Google documents the question_and_answer attribute for Merchant Center: up to 30 pairs of questions and answers per product, 1,000 characters each, 10,000 characters total, submittable in every country through the primary feed, a supplemental feed, or the Merchant API. It belongs to a family of six attributes known as conversational attributes, built for AI-driven surfaces such as AI Mode. Adding it has no impact on the approval status of products already live.
Contents
Google documents a question_and_answer attribute in the Merchant Center product data specification: it lets merchants add up to 30 questions and answers per product, within a total limit of 10,000 characters, usable in every country.1 It belongs to a family of six optional attributes known as conversational attributes, designed to complement the primary product data and help customers find information on AI-driven surfaces such as AI Mode, while also enhancing traditional search.2 That family was announced at Google Marketing Live in May 2026, after Google first teased it on X in January with a promise of dozens of new attributes for the conversational commerce era.3 This article covers the format of the question_and_answer attribute, the rules for filling it in, and a method for picking questions from what a store already has.
01The question_and_answer attribute and its family
The attribute is meant for frequently asked questions about a product, authored by users, merchants, or manufacturers, to guide research and buying decisions.1 It is optional for each product.1
Format and limits
The attribute is a group attribute with two sub-attributes, repeatable up to 30 times per product.1
| Sub-attribute | Limit |
|---|---|
Question [question] (required) | string, up to 1,000 characters |
Answer [answer] (required) | string, up to 1,000 characters |
| Repetition | up to 30 pairs per product |
| Total per product | 10,000 characters maximum, all pairs combined |
The attribute has no Schema.org property equivalent.1 A product page can separately carry its own visible FAQPage markup, like the FAQ schema covered for informational content: the two mechanisms sit side by side, one in the Merchant Center feed, the other in the page's code.
A family of six conversational attributes
Question and answer [question_and_answer] shares its documentation with five other attributes: document link [document_link], related product [related_product], item group title [item_group_title], variant option [variant_option], and popularity rank [popularity_rank].2 All are optional, built to complement a product listing rather than replace it, and can be added via a supplemental data source (Google's recommendation), the primary data source, or the Merchant API.2 Submitting them has no impact on the approval status of products already live.2
This data layer adds to the channels already open to shopping agents on the merchant side, such as the Merchant Center hub of the Universal Commerce Protocol, which covers the transaction rather than the product listing itself.
02What belongs in the attribute, and what doesn't
Google sets minimum requirements so the attribute stays useful and doesn't duplicate the rest of the product listing.1
Google's minimum requirements
Both the question and answer sub-attributes must be submitted together. Content should cover only the product: no time-related information such as price or dates, which belongs with other attributes. Google asks merchants to avoid listing keywords or search terms, and to give correct, informative answers, since shoppers rely on this data to make buying decisions.1
Don't duplicate other attributes
Google advises against repeating information already submitted in the title, description, product detail, or product highlight attributes.1 If a [document_link] PDF is also submitted and contains the same information, the matching question-and-answer entry should be left out: Google will extract that information from the document instead.1 Avoiding the same data in several fields of a product listing ties into a broader principle for product page work on an online store, where clean markup often counts for more than a long text block.
The comma trap in a text feed
In a TSV or CSV file, colons, commas, and quotes inside a question or answer must be escaped to match the feed's format, or a single value can be split into several pieces.1 Google recommends TSV over CSV to limit that risk, or escaping characters with a backslash in Google Sheets.1 Google's own example packs two pairs into one cell: "Does it have a headphone jack?":"This version doesn't have a headphone jack.", "Does it support Bluetooth?":"It has full Bluetooth 6.0 support."12
03Picking questions from what already exists
The 30 available slots fill in better from what a store already has than from a list written in one sitting.
Start from what's already been asked
Three sources surface questions real buyers actually ask: customer support or live chat threads, questions left under reviews or on a marketplace listing, and any FAQ already published on the product page. A question that keeps coming back in those exchanges deserves a slot before a hypothetical one.
Prioritize compatibility, size, and upkeep
For a physical product, the questions that most often block a buying decision are about compatibility with other equipment, dimensions and fit, and upkeep or how long it lasts. Those are also the hardest to guess from a short title or description, so the attribute adds the most value there.