Google Ads for Ecommerce: PMax and Shopping

Most ecommerce brands running Google Ads think they have a bidding problem. They almost always have a feed problem. When we take over an account, the Shopping feed is where we find the fastest wins, because it decides which products show, for which searches, at what price against competitors. Google Ads for ecommerce lives or dies on that feed long before you touch a bid strategy or a budget cap.

So let's start there, then work out to Performance Max, search structure, and the ROAS targets that keep you from buying revenue you can't afford.

The shopping feed is the real lever

Your feed is a product database Google reads to match queries. Titles, product types, GTINs, images, price, availability. Google matches search terms against that data, so a vague title like "Men's Jacket – Style 4471" gets shown for almost nothing worth having. Rewrite it to "Barbour Men's Waxed Cotton Jacket, Navy, Medium" and you qualify for the searches that convert.

A few things we change on nearly every account:

• Front-load titles with the terms people search: brand, product, key attribute, size or colour. The first 30 characters carry the most weight.
• Fill product_type and google_product_category properly. They help Google understand the item and give you cleaner segmentation later.
• Fix disapprovals and "missing GTIN" warnings before spending another dollar. Disapproved items just quietly don't serve.

Feed work is unglamorous and it compounds. A tighter feed cuts wasted impressions, which pulls your cost per acquisition down without touching bids. If cutting CPA is the goal, the feed is usually a bigger lever than the auction, and we go deeper on that in our guide to lowering CPA.

Where Performance Max fits, and where it bites

Performance Max is Google's black-box campaign type. One campaign spans Shopping, Search, Display, YouTube, Gmail, and Discover, with the machine deciding placement and bid. For a well-fed catalogue it can produce strong ROAS. We run it on plenty of accounts. But go in with your eyes open.

The main risk is that PMax eats your brand traffic and reports it as prospecting. Someone searches your brand name, PMax serves them, and the campaign takes credit for a sale that would have happened anyway at a fraction of the cost. Your blended numbers end up worse than the dashboard claims. Two defences we use: add brand terms as account-level negatives (or use the brand exclusion list) so PMax chases new demand, and keep a separate branded search campaign you actually control.

The second problem is opacity. You get limited search term data and almost no placement control out of the box. Use the "Insights" tab and the supplemental feed to steer it, feed it strong creative and accurate product data, and don't expect it to fix a weak catalogue. PMax amplifies what you give it, good or bad.

Structuring search so it doesn't overlap

Shopping and PMax cover product intent. Your search campaigns should cover the rest without colliding. We keep it simple.

Brand search gets its own campaign, tightly matched, cheap, and defended. This is where you protect the traffic PMax and competitors both want. Non-brand search covers category and problem-aware terms ("waterproof hiking boots", "gift for new dad") where Shopping alone doesn't tell the whole story. Keep those in exact and phrase match with a real negative keyword list, because broad match on non-brand terms is how budgets evaporate.

One rule that saves money: decide the hierarchy between Shopping, PMax, and Search so they aren't bidding against your own ads. If you run standard Shopping and PMax on the same products, PMax wins the auction by default, which hides whether Shopping was the more efficient buy. Pick one primary vehicle per product set. This is the same discipline behind consolidating campaigns instead of splintering budget across dozens of thin ad groups.

Feed optimisation past the basics

Once titles and categories are clean, the next gains come from custom labels and merchant promotions. Custom labels let you tag products by margin, season, bestseller status, or price band, then build campaigns or PMax asset groups around those tags. That is how you bid harder on your 60% margin items and pull back on the 12% ones instead of treating the whole catalogue as one blob.

Other feed moves worth the effort:

• Add merchant promotions so a "20% off" badge shows on your listing. Higher click-through at no extra CPC.
• Keep pricing and availability synced in near real time. A price mismatch between feed and landing page gets items disapproved fast.
• Test your main product image against a lifestyle shot where policy allows. Image is most of the click decision in Shopping.

Feed and landing page have to agree, so this work pairs directly with landing page optimisation. A great feed sending traffic to a slow, mispriced page just buys you expensive bounces.

Setting ROAS targets that hold margin

Here is where most accounts leak profit. A target ROAS is a revenue-to-spend ratio, and revenue is not margin. If your product carries 40% gross margin, a 4x ROAS campaign is roughly breakeven after cost of goods, before you have paid for shipping, returns, or overhead. Hitting your ROAS target and losing money at the same time is completely normal in Google Ads for ecommerce, and it catches a lot of people.

Work it backwards from margin. Take your true contribution margin per order, subtract the acquisition cost you can afford, and that gives you a break-even ROAS. Set your target above it with real headroom, not at the edge. Then watch blended numbers, not just the platform's self-reported figures, because PMax and Shopping both flatter themselves on brand and returning customers. We break down platform ROAS versus true blended performance in our note on ROAS versus MER, and it changes how you read every campaign.

One more thing: don't set tROAS too high too early. A punishing target starves the campaign of volume and the algorithm never gathers enough conversion data to optimise. Start closer to break-even, let it learn, then tighten. This is a core piece of how we run Google Ads for ecommerce clients, and it is usually the first thing we fix.

If you want a straight read on where your feed, PMax, and ROAS targets are quietly costing you, we offer a free ad account audit. We will tell you what we would change in the first thirty days and what we would leave alone.

FAQ

Should I run Performance Max or standard Shopping? Run whichever you can control against your goals, but not both on the same products at once, since PMax will win the auction and hide your real Shopping efficiency. Many accounts do well with PMax for prospecting plus a defended branded search campaign alongside it.

What ROAS target should I set for ecommerce? Set it above your break-even ROAS, which comes from your true contribution margin, not a generic number like 4x. Start near break-even so the algorithm gathers data, then tighten once it has volume.

Why are my products not showing in Shopping? Almost always a feed issue: disapprovals, missing GTINs, a price or availability mismatch between feed and landing page, or thin titles. Check the Merchant Center diagnostics tab first, before you assume it is a bidding or budget problem.