The Maze: Performance Max did not remove the work from ecommerce advertising. It moved the work. As Google automates bidding, targeting and placement, the performance ceiling is increasingly set by what the business gives the system: an offer worth buying, a page that closes the sale, product data that explains the catalogue, and conversion signals tied to real value. Account settings still matter. They just cannot manufacture demand, margin or trust.
The offer sets the economic ceiling. A bidding system can move spend toward the outcomes it is shown, but it cannot make an undifferentiated product more compelling or turn weak unit economics into profitable growth. Google's own retail guidance starts with business goals and recommends transaction-specific conversion values when products carry different prices. That matters because revenue is not profit. The account needs targets that reflect margin, customer value and the reason a shopper should choose this product now. Otherwise automation becomes very efficient at scaling the wrong sale.
The feed and landing page form one promise. Product data decides how Google understands the catalogue before the click. The page decides whether the promise survives after it. Google says product data shapes how ads and free listings behave, and it asks retailers to keep titles, attributes, prices and availability accurate while matching that information to the landing page. This is not catalogue housekeeping. A vague title weakens matching. A price mismatch kills trust. A slow or confusing checkout wastes the click. Better bidding cannot repair a broken handoff from query to product to purchase.
Measurement teaches the machine what winning means. Conversion tracking is not only a reporting layer after the campaign. It is the feedback loop used to allocate the next pound, euro or dollar. Google's enhanced-conversions guidance uses consented first-party data to recover more observable conversions and improve bidding. But coverage alone is not the goal. The event must represent the right action, carry the right value and avoid rewarding low-quality outcomes. Feed poor signals into automation and it does not become random. It becomes consistently wrong.
Platform levers belong second, not nowhere. Campaign structure, Smart Bidding, creative, Ad Strength, exclusions and experiments still change performance. Google recommends consolidating Performance Max where budgets and objectives allow, because more shared data can speed learning. It also recommends strong creative coverage, value-based bidding, relevant exclusions and active reporting. The mistake is turning an Optimization Score or an Excellent Ad Strength label into the business objective. These are diagnostics and tuning tools. Use them after the offer, page, feed and measurement are good enough to amplify.
The operating model is now cross-functional. Merchandising owns the proposition and margin. Ecommerce owns the product page and checkout. Catalogue teams own feed accuracy. Data teams own the conversion signal. Paid media still owns the auction strategy, but it can no longer pretend the auction is the whole system.
Why it matters: The next Google Ads breakthrough may not come from a new campaign type. It may come from fixing the price architecture, rewriting product titles, speeding up the product page or assigning purchase values correctly. That changes where leaders should put time and budget. Stop asking the media buyer to rescue every weakness from inside the account. Give the algorithm a stronger business to optimize, then make the platform prove it can allocate better.
