Merchant Center Optimisation for Shopping Campaigns

Optimise Merchant Center in the right order

Merchant Center optimisation is not polishing every optional field. It is the work of increasing accurate, policy-compliant product coverage and keeping the data fresh enough that customers see the same offer after they click.

Use four layers: account trust and setup, product eligibility, offer accuracy, then enrichment and reporting. Moving straight to title rewrites while a shipping service is wrong or a data source is stale puts effort in the wrong place.

1. Measure commercially weighted coverage

In Products → Needs attention, export products that are limited or not approved by country and marketing method. Join item IDs to recent net revenue and strategic priority. Report both SKU coverage and revenue-weighted coverage.

A hypothetical merchant has 9,700 of 10,000 products approved: 97% SKU coverage. The 300 missing products include a flagship range that generated £80,000 of the last £500,000 in sales, so revenue-weighted coverage is only 84%. The second number determines urgency.

2. Fix account and website trust signals

Confirm the claimed and verified website, business name, address, contact routes and payment flow are current. Shipping and return information in Merchant Center must agree with the site. Test the full purchase path without an admin login and from each target market.

Use the exact issue shown in Merchant Center if there is a policy warning. Google’s issues guide distinguishes product-level and account-level problems. Do not reduce an account issue to one example product.

3. Make the data source observable

Record source method, owner, schedule, last successful processing time, target countries and transformations. Alert on missing files, item-count changes and parsing errors. A feed that quietly drops 20% of the catalogue can leave the account spending normally on the remainder.

Keep IDs stable. An ID should identify one product or sellable variant over time; changing it discards product history and can break campaign subdivisions. Never reuse an old ID for a different product.

4. Synchronise price and availability

Compare volatile products across source data, Merchant Center’s processed value, landing page, product structured data and checkout. For availability, use the supported value that reflects whether the exact submitted variant can be purchased. Google’s availability requirements also cover preorder and backorder display.

Enable automatic item updates where appropriate. They can update price, sale price, availability and condition from the site, but they are not a substitute for accurate scheduled data. If automatic corrections are frequent, shorten the source refresh or use an API-based workflow.

5. Add and test product structured data

Product-page structured data helps Google interpret current offer information. Google identifies price, priceCurrency, availability and condition as required schema.org values for automatic item updates. Its Merchant Center structured-data guide explains the mapping.

Test representative single products, sale products and variant pages in the Rich Results Test. Ensure the structured offer describes the same selected variant as the feed. A valid schema block with the parent product’s cheapest price can still create a mismatch for a more expensive submitted variant.

6. Improve identifiers and variant data

Submit manufacturer-assigned GTINs, brand and MPN accurately where applicable. Do not invent values. Group variants with a shared item_group_id, while giving every sellable variant its own ID and the relevant colour, size, material, pattern, age group or gender attributes where required.

Sample entire variant families rather than one parent. Check that the image, title, price, availability and landing-page selection all describe the submitted variant.

7. Improve titles and images by category

Build a category-specific title rule using reliable attributes. The title attribute accepts 1–150 characters and can be truncated, so front-load the product identity and important differentiators. Promotional text, repeated keywords and attention-grabbing capitals are not acceptable substitutes for product facts.

Use a clear primary image of the exact variant, without promotional overlays. Add alternate angles and detail images through additional image fields. Keep image URLs crawlable and stable, changing the URL when the image changes.

8. Use taxonomy for understanding and control

Populate Google product category with the most specific appropriate Google taxonomy value. Use product type for the retailer’s own hierarchy. Add custom labels for commercial states such as margin, stock and lifecycle. These fields are not interchangeable.

Validate counts after a rule change. If “Home > Lighting > Table Lamps” suddenly contains garden furniture, the taxonomy transformation has failed even if Merchant Center accepts the text.

9. Configure shipping, returns and promotions as data

Model the offer that a shopper actually receives. Test shipping rates by price threshold, weight, oversized class and region. Keep return policy details accurate and linked to the site. Use sale price and promotions features for genuine discounts rather than embedding “sale” or “free delivery” in titles.

Run test baskets before a campaign launch. The point is not just Merchant Center approval; it is avoiding a click that reveals an unexpected delivery charge or a price different from the listing.

10. Turn Needs attention into an operating queue

Review the queue daily for high-spend accounts and after every catalogue release. Prioritise by enforcement level, revenue affected, lost-click potential and fix confidence. Assign each issue to feed, ecommerce, policy or paid-media ownership.

Recheck performance after processing. Eligibility improvements should recover approved coverage and impressions; enrichment work should be measured on a defined product cohort. Merchant Center “optimisation score” or issue count is not the business outcome.

Field-level automation procedure

Choose one volatile field, such as availability. Record where stock is created, how reservations and cancellations change it, and how the storefront calculates purchasability. Map the source values to Google’s supported states. Decide how preorder and backorder dates are produced and displayed.

Then test timestamps at each stage: source update, feed generation, Merchant Center processing, page cache and checkout. If stock changes at 10:00 but Merchant Center receives it at 02:00 the following day, the optimisation is update frequency, not better copy. Add an alert for a source older than the agreed maximum age and a report of automatic corrections.

Run the same procedure for price, including VAT, currency, sale periods, member pricing and variant-specific differences. Avoid IP-based price switching that makes the crawler and customer see different offers. Use a test product and a high-velocity product in every release sample.

Shipping QA matrix

Record the Merchant Center configuration, product-level overrides and checkout result. A shipping setup is not accepted because one London postcode worked.

Failure cases and response

Manual edits keep disappearing: identify the primary source or rule that overwrites them and move the correction upstream. Automatic updates are frequent: shorten source latency and validate structured data rather than treating the corrections as success. Issue count falls but impressions do not recover: confirm country, destination, campaign inclusion and demand. New IDs appear after each sync: stop the release and restore stable ID generation.

For field definitions and conditional requirements, use the Merchant Center attributes guide. Set an acceptance threshold for every release: item-count reconciliation, priority approval coverage, no unexplained ID change and no new material issue family.

Measurement plan for Merchant Center work

Segment changes by purpose. Eligibility work uses approved product count, revenue-weighted coverage and recovered impressions. Freshness work uses mismatch recurrence, automatic corrections and cancelled orders. Enrichment work uses a fixed product cohort and item-level impressions, CTR and post-click quality. Shipping work uses rate accuracy, conversion and customer-service incidents.

Report the numerator and denominator. “Issues down 50%” is weak if the account moved from two issues to one; “priority approved revenue coverage increased from 76% to 97% across 420 launch products” is decision-grade evidence.

Add a monthly catalogue reconciliation: active products in the ecommerce platform, products submitted, products processed, products eligible and products receiving impressions. Each step can lose inventory for a different reason. The reconciliation shows whether the next investigation belongs in source selection, Merchant Center, campaign inclusion or demand.

Review automation permissions and owners quarterly. A forgotten supplemental source or old rule can continue overwriting correct values long after the original project ends. Remove obsolete transformations only after exporting them and confirming no product cohort still depends on their output.

Document each removal in the release log and monitor the next complete source-processing cycle for unexpected value changes or lost product coverage.

Sources

• Google Merchant Center: Issues in Merchant Center
• Google Merchant Center: Automatic item updates
• Google Merchant Center: Structured data
• Google Merchant Center: Product data specification