paste last year's description.
Then nobody owns the drift. The brand assumes the retailer keyed what was sent; the retailer assumes the brand would flag anything wrong; the shopper assumes somebody checked. When the brand updates a recipe, some retailers re-key it that week, some that quarter, and some never — so even listings that started identical diverge over time.
Monitoring tools can report the problem — a dashboard that tells you your listings disagree. But reporting isn't repair: the fixing stays manual, one email and one retailer portal at a time, and the drift re-accumulates as fast as it's chased down. The gap persists because the pipe is broken, not because nobody has noticed the puddle.
Ask for a drift review and we'll show you where your own listings diverge today — a concrete starting point for weighing up what the gap is costing you.