Text version of this lessonExpand
Many teams treat Merchant Center like a place to upload products for ads. For ecommerce, it is closer to Google’s product-understanding layer. Titles, images, price, stock, attributes, policy pages, and landing-page consistency shape whether Google can show your products and whether it shows them to the right people. This lesson gives you a product feed launch checklist and a Merchant Center issue triage simulator: make priority SKUs understandable, sellable, and trustworthy before Shopping / PMax decisions.
Lesson output: product feed launch checklist
Product upload success does not prove that Google understands the product. Before launch, sample priority SKUs and review fields, pages, checkout, policies, and Merchant Center diagnostics in one checklist.
| Layer | Check first | Healthy sign | Do not do yet |
|---|---|---|---|
| Title | Does title explain product type, brand, spec, use case, and key attributes? | Both shopper and system can understand the product quickly. | Do not use only internal SKU names, slogans, or new arrival language. |
| Image | Does image_link clearly show the actual sold variant? | Ad card and product-page hero are consistent. | Do not force Shopping with a lifestyle image that hides the item. |
| Price and availability | Do feed, product page, and checkout agree on price / availability? | Promotion, out-of-stock, preorder, and restock states sync reliably. | Do not keep advertising SKUs with wrong price or stock state. |
| Product identity | Do brand, GTIN, MPN, SKU, variant, and product_type match the real product? | Standard products use real GTINs; private or custom items do not invent barcodes. | Do not use SKU as GTIN or guess a barcode. |
| Link and policy | Do product link, shipping, returns, contact info, and policy status support each other? | The shopper lands on a buyable, trustworthy page. | Do not ask ad budget to solve page-trust problems. |
Plain terms: Merchant Center, feed, GTIN, and diagnostics
Merchant Center: Google’s platform for receiving, reviewing, and managing product data. You see products, diagnostics, data sources, shipping/tax, returns, and website verification there.
Product feed: the set of product fields submitted to Google through Shopify sync, sheets, files, APIs, or supplemental sources. Common fields include title, image_link, price, availability, brand, GTIN, product_type, and link.
GTIN: a Global Trade Item Number, the barcode identity on a product package. Common forms include UPC, EAN, JAN, or ISBN. SKU is store-owned and is not GTIN. Use the real manufacturer barcode when it exists; do not invent one for private or custom products.
Checkout: the path where a shopper confirms price, discount, shipping, tax, stock, and buyability before payment. Merchant Center and ad decisions are affected by product-page and checkout facts, not only the feed row. If the feed says in_stock but checkout cannot buy the SKU, this is a broken product fact, not an ad optimization problem.
Diagnostics / issue details: the Merchant Center surface for product disapprovals, warnings, and data issues. It is not generic technical noise. It tells you whether the product can show, whether matching quality is weak, or whether account risk exists.
Filled fields do not prove consistent facts
Product data is not just a machine file. The buyer-facing price, availability, image, and promise must match the feed. Otherwise Shopping / PMax learns from unstable product facts.
| Fact | Compare | Risk | Proof |
|---|---|---|---|
| Price fact | Feed price, product-page price, checkout price, sale-price effective period. | Price mismatch hurts trust, review stability, and conversion quality. | Screenshot product and checkout pages, then record feed update time. |
| Availability fact | Feed availability, Shopify inventory, product-page button, preorder/out-of-stock copy. | Sending budget to out-of-stock items wastes demand. | Sample priority SKUs and record out-of-stock, preorder, and restock dates. |
| Product identity fact | SKU, GTIN, brand, variant, product_type, and product title. | Identity confusion makes the system place products in the wrong comparison set. | Create one sample row for a standard product and one for a private/custom product. |
| Landing-page fact | Ad-card image, product-page main image, product description, shipping and returns. | Promise mismatch reduces post-click buying intent and can trigger issues. | Save product hero, policy pages, and Merchant Center issue details. |
For large catalogs, do not chase full perfection first
Large catalogs create paralysis because there are too many issues. The steadier move is not to clean everything at once. Start with the products that matter most commercially.
- High-impression SKUs: they already get exposure, so title, image, or price issues directly affect clicks and matching.
- High-margin SKUs: even with lower volume, they deserve cleaner attributes, images, and policy consistency first.
- Promotion-priority SKUs: check sale price, stock, discount page, shipping promise, and fetch timing before launch week.
- Issue-heavy categories: when the same issue repeats across a category, set category rules first, then review core SKUs in batches.
Merchant Center issue triage simulator
When Merchant Center flags an issue, do not fix by issue count and do not jump into the ad console. First decide whether the issue blocks eligibility, product identity, pre-click trust, or page/policy trust.
| Symptom | Severity | First proof | Fix first | Do not do |
|---|---|---|---|---|
| Price or availability mismatch | Eligibility blocker or high-risk quality issue. | For the same SKU: feed row, product page, checkout page, Shopify stock, and latest update time. | Align price and availability facts; pause broken SKUs if needed. | Do not raise budget, change tROAS, or blame learning first. |
| GTIN / brand identity conflict | Traffic-quality issue that can become disapproval. | Supplier barcode, package photo, Shopify fields, gtin field, and product-page brand spelling. | Use the real manufacturer GTIN when it exists; do not invent one when it does not. | Do not copy a random barcode. |
| Weak title or image | Traffic-quality issue. | High-impression low-click or high-click low-conversion SKUs, title, image_link, page hero, and search demand. | Start with high-impression or high-margin SKUs; title should include category, spec, use, and key attributes. | Do not use bids or budget to hide a product card shoppers cannot understand. |
| Link, policy, or page trust risk | Eligibility or account-risk issue. | Product link, mobile hero, shipping policy, return policy, contact page, and issue details. | Fix accessibility and policy facts first, then check structured data. | Do not ask ad budget to solve page-trust problems. |
20oz SKU feed QA lab: choose the SKU issue, then choose the repair
The worst feed review habit is calling every problem bad product data. The real skill is triage: is this SKU blocked by an availability mismatch, unclear product identity, a weak product card, or missing page policy proof? Once that is clear, you can decide whether to pause the SKU, fix identity, rewrite title / image, fix page policy, or keep observing.
| SKU scenario | Hidden issue | Correct action | First proof | Write back to checklist |
|---|---|---|---|---|
| MUG-20-BLK: 20oz tumbler feed says in_stock, the black variant is sold out on the page, checkout cannot buy it, and Shopping still gets clicks with 0 purchases. | Availability facts conflict. This is not a bidding or learning-period problem. | Pause the problem SKU or exclude the variant, then fix inventory sync. | Feed row, product-page variant, checkout, Shopify inventory, and latest Merchant Center processing time. | Availability facts mismatch; hold Shopping / PMax judgment until feed, page, and checkout screenshots align. |
| BRUSH-ELEC-2PK: a 2-pack electric toothbrush bundle uses the single-item package GTIN in the feed, and issue details flag a possible identifier problem. | Product identity has not passed. Google may understand the bundle as a single item. | Fix GTIN / bundle / brand identity. Do not guess when unsure. | Supplier GTIN, package photo, bundle rule, Shopify field, feed gtin, and product-page title. | Validate identity before the SKU returns to product-group learning. |
| DOG-COVER-L: dog car cover has high impressions but weak clicks. The title only says Dog Cover, and image_link is a distant lifestyle image. | The pre-click product card is unclear. This is not a budget shortage. | Rewrite title and main image so the card clearly shows category, size, use, and key attributes. | High-impression low-CTR SKU, title, image_link, product-page first screen, search demand, and competitor cards. | Enter title / image_link repair queue; validate product-card clarity before budget decisions. |
| RUG-8X10-SAGE: price and stock align and the item can be added to cart, but the return policy page is 404 and shipping details are scattered in FAQ. | Page and policy facts do not support the ad click. | Fix return page, shipping promise, contact info, and structured data. | Product link, mobile first screen, return page, shipping page, contact page, structured data, and issue details. | Page / policy trust gate failed; add proof before ad judgment. |
Do not ask ad budget to solve feed problems
If the SKU cannot be bought, identifiers are not trusted, the product card is unclear, or the policy page is inaccessible, fix the facts first. More budget, a tROAS change, or a new PMax goal cannot replace product-fact validation.
Read Merchant Center issues by business impact
Issue details help identify, understand, and resolve issues affecting products. Operationally, group them into three layers instead of reading only issue count.
| Level | Signal | Impact | Responsible |
|---|---|---|---|
| Eligibility blockers | Disapproval, account risk, unreachable links, severe price or availability mismatch. | Products cannot show reliably; fix before scale talk. | Product-data lead + site lead. |
| Traffic-quality issues | Generic titles, weak images, missing attributes, brand or GTIN confusion. | Products can show, but matching quality is weaker. | Merchandising + ads lead. |
| Optimization opportunities | More attributes, better images, richer details, stronger structured data. | Not always urgent today, but affects later scaling efficiency. | Merchandising + content lead. |
A feed edit is not done until impact is recorded
If performance improves, the team needs to know what to copy. If it worsens, the team needs to know what to roll back. Without records, feed optimization becomes unexplained field drift.
- Changed field: title, image_link, price, availability, GTIN, category, or link.
- Affected SKUs: record SKU, product name, category, and whether it is a priority product.
- Reason: which diagnostic, mismatch, or business priority drove the change.
- Update time: source-system update time and Merchant Center fetch or processing state.
- Observation window: how many days, which SKUs, and which metrics decide continue or rollback.
30-minute feed launch review: validate 10 priority SKUs first
A beginner does not need to clean the full catalog on day one, but the team must be able to explain the most important 10 SKUs. Use 30 minutes to align fields, pages, diagnostics, and ad readiness.
| Time | Question | Output |
|---|---|---|
| 0-5 minutes | Which 10 SKUs are sampled today: high impression, high margin, promotion priority, or issue-heavy category? | SKU list and priority reason. |
| 5-12 minutes | Do title, image_link, price, availability, GTIN, and link have obvious breaks? | Mark each SKU as pass, repair, or pause. |
| 12-18 minutes | Do product page, checkout, shipping, returns, and contact information match the feed? | Page proof and issue-detail evidence. |
| 18-24 minutes | Is each diagnostic an eligibility blocker, traffic-quality issue, or optimization opportunity? | Responsible person and first repair action. |
| 24-30 minutes | After the edit, how long do we watch, and what triggers rollback or continuation? | Observation window, review metric, and next check date. |
Three SKU repair examples
MUG-20-BLK: title only says Travel Mug, with no capacity, material, or leak-proof value. Repair by adding capacity, main use, and key attributes to the title, then matching the product-page first screen.
RUG-8X10-SAGE: feed says in stock, but the product-page variant is sold out. Fix inventory sync cadence, pause sold-out variants first, then review Merchant Center diagnostics.
DOG-COVER-L: main image is lifestyle-heavy; product body, size, and installation are unclear. Use a clear main product image, add installation images, and ensure image_link points to the main image.
Stop / Go rules
Do not let Shopping / PMax carry budget before these proofs are ready
- If high-risk diagnostics, disapproval, or account status issues remain, fix eligibility first.
- If price, availability, image, or landing page conflicts with feed, validate fact consistency first.
- If priority SKUs have weak titles, confused GTINs, or missing attributes, complete at least 10 sample checklist rows first.
- If feed changes lack affected SKUs, timing, and rollback condition, write the change log first.
When the feed is stable, the ad system has a chance to learn correctly. Otherwise, the next channel-split lesson starts from bad product facts.
Copyable lesson notes: make feed validation reviewable next time
Do not leave this lesson with a vague note like improve the feed. Useful feed work records current pressure, first proof, repair action, stop boundary, and review window. When CTR, CVR, Shopping impressions, or PMax product distribution changes later, the team can return to the same SKU evidence instead of guessing again.
| Note item | What to write |
|---|---|
| Current pressure | Example: 3 of 10 priority SKUs have price, stock, or checkout fact conflicts; hold them from Shopping / PMax judgment first. |
| First proof | For the same SKU: feed row, product page, checkout page, Shopify inventory, latest Merchant Center processing time, and issue-detail screenshots. |
| This week action | Fix buyability and eligibility issues first, then title, image, GTIN, product_type, and policy-page issues. |
| Stop action | If the SKU is out of stock but still advertised, GTIN is guessed, policy page is 404, or checkout cannot buy, do not raise budget or blame PMax learning first. |
| Review window | Record feed update time, Merchant Center processing state, and watch 3-7 days of impressions, clicks, CVR, disapprovals, and product distribution. |
| Next route | After feed validation passes, move to the next lesson and decide the roles of Search, Shopping, and PMax. |