Google Shopping Ads occupy a strange spot in most advertisers' accounts: they're often the highest-ROAS campaign type running, yet they're also the one most people set up once, half-configure, and never revisit. Unlike Search campaigns, Shopping ads don't run on keywords you write. They run on your product feed, which means the quality of your setup is determined largely before you ever touch Google Ads itself. After years of auditing accounts across ecommerce PPC scaling engagements, feed quality is consistently the single biggest lever most businesses leave untouched.
This guide walks through the full setup process (Merchant Center, feed structure, campaign configuration, and the optimization levers that actually move performance), along with an honest look at where Performance Max fits into the picture now.
How Google Shopping Ads Actually Work
Shopping ads pull directly from a product feed you submit to Google Merchant Center, not from ad copy or keywords you write. Google's algorithm matches your product data (title, description, category, attributes) against user search queries automatically. This is the core difference from Search campaigns: your feed data functions as your targeting.
That means two accounts running identical bids and budgets can perform completely differently based purely on feed quality. A well-structured feed with accurate, keyword-rich titles and correct product categorization will out-earn a sloppy one every time, regardless of how sophisticated the bidding strategy is layered on top.
Setting Up Google Merchant Center
Before any campaign exists, your product data needs to live in Merchant Center in a format Google can parse reliably. The core setup steps:
- Verify and claim your website URL through Merchant Center's verification process, which confirms you own the domain the products link to.
- Set up tax and shipping settings accurately at the account level, since incorrect shipping costs are one of the most common causes of product disapprovals and poor performance.
- Choose a feed submission method: a scheduled feed fetch from a hosted file (most common for small-to-mid catalogs), a direct API integration (better for large or frequently changing catalogs), or a native platform integration if you're on Shopify, WooCommerce, or similar (usually the simplest starting point).
- Link Merchant Center to your Google Ads account so campaigns can pull from the feed.
Feed Optimization: The Part Everyone Skips
This is where most Shopping campaigns are won or lost, long before a campaign is even created.
Product Titles
Titles are the single highest-weighted field for query matching. Front-load the most important, searchable terms in roughly this order: brand, product type, then key attributes like size, color, and material. Generic titles like "Blue Shirt" waste the opportunity; "Men's Slim Fit Blue Oxford Shirt - Long Sleeve" gives the algorithm far more to match against.
Product Categories and Attributes
Google's own product taxonomy (google_product_category) should be set as specifically as possible, not left at a broad top-level category. Combine this with accurate attributes such as color, size, material, gender, and age group, since these fields both improve match quality and unlock filtering in the shopping experience itself.
Images
High-resolution, clean product images on a plain or white background consistently outperform lifestyle shots for Shopping placements specifically (lifestyle imagery tends to do better in Performance Max's broader placements). Avoid overlaid text, watermarks, or promotional badges baked into the image, since these are increasingly flagged as feed policy violations.
Custom Labels
The five custom label fields are underused by most advertisers. Use them to segment products by margin tier, bestseller status, seasonality, or clearance status. This segmentation becomes the backbone of a mature campaign structure where you can bid differently on high-margin versus low-margin products instead of treating the whole catalog identically.
Campaign Structure for Shopping
A tiered structure using custom labels or product groups tends to outperform a single flat campaign with all products lumped together, since it lets you allocate budget and bids according to actual business priorities rather than treating every SKU the same:
| Tier | Typical Contents | Bid Approach |
|---|---|---|
| High priority | Bestsellers, high-margin products | Higher bids, more budget allocation |
| Mid priority | Steady sellers, average margin | Moderate bids, standard budget |
| Low priority / clearance | Slow movers, low margin, excess inventory | Lower bids or exclude entirely if unprofitable |
A common approach: use a "priority" campaign setting (low, medium, high) combined with shared negative keyword lists and product group subdivisions by custom label, so budget naturally flows toward the products that matter most to the business rather than whatever happens to match the most queries.
Bidding Strategy for Shopping Campaigns
Manual CPC still has a role for brand-new campaigns with no conversion history, since it gives you direct control while you gather data. Once a campaign has enough conversion volume (generally a few dozen conversions in a recent window), automated strategies typically outperform manual bidding:
- Maximize Conversions: a reasonable default once you have baseline volume and want the algorithm to find as many conversions as possible within budget.
- Target ROAS: the better choice once you have enough conversion value data and a clear target return, particularly for catalogs with meaningfully different product margins.
- Target CPA: useful when your business cares more about a consistent cost per acquisition than maximizing total conversion value.
Whichever strategy you choose, give it time to exit the learning phase before judging results, since frequent bid strategy changes reset that learning process and tend to produce worse results than leaving a reasonable strategy alone longer than feels comfortable.
Negative Keywords in Shopping Campaigns
Shopping campaigns don't use keywords for targeting, but negative keywords still apply and matter more than most advertisers realize, since they're your only lever to prevent irrelevant query matches. Review the Search Terms report regularly and add negatives for:
- Queries for products you don't sell but that share terminology with ones you do.
- Research-intent terms ("reviews," "how to," "vs") that rarely convert for transactional products.
- Competitor brand names, if you don't want to compete on those terms (or conversely, monitor performance closely if you do).
A shared negative keyword list applied across all Shopping campaigns saves significant maintenance time compared to managing negatives per campaign.
Where Performance Max Fits In
Google has been pushing advertisers toward Performance Max as the successor to Standard Shopping campaigns, and for many accounts it now outperforms standalone Shopping on pure conversion volume, since it also draws from Search, Display, YouTube, and Gmail inventory using the same feed. The tradeoff is reduced visibility and control: you lose the granular Search Terms report and much of the manual negative keyword control that Standard Shopping offers.
A practical approach many advertisers use: run Performance Max for broad reach and volume, while keeping a smaller Standard Shopping campaign active specifically to retain search term visibility and negative keyword control, using insights from the Standard campaign to inform Performance Max exclusions where possible. This isn't the only valid approach, but it addresses the biggest complaint about a Performance Max-only setup: flying blind on what's actually triggering your ads.
Reporting: What to Actually Watch
Shopping campaigns generate a lot of surface-level data (impressions, clicks, CTR) that can look healthy while masking a weak underlying setup. A few metrics worth prioritizing over the vanity ones:
- Benchmark CTR and impression share by product group, not just campaign-wide. A campaign average can hide individual products with near-zero visibility due to feed issues or poor bids, dragged up by a handful of strong performers.
- Click share and lost impression share (rank vs. budget) tell you whether you're losing visibility to competitors on bid strength or simply running out of budget, two very different fixes.
- Conversion rate by product group or custom label, since a segment with a low conversion rate relative to the rest of the catalog often points to a landing page, pricing, or availability problem rather than a targeting problem.
Connecting Shopping performance data back to actual revenue and margin, not just conversion count, is where PPC ROI metrics and attribution becomes especially relevant for Shopping specifically, since a catalog with wide margin variance can show a healthy blended ROAS while individual product lines are actually unprofitable.
Seasonal and Inventory Considerations
Shopping campaigns are more sensitive to inventory and seasonality shifts than Search campaigns, since out-of-stock products are automatically excluded from serving, and that exclusion happens at the product level rather than requiring manual pausing. A few practices that matter more here than on other campaign types:
- Keep stock status accurate and current in the feed, since serving ads for out-of-stock products wastes budget and damages the user experience, while incorrectly marking in-stock products as unavailable silently kills their eligibility.
- Adjust custom label priorities ahead of seasonal peaks (holiday shopping, back-to-school, industry-specific cycles) so budget shifts toward seasonally relevant products before demand spikes rather than reactively after.
- Watch for feed processing delays around high-traffic periods: Merchant Center processing can lag during peak submission volume across the platform, so push critical pricing or availability updates with enough lead time rather than same-day.
Common Shopping Ads Mistakes
- Treating the feed as a one-time setup task instead of an ongoing optimization surface. Titles, categories, and images should be revisited regularly, not just at launch.
- Ignoring product disapprovals in Merchant Center, which silently remove products from eligibility without necessarily triggering an obvious account-wide alert.
- Running the entire catalog at a single bid instead of segmenting by margin or priority, which wastes budget on low-value products at the expense of bestsellers.
- Neglecting the Search Terms report because Shopping campaigns feel more "automated" than Search. The negative keyword discipline still matters.
- Switching bidding strategies too frequently, never letting any single strategy accumulate enough data to actually optimize.
Google's own Google Ads Help Center and Think with Google both publish regularly updated guidance on Shopping and Merchant Center policy changes, worth checking periodically since feed requirements and eligibility rules shift more often than most advertisers expect.