Google Shopping & Performance Max

Google Shopping lets retailers show product listings, complete with a photo, price, and store name, directly in Google search results. You do not write ad copy. Instead, you upload a product feed (a structured file that describes your inventory), and Google decides which products to show for which searches.

If someone types "women's trail running shoes under $80," Google scans your feed and surfaces the best-matching products automatically.

Performance Max (PMax) is Google's next step beyond Shopping. It takes your product feed and adds creative assets, headlines, images, and videos, then distributes ads across every Google surface at once: Search, Shopping, YouTube, Display, Discover, Gmail, and Maps.

One campaign can reach a shopper who is researching on Google in the morning, watching YouTube at lunch, and browsing a news site in the evening.

Quick Summary

Google Shopping uses a product feed, not keywords, to match your products to search queries
Performance Max (PMax) extends Shopping to run across all Google channels from a single campaign
Both campaign types rely on Smart Bidding (Target ROAS or Target CPA) powered by Google's AI
Feed quality is your biggest lever, better product data beats bigger budgets
PMax needs at least 30-50 monthly conversions to learn effectively; launch without this and you waste budget

Why It Matters

Paid search used to mean picking keywords and writing ads manually. Shopping and PMax represent a structural shift: the algorithm now handles placement, bidding, and creative combinations. Marketers who understand this system can scale to thousands of products without writing a single ad.

Note

The scale advantage is enormous. An e-commerce store with 5,000 products would need thousands of ad groups to manage them manually. With a well-structured product feed, one PMax campaign covers all 5,000 SKUs, automatically showing the right product for every relevant search query.

How Google Shopping Works

Shopping campaigns start inside Google Merchant Center, a separate free account where you upload and maintain your product feed. The feed is typically a spreadsheet or automated export with fields like:

title, the product name (most important field for matching queries)
description, supporting detail and keywords
price and availability, must stay accurate or ads get disapproved
image_link, the photo shown in the listing
google_product_category, Google's taxonomy code for your product type
gtin or mpn, the barcode or manufacturer part number (boosts eligibility)

Google ingests this feed, validates it, and makes those products eligible to appear in Shopping results. Feed quality is everything.

A title like "Blue Shoe Men" will barely match anything. A title like "Nike Air Zoom Pegasus 41 Men's Running Shoe Blue Size 10" matches dozens of relevant search queries because it mirrors how real shoppers search.

How Performance Max Works

PMax builds on your Merchant Center feed but adds an asset layer. You supply:

Up to 15 headlines
Up to 5 descriptions
Up to 20 images (multiple aspect ratios)
At least 1 video (Google auto-generates one if you skip this, but it is usually low quality)

Google's AI assembles thousands of creative combinations from these assets and tests them across every channel. You also provide audience signals, these are hints, not restrictions. You might upload your customer email list and say "find more people like these." The algorithm uses that as a starting point, then expands beyond it once it finds higher-converting patterns.

Smart Bidding drives the whole system. Target ROAS (return on ad spend, how many dollars you earn per dollar spent) tells Google to maximize revenue while hitting your revenue goal. Target CPA (cost per acquisition, what you pay to get one customer) tells it to maximize conversions within a cost ceiling.

The AI adjusts bids on every single auction, factoring in device type, location, time of day, user intent signals, and hundreds of other inputs simultaneously.

The Learning Period

After you launch PMax or make any major change to it, the campaign enters a learning period that typically lasts 1-2 weeks. During this time, the AI is running experiments and its efficiency is lower than normal. You will likely see higher CPAs and inconsistent results. This is normal.

What makes it worse: every significant change (new budget, new bid target, major feed changes, adding or removing asset groups) restarts the clock. This is why experienced practitioners plan changes carefully rather than tweaking campaigns daily.

Common Mistake

Do not launch PMax without sufficient conversion history. Google's own recommendation is at least 30-50 conversions per month per campaign before switching. Launch with less data and the algorithm guesses broadly, burning budget while it learns. The fix: run a Standard Shopping campaign first, build up conversion volume over 2-3 months, then migrate to PMax once the data is there.

Real Case Studies

KEH Camera: 76.3% Revenue Increase (2023)

KEH Camera is a US-based online retailer of secondhand camera equipment. Before PMax, they were running over 70 separate Smart Shopping campaigns. In 2022, they migrated to Performance Max using a structured five-step approach: pausing auto-generated PMax campaigns first, consolidating their best-performing Smart Shopping campaigns, then gradually shifting budget to PMax.

The results over Q1 2022 versus Q1 2023:

Revenue from ads increased 76.3%
Transactions increased 44.1%
Average monthly ROAS over six months: 9.93x (that means $9.93 earned for every $1 spent)
Their New Year's sale alone drove 16.1% of monthly PMax revenue at a 12.1x ROAS

The key: KEH did not dump all 70 campaigns into PMax at once. They tested with their best performers, measured, then expanded. Consolidation + patience was the strategy.

MoneyMe: 22% More Conversions, 20% Lower CPA (2024)

MoneyMe is an Australian fintech (financial technology) company offering personal loans. They ran Performance Max campaigns targeting users across Search, Display, and YouTube simultaneously. The result: 22% more conversions compared to their previous campaign setup, with a 20% reduction in cost per acquisition.

This matters because fintech is a high-competition, high-CPA category. Getting 20% cheaper conversions in that environment is a significant efficiency gain.

Real Example

How asset groups work in practice. Imagine you sell running shoes and hiking boots. You would create two separate asset groups inside one PMax campaign, one with running-focused headlines, images of runners, and descriptions about speed and cushioning; the other with trail-themed headlines, outdoor imagery, and descriptions about grip and durability. Google then shows the most relevant asset group to each user based on their intent signals. This is the closest thing PMax has to ad group segmentation.

The Product Feed Is Your Real Creative

In traditional search ads, your competitive advantage is your ad copy. In Shopping and PMax, your advantage is your product data. Here is what separates high-performing feeds from mediocre ones:

Product Titles (most important): Mirror the exact language your customers use. Include: brand + product name + key variant details (color, size, material, model number). Google reads titles to match your products to search queries.

Product Images: Shopping is visual. Clean, white-background images perform best for most categories. For apparel and lifestyle products, on-model images can lift CTR (click-through rate, the percentage of people who click after seeing your ad).

GTINs and MPNs: If your product has a barcode (GTIN) or manufacturer part number (MPN), include it. Google uses these to match your products to structured catalog data, which improves eligibility and quality scores.

Custom Labels: These are fields you can populate with your own data to help you segment campaigns. Common uses:

Margin tier (high-margin vs. low-margin products)
Inventory level (overstocked items to push)
Seasonality (summer vs. winter products)

Smart Bidding: What You Control vs. What the Algorithm Controls

Understanding this boundary is critical. Here is what you control:

The conversion goal (ROAS target or CPA target)
The daily budget
The product feed and creative assets you provide
Audience signals (who to use as a starting point)
What to exclude (negative keywords at campaign level, brand exclusions)

Here is what the algorithm controls:

Which auction to enter and at what bid price
Which creative combination to show each user
Which channels to prioritize for each user
How to expand beyond your audience signals

The mistake most beginners make is trying to control too much: micromanaging bids, changing targets every few days, or running too many campaigns that cannibalize each other. The mistake at the other extreme is giving the algorithm too little guidance, no audience signals, poor creative assets, weak feed data. The skill is finding the right balance.

Pro Tip

PMax will spend budget on branded searches (people searching your company name directly) that would have converted anyway. This inflates ROAS numbers but does not represent real incremental growth. Use the Search Terms Insights tab inside PMax to spot this. Many practitioners add a separate branded Search campaign with a higher priority so branded traffic routes there instead, keeping PMax focused on finding new customers.

2025 Updates That Change How You Work with PMax

Google has added features in 2024-2025 that give advertisers more visibility and control:

Asset-level performance data: You can now see which individual headlines, images, and descriptions are performing best, previously all assets were reported as a group
Campaign-level negative keywords: You can now exclude specific search terms from PMax at the campaign level (not just account level), which was a major advertiser complaint for years
Channel reporting: Breakdown of where your PMax budget is actually going across Search, Shopping, YouTube, Display, and other placements
Search term insights: A cleaner view of the actual queries triggering your PMax ads

These updates matter because they address the biggest criticism of PMax: it was a black box. Advertisers are now getting more transparency into what the algorithm is actually doing.

Common Mistakes

Common Mistake

Neglecting feed maintenance after launch. The product feed is not a one-time setup, it is a living document. Prices change, products go out of stock, new items launch. A feed that is 90 days stale will have disapproved products, wrong prices, and missing inventory. Set up automated feed refreshes (daily is ideal for active catalogs). Monitor the Merchant Center diagnostics tab weekly for disapprovals.

Campaign Structure Principles

Keep these rules when structuring Shopping and PMax campaigns:

One PMax campaign per distinct goal, if you have products with a 60% margin and products with a 10% margin, they should not share a campaign or a ROAS target
Segment asset groups by product category and audience, do not mix running shoes and hiking boots into one asset group; the algorithm cannot serve relevant creatives for both
Provide at least one video asset, if you skip this, Google auto-generates a video from your images that is often visually weak; a simple 15-second product video you record yourself beats it
Set realistic ROAS targets, if your current Smart Shopping achieves 400% ROAS, do not launch PMax at a 700% target; start at a match or slightly below and tighten as the algorithm learns

The One-Line Takeaway

Your product feed is your ad copy, invest in its quality first, then let the algorithm handle the rest.

Bidding Strategies, Smart Bidding (Target ROAS, Target CPA) is the engine inside every PMax campaign; understanding how the auction works explains why feed quality and conversion volume are prerequisites
Google Search Ads, Shopping replaces keyword-level control with feed-level control; knowing traditional Search campaigns helps you recognize what PMax handles automatically versus what you still need to manage manually
MMM vs MTA, PMax runs across Search, YouTube, and Display simultaneously, which creates attribution complexity that last-click models misread; understanding measurement frameworks helps you evaluate its true incremental contribution

You Might Also Like