Google Ads for Ecommerce: Structure, Feed, PMax & Shopping Guide

Google Ads drives about 18% of all ecommerce revenue. It's the highest-intent paid channel available — people are actively searching for products they want to buy. But the platform has gotten complicated. Performance Max, Standard Shopping, Search, Demand Gen, AI Max — the campaign type list keeps growing and Google's documentation is designed to get you spending, not to help you spend well.

This guide covers the campaign architecture, feed mechanics, and bid strategy decisions that actually matter for ecommerce stores. No fluff about "setting up your first campaign." If you're reading this, you already have a Google Ads account. The question is whether it's structured to scale.

The campaign hierarchy

Every ecommerce Google Ads account needs three layers, in this order of priority:

Layer 1: Google Shopping (Standard Shopping campaigns). These show your product images, prices, and store name directly in search results. They capture high-intent product searches — someone typing "black leather boots size 9" or "wireless noise cancelling headphones under $200." Standard Shopping gives you more control over bidding, search term visibility, and budget allocation than Performance Max. Start here.

Layer 2: Performance Max. PMax runs across all Google surfaces — Search, Shopping, Display, YouTube, Discover, Gmail, Maps. It's Google's black box: you provide assets (images, videos, headlines, descriptions) and audience signals, and the algorithm decides where to show them. PMax is good at reaching audiences you wouldn't find through Shopping alone, but it cannibalizes branded search and gives you limited visibility into what's actually working.

Layer 3: Search campaigns. Text ads for non-product queries: "best [category] for [use case]," competitor brand names, informational queries related to your products. Search campaigns are how you reach people whose queries don't trigger product-format results.

Most stores run PMax as their only campaign and let Google handle everything. This is a mistake. You lose control over budget allocation between branded and non-branded queries, you can't see which search terms are driving conversions, and you can't separate prospecting from retargeting performance. The three-layer structure gives you control where it matters and lets automation handle the rest.

Feed optimization: the thing that matters most

Your product feed controls which search queries trigger your ads, what your Shopping listing looks like, and how Google categorizes your products. Most underperforming ecommerce Google Ads accounts have a feed problem, not a bid strategy problem.

Here's what to get right:

Product titles are the most important field. Google matches search queries to product titles, so your titles need to contain the words people actually search. The format that works: Brand + Product Type + Key Attribute + Size/Color/Variant. "Nike Air Max 90 Running Shoes Men's Black Size 10" beats "Air Max 90 Black" every time.

For fashion and apparel, front-load the product type: "Men's Running Shoes Nike Air Max 90 Black." For electronics, lead with the brand: "Sony WH-1000XM5 Wireless Noise Cancelling Headphones." Match the format to how people in your category actually search.

Product descriptions matter less for matching but affect Quality Score. Write them for humans. Don't keyword-stuff.

Set your Google Product Category to the most specific level available. Don't use "Apparel & Accessories" when "Apparel & Accessories > Clothing > Shirts & Tops > T-Shirts" exists. Specific categorization improves matching accuracy.

GTINs (Global Trade Item Numbers) are required for resellers and strongly recommended for everyone else. If you manufacture your own products, apply for a GTIN exemption in Google Merchant Center. Missing or incorrect GTINs cause disapprovals, and disapprovals mean your products don't show.

Product images should be clean, high-resolution, white-background shots for the primary image. Lifestyle images go in the additional_image_link field. Shopping is a visual format — a bad product photo loses the click before the bid matters.

Use custom labels to segment products by margin tier, price point, seasonality, or bestseller status. This lets you set different ROAS targets for different product groups instead of applying one target across everything.

Automate your feed management if you have more than 50 SKUs. Shopify's Google & YouTube channel handles basic syncing. For more control, use DataFeedWatch, Feedonomics, or GoDataFeed. Manual management breaks at scale.

Performance Max: how to use it without losing control

PMax is a good campaign type used badly by most stores. The default setup — dump all products into one asset group with no audience signals — is how Google wants you to use it, because it maximizes Google's revenue. It's not how you should use it.

Segment by product category or margin. Create separate asset groups for different product lines. A skincare brand should have one asset group for cleansers, one for moisturizers, one for serums. Each gets its own creative assets and its own performance data. This lets you see what's working and allocate budget accordingly.

Add audience signals. PMax uses audience signals as starting hints for its targeting algorithm. Add your customer lists (purchasers, email subscribers), your website visitors, and relevant in-market and affinity audiences. Google will expand beyond these, but the signals improve the starting point.

Exclude branded keywords. By default, PMax captures your branded search traffic and takes credit for it. This inflates PMax's reported ROAS because branded searchers would have found you anyway. Add brand terms to your account-level negative keyword list (you can now do this in Google Ads settings) and run branded traffic through a dedicated Search campaign where you can see it clearly.

Watch for cannibalization. If you're running Standard Shopping and PMax simultaneously, PMax will try to win the same auctions. Check your search terms report in Standard Shopping — if impressions drop after launching PMax, it's cannibalizing. The fix: use campaign priority settings and shared budgets to control which campaign wins which queries.

Don't judge PMax by PMax-reported ROAS. PMax will always report good numbers because it's capturing your easiest-to-convert traffic (branded search, retargeting) alongside genuine prospecting. Judge PMax by its incremental contribution: did total account revenue increase when you added PMax, holding everything else constant? If total revenue went up but Standard Shopping revenue went down by the same amount, PMax isn't adding anything — it's just stealing credit.

Search campaigns for ecommerce

Search campaigns play a supporting role in ecommerce, but they're the best channel for three specific use cases:

Competitor conquesting. Bidding on competitor brand names when someone searches "Allbirds shoes" or "Glossier skincare" and showing your ad instead. This works when you have a genuine alternative and your landing page makes the comparison obvious. It doesn't work when your ad just says "Buy Our Shoes" with no reason to switch.

Category and informational queries. "Best running shoes for flat feet" or "how to choose a moisturizer for oily skin." These queries have purchase intent but no specific product in mind. Your ad can link to a category page or a buying guide. The conversion rate will be lower than Shopping but the audience is earlier in the funnel and less price-sensitive.

Branded search protection. Running ads on your own brand name to prevent competitors from stealing your traffic. This is defensive, not offensive. Keep bids low, use Target Impression Share as the bid strategy, and make sure you're showing for your own name.

Use phrase match and exact match for non-branded keywords. Broad match in Search campaigns for ecommerce generates garbage traffic unless you have a large budget and strong conversion data for Smart Bidding to learn from. If you're spending under $10K/month on Search, stick with tighter match types.

Bid strategy decisions

Standard Shopping: Start with Manual CPC or Maximize Clicks to gather conversion data. Once you have 30+ conversions per month in the campaign, switch to Target ROAS. Set Target ROAS based on your actual break-even economics, not on a benchmark.

Performance Max: Always runs on automated bidding. You choose between Maximize Conversion Value (no cap — Google spends your budget to maximize revenue) and Maximize Conversion Value with a Target ROAS (Google tries to hit your ROAS target). Start without a ROAS target for 2–4 weeks to let the algorithm learn, then add a target once you have data.

Search campaigns: For branded, use Target Impression Share (aim for 90%+ on your own terms). For non-branded, use Target CPA or Target ROAS once you have enough conversion data.

The common mistake: setting Target ROAS too high and starving the campaign. If your Target ROAS is 8x but your actual ROAS is 4x, Google won't be able to find auctions that meet your target and will reduce spend dramatically. Start with a target 10–20% below your current ROAS and gradually increase.

The measurement problem nobody talks about

Google Ads reports ROAS using its own last-click attribution by default. This number is inflated — sometimes massively.

Here's why: if someone sees your Meta ad, visits your site, leaves, then Googles your brand name and buys, Google Ads claims that conversion. If someone sees your YouTube brand ad, remembers your name, searches for you three days later and buys, Google claims it. If someone would have bought anyway through organic search, but happened to click a branded PPC ad first, Google claims it.

This means your Google Ads ROAS isn't measuring what you think it's measuring. It's measuring conversions that touched Google Ads at any point, not conversions that Google Ads caused.

The fix: track blended ROAS (total revenue ÷ total ad spend across all channels) and compare it to platform-reported ROAS. The gap between the two tells you how much credit inflation is happening. If Google reports 6x ROAS but your blended ROAS is 2.5x, there's a lot of overlap and credit-stealing between channels.

For more on getting measurement right across your entire ecommerce marketing mix, read our ecommerce marketing strategy guide.

The brand effect on Google Ads performance

Most Google Ads guides don't cover this because it's outside the ad account.

Your Google Ads performance has a ceiling set by your brand awareness. A brand nobody has heard of will have lower CTR on Shopping ads because nobody recognizes the logo. Conversion rates will be lower because there's no trust. CPA will be higher across every campaign type. You can optimize the feed, fix the bid strategy, perfect the campaign structure — and still underperform a worse-run account from a brand people actually know.

This is why we treat Google Ads as one component of a broader ecommerce marketing strategy, not as a standalone channel. The brands that invest in fame and mental availability see their Google Ads performance improve as a side effect — higher CTR, higher CVR, lower CPA — without changing anything in the ad account itself.

Google Ads captures demand. It doesn't create it. If nobody is searching for you or your product category, the best campaign structure in the world won't change that.

Checklist: Google Ads account audit for ecommerce

If you're running Google Ads for an ecommerce store and want to know if the setup is sound, check these:

• Product titles include the words people search (not just your internal product names)
• Google Product Categories are set to the most specific level available
• GTINs are present and correct (or exempted)
• Product images are high-quality, white-background primary shots
• Custom labels segment products by margin tier at minimum
• Standard Shopping campaigns run alongside PMax (not PMax only)
• PMax has segmented asset groups by product category
• Branded keywords are excluded from PMax
• Search campaigns cover branded terms with Target Impression Share
• Conversion tracking uses both the Google tag and server-side tracking (GA4 + Ads tag)
• Blended ROAS is tracked alongside platform-reported ROAS
• Target ROAS settings match actual break-even economics, not arbitrary benchmarks