Merchants keep treating the product feed as a formality to clear before launch, but the checkout research on field counts and naming shows how small a mismatch between what a form asks for and what a shopper expects can cost, and that same gap is what breaks feed submissions.
Most merchants treat a rejected Merchant Center feed as a data problem, something to patch with a bulk find-and-replace and resubmit. The more useful way to see it is as the same failure that shows up in checkout forms: fields that exist on paper but don’t match how information actually flows in practice. Checkout research from Baymard puts a number on this for the purchase flow, 11.3 fields on the average form and 17% of shoppers abandoning specifically because of that complexity. A product feed is a form too, just one filled out by a merchant instead of a shopper, and it fails for the same structural reasons.
Baymard’s testing on checkout forms found that 89% of sites in its e-commerce UX benchmark still split a single name into two separate fields, first and last. That split looks harmless on a spec sheet. In practice, 42% of participants typed their full name into the first-name box at least once, meaning the field did not match how people actually think about their own name. Compare that to sites using one combined name field, where only 4% of users showed any hesitation before typing. The field count was lower and the friction was lower too.
A product feed has the same trap. A field can be technically filled in, present, non-empty, correctly typed as a string or number, and still be the wrong shape for what the matching system expects. A price field with a currency symbol embedded in the string, a GTIN copied from the wrong variant, a description field padded with keywords that don’t describe the item, all of these pass a basic “is this field empty” check while still causing a rejection somewhere downstream. The problem isn’t missing data. It’s data shaped for the wrong expectation, the same mismatch Baymard documents at the checkout level.
It matters to be precise about what these figures cover, because none of the checkout research or the USPS rate sheet in the source material describes Google Merchant Center directly. The 11.3 average field count and the 17% abandonment figure come from checkout flow testing, not from feed audits. The 89% figure, the 42% figure, and the 4% figure all come from how shoppers behave when typing a name into a form, not from how Google’s matching algorithm parses a title attribute.
What transfers is the underlying lesson, not the numbers themselves: unnecessary fields and mismatched field structures cost something measurable, whether that’s a completed purchase or an approved feed line item. The counts rest on Baymard’s published checkout usability testing across its tracked site benchmark, and on USPS Notice 123 as loaded September 9, 2026, for the postal rate figures used elsewhere on this site. Neither document mentions Merchant Center by name, so any application of these numbers to feed design is an analogy, not a direct measurement, and should be read that way.
| Measured behavior | Figure | Source |
|---|---|---|
| Average checkout form field count | 11.3 fields | Baymard checkout research |
| Users abandoning due to checkout complexity | 17% | Baymard checkout research |
| Sites with more than one name field | 89% | Baymard e-commerce UX benchmark |
| Users typing full name into first-name field | 42% | Baymard checkout research |
| Hesitation rate with a single combined name field | 4% | Baymard checkout research |
| USPS authorized publication rate | 5% | USPS Notice 123 |
| USPS nonadvertising adjustment factor | 1% | USPS Notice 123 |
| USPS single-issue purchase price | 90% | USPS Notice 123 |
The postal figures sit in this table because they come from the same fact set gathered for this research cycle, not because they describe feed behavior. They are included for completeness and should not be read as related to Merchant Center field requirements.
Neither the checkout data nor the postal rate sheet gives a merchant a numeric Merchant Center disapproval rate, so nobody should quote one that isn’t published on Google’s own current spec pages. What the checkout research does give is a repeatable pattern: fields that don’t match how the underlying information is actually structured cause friction that’s disproportionate to how small the field looks. A split name field seems minor and still causes 42% of people to type in the wrong box. A feed attribute formatted slightly off from spec seems minor too, and it can still block an entire product listing.
Check every field in a feed against the current published Merchant Center spec, not against habit or a template copied from a previous store. Where a field’s purpose is unclear, treat it the way Baymard’s testing treats a redundant name field, as a candidate for removal or simplification rather than a box to fill just because it’s there.
Google's own spec calls for an id, title, description, link, image_link, availability, price, and condition on every item, along with brand and GTIN or MPN where the product category expects them. None of those field counts are numbers this article can independently verify beyond what's published on Google's own spec pages, so treat the list as structural rather than numeric. The lesson from checkout research still applies here: extra fields that don't change how the item is matched or shown just add friction to the feed-building process, the same way extra checkout fields add friction at purchase.
A field being present is not the same as a field being usable, and that gap is exactly what the checkout research documents at the form level: 89% of sites in the Baymard e-commerce UX benchmark have more than one name field even though a single combined field causes hesitation in only 4% of users. Feed rejections often come from the same category of problem, values technically present but not matching what the field expects, such as a price formatted with letters or a GTIN that doesn't match the product type. There is no numeric disapproval-rate fact available here to cite, so the safest move is to check each field against Google's own current spec page rather than guess.
The checkout data says combining redundant fields helps: a single name field causes brief hesitation in only 4% of users, while splitting it causes 42% of people to put their full name in the wrong box at least once. Feeds don't let a merchant invent new combined fields, since Google defines the schema, but the underlying discipline transfers, meaning don't pad a feed with optional attributes that don't serve matching or display. Every field you populate is one more place a formatting mismatch can trigger a rejection.
Baymard's checkout research puts the number at 17% of users abandoning specifically because of checkout complexity, against an average of 11.3 form fields across the sites tested. That number describes checkout, not the merchant feed itself, but it is the closest measured figure available for how shoppers react to unnecessary data entry. A feed that generates disapprovals delays the same purchase path that a slow checkout interrupts, so the two problems compound rather than staying separate.
The facts available here come from USPS Notice 123, which covers postal rates and package dimensions rather than Merchant Center fees, so nothing in that document applies to feed submission costs. Notice 123 does list figures such as a 5% rate for authorized publications, a 1% nonadvertising adjustment factor, and a 90% single-issue purchase price, but those describe mail classification, not ecommerce data feeds. Anyone looking for Merchant Center's actual fee structure needs to check Google's own current pricing documentation rather than infer it from unrelated postal rate sheets.