If Google Shopping performance is weak, do not start by changing the bid strategy again. First confirm that the product feed accurately describes what is sold, that products are eligible and in stock, that conversion values reflect real revenue, and that the catalog is segmented by business economics. Bidding can only optimize the product information and outcome signals it receives.
This matters especially for local and specialty retailers with mixed margins, seasonal inventory, store pickup, limited quantities, or large catalogs. One campaign-level return target can hide a catalog where a few products subsidize many items that cannot afford the click cost.
Why the Feed Is Media Strategy
Shopping ads are assembled from Merchant Center data rather than a manually written keyword ad for every item. Titles, descriptions, images, identifiers, price, availability, brand, product category, shipping, and other attributes help Google understand the product and match it to relevant searches. Inaccurate or thin data narrows the system’s ability to make a useful match.
Google’s product-data guidance states that properly formatted attributes help products appear in ads and free listings and match the right queries. That makes feed work more than technical maintenance. It shapes reach, relevance, eligibility, customer expectations, and the campaign structure available to the retailer.
Start With Eligibility and Accuracy
Review the Merchant Center “Needs attention” area and product status before diagnosing bidding. Resolve disapprovals, missing identifiers, mismatched price or availability, image problems, broken landing pages, and country or shipping inconsistencies. A campaign cannot recover demand for products that are ineligible or misrepresented.
Protect stable product IDs when possible so reporting and conversion history remain coherent. Confirm that inventory updates quickly enough for the business. A local retailer advertising an item that is no longer available creates wasted spend and a poor store or pickup experience even if the ad earns a click.
Rewrite Product Titles for How People Shop
A useful title identifies the product with the attributes a buyer uses to distinguish it: brand where relevant, product type, model, material, size, color, compatibility, or another high-value feature. Put the most important differentiators early while staying accurate and readable. Avoid promotional copy, keyword repetition, and details that contradict the landing page.
Descriptions should add helpful attributes and use cases, not repeat the title five times. Strong images need clear product visibility and should reflect the actual variant. For local inventory, pickup availability and store data must stay aligned with reality. Feed quality is a promise to the shopper, not merely a way to trigger more impressions.
Segment the Catalog by Economics
Two products with the same revenue can have very different allowable acquisition costs. One may have high margin, reliable inventory, and repeat-purchase value; another may be low margin, frequently returned, or expensive to ship. If both receive the same treatment, the campaign can chase attractive platform revenue while eroding profit.
Google supports up to five custom-label attributes for grouping products in Performance Max, Shopping, or Demand Gen campaigns. Retailers can use a consistent label scheme for margin tier, season, inventory depth, bestseller status, price band, or strategic category. Use labels for reporting and bidding decisions; do not create a new campaign for every small distinction.
Consolidate When Data Is Thin
Granular structure feels controlled, but it can leave each campaign with too little conversion evidence. If the account produces only a handful of monthly purchases, splitting by brand, category, priority, price, and geography may fragment the signal. Consolidation can help the system learn across more activity while custom labels preserve diagnostic visibility.
Separate only where the business needs a materially different budget, return target, market, or inventory strategy. A high-margin bestseller group may deserve its own treatment. A long tail of similar low-volume items may learn better together. Structure should reflect decision rights and data density, not an urge to make the account look organized.
Feed Better Conversion Signals Back
Revenue alone may not represent value. Where the platform and technical stack allow, pass accurate transaction values, exclude canceled orders, and analyze profit, returns, new-customer value, store pickup, and repeat behavior outside the ad interface. At minimum, compare platform return with actual contribution margin.
Retailers should also review whether one product earns the click while a different product completes the basket. Product-level reporting needs to inform merchandising and cross-sell decisions, not just item-level bid changes. The business wins on the order and the customer relationship, not on a single SKU’s dashboard row.
A 30-Day Feed-First Recovery Plan
In week one, audit eligibility, tracking, price, availability, shipping, identifiers, and landing-page consistency. In week two, improve titles, descriptions, images, and category data for the products with the most opportunity. In week three, apply a small, documented custom-label system based on margin, inventory, seasonality, and performance. In week four, simplify unnecessary campaign splits and establish a scorecard for revenue, margin, conversion volume, and product coverage.
Then let the cleaner structure collect enough evidence before making another major change. If Shopping activity is not translating into profitable sales, book a growth session to review the feed, measurement, and campaign structure as one retail acquisition system.
Frequently Asked Questions
What product-feed fields matter most?
Accurate titles, descriptions, images, identifiers, price, availability, brand, category, shipping, and variant data are foundational.
What are custom labels used for?
They group products for reporting and bidding using retailer-defined values such as margin tier, season, price band, or sales status.
Should every product category have its own campaign?
No. Split only when a category needs a distinct budget, goal, market, or economic treatment and has enough data to support the separation.
How often should a feed be updated?
Update as often as necessary to keep price, inventory, availability, and product details accurate. Fast-changing inventory may require frequent synchronization.
