Shopping
Part of Shopping campaign operations
Preparing a product feed for paid shopping traffic
Prepare a Shopping feed with stable product IDs, accurate offers, Australian shipping details and a practical upload check.
Prepare a Shopping feed by making each product record an accurate, purchasable offer. Give Merchant Center stable product IDs, useful descriptions, and working image and landing-page URLs. Check that price, availability and shipping match what an Australian customer sees.
Work through the feed in order: assemble product data, compare offers with the live site, submit the data, then check it again after Merchant Center processing. Correct mismatches at the source and recheck the affected items.
Build from dependable product data
Before exporting, decide which system owns each fact. A product information system may own names and variants; a commerce platform owns current price and stock. Resolve conflicting values at their source, because a correctly formatted feed can still carry stale information.
Choose a product data method suited to your catalogue and how often it changes, then prepare a product file for Merchant Center. Use Google’s current Product data specification to check required and conditional attributes; do not assume a template covers every product.
Give each sellable item a unique, stable id. Its title and description should identify the actual product and relevant variant without promotional filler. The link should lead to that product’s page, and image_link should point to an accurate main image Google can crawl.
For new, in-stock brand-name products with a manufacturer-assigned GTIN, provide the correct GTIN and corresponding brand for Australian customers. GTIN means Global Trade Item Number; it identifies a specific product version and is commonly found next to the product barcode.
For used, custom, handmade or vintage products without a product code, the identifier guidance allows the GTIN field to be set to identifier exist=false. Check identifiers and variant fields where they apply. Unnecessary ID changes also make later sales reconciliation harder.
Reconcile the customer offer
Sample ordinary items, variants, sale items, out-of-stock items and products with unusual delivery charges. Compare the submitted data with the exact product page, structured data where used, and checkout:
| Field | Check |
|---|---|
| Price | The correct product and variant, current amount and currency, including applicable GST. |
| Availability | The stated stock status agrees with the page and ability to order. |
| Shipping | Merchant Center settings or product overrides reflect the relevant charge and delivery area. |
| Landing page | The advertised item can be found without a misleading redirect. |
| Image | The main image accurately shows the offer and has no promotional overlay or generic substitute image. |
For a sale or variant, check the exact offer rather than only its parent product. Check that the shipping information submitted to Merchant Center reflects the relevant charge and delivery area.
Key fields to verify between feed and live site
- PriceMust match current amount in AUD, include GST, and reflect correct variant pricing.
- AvailabilityStock status must align with live site and ability to order (e.g., 'in stock', 'out of stock').
- ShippingCharges and delivery areas must reflect actual Australian shipping rules and Merchant Center settings.
- Landing pageDirect link to the exact product page without redirects or misleading content.
- ImageMain image must show the actual product with no promotional overlays or generic substitutes.
Check the processed data
Before upload, check file formatting and URL encoding. After Merchant Center processes the data, review its diagnostics for issues affecting your product data.
Check the processed data against the sampled live offers, including price, availability and shipping. Recheck the exact product page and checkout so the submitted offer still matches what Australian customers can see.
Correct the catalogue, site or mapping that owns the error, refresh the data source and check the affected items again. Record who owns the source and when it changed so a later mismatch can be traced.
