Make pack size and price units clear on product pages
State what one purchase quantity buys: a single item, a pack or another selling unit. Then check that the title, selected variant, image and displayed price describe those same contents.
A shopping assistant can repeat an ambiguity already present in the catalogue. Resolve it at the source before adjusting the exported wording.
Put the selling unit into a sentence a buyer can check
Consider a fictional product page headed “Coffee filters, 100”. Does the price buy one box containing 100 filters, 100 boxes or a pack containing several smaller bags? A person may infer the answer from the photograph. An assistant comparing extracted titles and prices may receive less context.
Write the offer explicitly: “One box containing 100 filters”, if that is what the confirmed catalogue says. Keep the sentence consistent with the selected variant and the structured representation. A clearer sentence cannot repair an incorrect pack count in the underlying data.
The OpenAI product-feed contract describes the fields accepted by that destination. Use those definitions when mapping the offer; do not create a field name merely because it would make a convenient example. The buyer-facing description and the destination's machine-readable contract have different jobs, but they should describe the same purchase.
| Question | Example of an unambiguous answer in this fictional case |
|---|---|
| What is sold? | One box |
| What is inside it? | 100 filters |
| What does the displayed amount buy? | The complete box |
| Is another pack selected? | The selected option must state its own contents |
| Is a comparison price shown? | Its denominator must be stated separately |
Avoid inferring legal unit-pricing requirements from this table. It is a content review method; applicable rules depend on the product and market.
Keep arithmetic and commercial meaning together
Suppose a synthetic offer sells a pack of six items for 24 euros. Dividing 24 by six gives four euros per item. That calculation is valid only if the six items are the relevant equal units and the 24-euro amount actually buys the whole pack. It does not prove what taxes, delivery charges or conditions apply.
If a page also offers a pack of twelve, the selected option must carry its own price and contents. Do not let a general paragraph describing six items remain beside a twelve-item selection without explaining the difference. The ambiguity can survive even when each individual field contains a plausible value.
During review, ask someone to restate the purchase using only the information available in the selected page state. If their answer differs from the catalogue, locate the misleading field. This is more diagnostic than asking whether the page “looks clear”.
Follow the same unit across representations
Check the title, description, option labels, feed and structured data. A bulk find-and-replace can change the prose while leaving the exported count untouched. A template can also reuse a parent description after the variant changes.
The correction should therefore start at the first source of ambiguity and travel through the representations that depend on it. Preserve the affected item as a test case. On the next catalogue change, verify the complete selling-unit sentence again, including the selected option, rather than merely searching the page for the number 100.
Compare title, image and description
A photograph may show accessories that are not included. A title may describe the product family while the selected variant changes the pack size.
State the contents in the page's main product information and check the selected variant. Do not rely on a distant FAQ to reverse the impression created by the title.
Include quantity misunderstandings when investigating avoidable returns, a topic covered in our French dossier. Verify the actual cause in the order evidence before attributing a return to the page.
Check the destination contract
Different feeds and structured-data formats represent product information differently. Consult the OpenAI product catalogue specification or Google's product documentation according to the destination.
Do not invent a field or place an unsupported meaning into a convenient text property. If a destination lacks a suitable representation, make the limitation explicit and preserve clear buyer-facing wording.
A mapping document should identify the source for each quantity fact and the exact field or passage carrying it.
Test a buyer question
Ask a reviewer to determine what arrives, how many units the price covers and which variant is selected using only the public page. Then compare their answer with the confirmed catalogue facts.
If an assistant is part of the test, preserve its question, response and cited source. A correct answer in one trial does not establish reliable behaviour across all systems.
The Verity Score product-markup guide offers adjacent context on representing product facts. It cannot establish the contents of your own package; that requires your authoritative product data.
