# Product Structured Data: Prices Availability and Variant Consistency By AI Visibility editorial team AI-assisted article. A shopper opens a link for a large tote. The selected size is Large, the page shows AED 99 and the buy button is disabled because it has sold out. The Product markup still reports AED 79 and in stock. Those values belong to the small tote. Every field can look reasonable when checked on its own. The error becomes obvious once you compare the selected SKU, the visible offer and the structured offer together. Use this audit when two or more variants share a template and can differ in price or availability. The existing AI Visibility product-information guide explains which fields a UAE store should publish and how variants are grouped. This walkthrough starts after those fields have been chosen. It shows how to detect the copied-offer error by opening each variant URL directly and checking the values together. AI Visibility product-information guide (/article/what-product-information-should-you-publish-ai-search-for-uae-e-commerce/) Start with the variant a customer can buy Build a small sample from the catalogue. Choose a default variant, a variant at a different price and one that cannot currently be bought. Two SKUs can cover all three cases if the default is cheaper and in stock and the other is sold out. If every item in the test has the same price and stock status, the wrong offer can pass a superficial comparison. Keep the variant’s expected option, item identifier, displayed currency, current price and stock state together in the record. A ProductGroup represents the family of closely related products. Use the group to hold genuinely shared characteristics and identify the options that distinguish the variants. The Product describes the specific option or combination, and the Offer describes its commercial terms. Google’s variant guidance documents ProductGroup, hasVariant, variesBy and productGroupID for grouping. Its examples cover a single page with directly selectable options and variants presented on different pages. Choose the pattern that fits how buyers reach the item. Schema.org ProductGroup (https://schema.org/ProductGroup) · Google product variant guidance (https://developers.google.com/search/docs/appearance/structured-data/product-variants) Field What to compare Error to look for Variant URL Open it in a fresh session and note the selected option The page opens with the default size Item identifier Compare the selected SKU with the Product node The parent group ID stands in for the child SKU Offer price and currency Compare the amount and currency on the same variant A cheaper sibling’s price or another market’s currency appears Availability Read the visible message and check the purchase control A sold-out option inherits in-stock status Offer destination Check where the markup’s URL resolves The offer sends readers to a different option If shoppers see a price range before choosing a size, record that as a group-level display state. Then open the URL for an individual option and compare its own price. A lower figure elsewhere on the page can describe the cheapest variant while the selected variant costs more. Keep both observations in context so the person fixing the template can see which price belongs in the variant offer. Open each variant URL as a shopper would Start with a fresh browser session. Paste the full URL for the non-default variant and record the final address after redirects, the selected size or colour and the visible SKU when present. Read the displayed price and stock message. Check whether the buy button is available in that state. Compare the rendered JSON-LD against the Product node for that exact item. If the page carries several Product nodes, use the SKU or another stable variant key to find the matching one. Reading the first Offer in the document can give you the wrong result. Then choose a different option using the visible controls. Watch the selected option, price, stock message and address. If selecting a variant updates the URL, copy it into a separate session and confirm it restores the same item. Google documents a single-page example where the page contains structured data for multiple variants and the markup stays in place while the visible selection changes. In that design, the important comparison is between the selected item and the correct Product and Offer in the group. Google product variant guidance (https://developers.google.com/search/docs/appearance/structured-data/product-variants) If a non-default address redirects to the group URL and selects the first option, inspect the redirect rules and the variant selector before editing the structured data. If the right option appears on the page but its Offer belongs to a sibling, trace how the markup gets its values. The visible component may read the selected SKU while the server template uses a default product object. Capture the original response and rendered state if they differ. That comparison tells the implementer where to reproduce the fault. Check sale prices and stock at the same moment Capture the active price and currency for the chosen item at a known time. A regular price, an active sale, a member offer and a conditional discount can be shown together. Keep the wording that explains each figure and inspect the markup field representing the active offer. Google’s merchant listing documentation describes supported price and availability fields and more specific pricing cases. Match your implementation to the type of offer actually shown. Google merchant listing structured data (https://developers.google.com/search/docs/appearance/structured-data/merchant-listing) Watch for the moment when a sale starts or ends. The storefront may switch price immediately while cached markup or a product feed updates later. Save the relevant snapshots with their timestamps and check the catalogue record for the same SKU and market. If the base and sale prices have switched places, reproduce the issue before and after the transition. If only one customer segment qualifies for a discount, record what an unsigned visitor sees separately from the member price. Likewise, a product family can include both purchasable and unavailable options. The selected variant’s visible message and purchase controls can provide separate signals. When an option is described as sold out yet its button appears enabled, record the conflict and ask the commerce owner to confirm the intended status. Preorders and backorders have their own meanings and may have different purchase paths. Avoid flattening them into a single stock state just to fit a template. Google Merchant Center availability guidance (https://support.google.com/merchants/answer/6324448?hl=en) A previously verified official Brewing Gadgets Instagram post illustrates the scope of a retail promotion. Its caption applied the offer to selected Compak grinders in the UAE and specified an end date. When a promotion uses that kind of wording, check which exact products and variants qualify before applying its price in markup. Keep the promotion and each variant’s stock status as separate pieces of evidence. Official Brewing Gadgets Instagram post (https://www.instagram.com/brewing_gadgets/p/DfWOw78j2CY/) View the selected official Instagram post (https://www.instagram.com/brewing_gadgets/p/DfWOw78j2CY/) Brewing Gadgets’ dated UAE promotion on selected Compak grinders. Reproduce a copied offer with two controlled variants On 6 November 2026, a local Chrome test loaded simulated Small and Large tote URLs. The browser intercepted those fictional URLs and served a controlled page for each. Both pages contained a ProductGroup with two Product variants. The visible pages selected the correct size and showed each variant’s own SKU, price and stock status. Small displayed AED 79.00 and in stock. Large displayed AED 99.00 and sold out, with the add-to-bag control disabled. The first fixture gave both variant records the Small offer. Chrome opened both variant URLs and the checker looked up each matching Product by the visible SKU. It found three mismatches for Large. The offer URL targeted Small, the price was AED 79.00 instead of AED 99.00, and availability was in stock despite the sold-out message. Small passed because its own values happened to be the template defaults. After attaching the correct offer to each variant, the same checks passed on both URLs. Here is the corrected Large offer from the fixture. The domain, item and amount are fictional. { "@type": "Offer", "url": "https://shop.example/tote?size=large", "price": "99.00", "priceCurrency": "AED", "availability": "https://schema.org/OutOfStock" } The test parsed the rendered JSON-LD and compared selected values with the page. It did not emulate Google’s validation or submit a feed. It also left the purchase control untouched. The result shows how a comparison detects a mismatch in a controlled example. Use a separate platform test to learn which search features recognise the complete markup and which requirements still apply. Sign off the URL, markup and feed as one comparison Make a Variant offer record with one row per tested item. Include the requested and final URL, selected variant, SKU, visible price and currency, stock message, button state and the values from its matched Offer. Capture when each observation was made. Add the current feed row if the store uses Merchant Center. Merchant Center’s guidance distinguishes the item identifier from the shared item group identifier and expects the landing page to present the relevant variant. Use both identifiers to trace the route from the catalogue to the submitted item. Google Merchant Center item grouping (https://support.google.com/merchants/answer/6324507?hl=en) · Google Merchant Center landing page requirements (https://support.google.com/merchants/answer/4752265?hl=en) Diagnose the first point where the data diverges. A wrong price in the catalogue needs a different fix from a correct catalogue price paired with a stale cached response. A wrong default on direct landing needs routing or selection work. A group-level identifier submitted as an item identifier needs a feed correction. Ask the owner to regenerate the affected output and replay the same two-variant comparison, including the variant that previously passed. When you test a real store, add a second case that differs from the first in how customers reach it. A direct product URL can work while a category link loses the variant parameter. A feed URL can carry the right option while a redirect silently drops it. Open the exact URL each source uses. Save any redirects and confirm the final selection, amount and currency in a signed-out session for the intended market. Where the store serves more than one market, keep each market’s run separate so values from different storefronts cannot be mistaken for a mismatch. Run the relevant structured-data validation after the comparison. Record which properties were recognised and which eligibility checks were reported. Search appearance, indexing and feed approval are separate outcomes. Keep the original test tied to the customer’s task. The buyer should land on the promised option and see the same terms that the site emits for that item. Google structured data policies (https://developers.google.com/search/docs/appearance/structured-data/sd-policies) For an inventory-driven store, repeat the comparison when the last sellable unit is taken out of stock in a test environment. Record when the catalogue changes, when the page begins showing the new state and when the outgoing data reflects it. A short delay can look like a permanent mapping error in a single capture. A sequence of timestamps helps the team decide which output needs invalidation or another refresh. Keep at least one failed case from the audit as a release check. Give the reviewer two variant links and the expected values for both. Check them after changes to the catalogue template, currency display or stock integration. If the store changes how it identifies a variant, update the test and confirm that the customer-facing link and the structured offer still refer to the same item. AI Visibility is published by Lunasol, which offers GEO and AEO services. The browser test used fictional variants and local responses. About AI Visibility (https://aivisibility.ae/about/)