Feed-Only Performance Max: How to Narrow Campaign Focus for Better E-Commerce ROAS

Every e-commerce advertiser running Performance Max has asked the same question at least once: where is my budget actually going? You set a ROAS target, upload your product feed, and hand Google the keys. Weeks later, the channel performance report reveals that 30-50% of your spend went to Display placements with conversion rates a fraction of what Shopping delivers. Search traffic, while typically highest quality, often represents only 20-35% of total Performance Max spending, while the Display Network can consume 30-50% of budgets with significantly lower conversion rates.

Feed-only Performance Max exists to solve this exact problem. A feed-only Performance Max campaign is built with zero manually added creative assets - no headlines, no descriptions, no lifestyle images, and no uploaded videos. By stripping away everything except the product feed, you force Google's algorithm to concentrate spend where purchase intent is strongest: Shopping placements and dynamic remarketing. This isn't a hack for beginners or a shortcut around proper campaign management. It's a precision instrument. And in 2026, with Google continuing to expand auto-generated assets and push spend toward upper-funnel placements, understanding when and how to deploy it is more relevant than ever.

What a Feed-Only Performance Max Campaign Actually Does

Performance Max draws from two fuel sources. Text, images, and video assets push the system to serve ads on YouTube, the Display Network, and Gmail. Your Google Merchant Center feed powers Google Shopping placements and targets bottom-funnel, high-intent queries.

When you remove every creative asset, the product feed becomes the only viable signal the algorithm has to work with. The campaign effectively becomes a Shopping-dominant engine. Ads are served exclusively in the Shopping Network and dynamic remarketing, in contrast to classic PMax campaigns which use all available channels such as Search, Display, YouTube, or Gmail.

Think of it as the spiritual successor to Smart Shopping. This method is very similar to the now-discontinued Smart Shopping campaign type and is often used to reproduce its functionality as closely as possible. Practitioners who remember the pre-2022 era, where Shopping and Display ran in a contained loop without the sprawl of YouTube and Gmail placements, will recognize the appeal. One critical distinction: feed-only does not guarantee Shopping-only anymore. A pure Shopping-only experience is harder to guarantee today. Since late 2023, Google has used feed images to auto-generate YouTube and Display assets, including interactive product tiles and Shorts placements. This means your feed images can appear in places you didn't intend. The realistic framing is "Shopping-dominant," not "Shopping-exclusive."

When Feed-Only Is the Right Call (and When It Isn't)

Feed-only PMax isn't a default campaign type. It is a precision tool, not a default campaign type. You should deploy it under specific conditions. Deploy feed-only when: - Your primary objective is to concentrate budget strictly on high-intent Shopping placements rather than upper-funnel awareness.

• Your brand lacks strong video or lifestyle imagery, and this setup prevents Google from generating low-quality, automated video ads that can damage brand perception.

• You demand strict isolation between Shopping spend and YouTube or Display spend because you already run dedicated campaigns on those networks. - An existing standard Performance Max campaign is spending too much in low-value networks and consistently missing target ROAS goals.

• You are hitting that traditional PMax wall of performance stagnation - things seemed to be working, but then you can't move the needle after 9-12 months.

Stick with full-asset PMax when: You have high-quality creative (video, lifestyle imagery, polished ad copy) and want Google's AI to find customers across every touchpoint. All-assets PMax campaigns have a broader reach and target audiences at all points of their customer journey, leading to more conversions over time. You also introduce your brand to completely new shoppers who weren't aware of your offerings.

A smart middle ground exists. You can use a full Performance Max setup for your bestsellers and a Shopping-only setup for all other products. This hybrid approach concentrates creative investment where it matters most while keeping long-tail products focused on Shopping efficiency.

Three Ways to Build a Feed-Only Campaign in 2026

The setup process isn't as straightforward as it used to be. Some guides explain that you can create a feed-only campaign by simply not adding any assets. If you have ever tried this, you will have noticed that this is no longer possible in most Google Ads accounts, because you have to store at least 3 headlines to save the asset group.

Here are the three methods that actually work.

Method 1: Create via Google Merchant Center

This is the cleanest workaround. Create the campaign directly via the Google Merchant Center. This allows you to bypass restrictions and set up a real feed-only campaign that only uses your product feed.

The steps are simple: navigate to Ad Campaigns in Merchant Center, select New Campaign, set your daily budget, and publish. Your campaign will start showing eligible products from your Merchant Center account. It may take up to 48 hours before your ads start showing.

One limitation matters here. You can only create one asset group this way, meaning that feed-only PMax campaigns created through Merchant Center are limited to a single asset group.

Method 2: Google Ads Editor Workaround

This method gives you more flexibility. Start by creating a standard PMax campaign with full assets in the Google Ads UI. Then open Google Ads Editor, locate the asset group section, delete the asset groups created during initial setup, and add a new empty asset group. Strip out all headlines, descriptions, images, and videos.

Method 3: Duplicate and Strip

If your feed-only asset group isn't working, try duplicating an existing asset group - find one that is already running, click the three-dot settings menu to duplicate it, then remove all non-feed creative assets manually.

Regardless of method, complete these post-setup steps: - Confirm the campaign objective is set to Sales and that location targeting is strictly restricted to "Presence" to avoid irrelevant traffic.

• Check that the asset group contains no text, images, or videos, and that both Final URL Expansion and Automatically Created Assets are completely disabled.
• Disable asset optimization, as this keeps more controls so that PMax feed-only will not expand to Display prospecting, Gmail, or YouTube.

That last point deserves emphasis. If you skip disabling auto-generated assets, the system will simply scrape your website and create assets on your behalf. Your "feed-only" campaign becomes a full-asset campaign in disguise.

Your Product Feed Becomes Your Entire Ad Strategy

With no headlines, no descriptions, and no images beyond what the Merchant Center provides, your product feed isn't just important - it's everything. The quality of the product data - primarily title, description, and images - is crucial for success.

Titles Are Your Highest-Leverage Optimization

Optimized product titles move more CTR than any other feed change. Google's algorithm matches Shopping queries to product titles before considering any other signal.

Google allows up to 150 characters, but in most cases only the first ~70 characters display. That means the front half of your title is prime real estate. Structure matters. Lead with the attribute that matters most for your product category - brand for recognized names, product type for generic brands. A concrete example: "Blue Shirt" will lose to "Men's Blue Oxford Button-Down Shirt" every time. The second title answers the query, matches long-tail searches, and signals precisely what the product is. Add the most important details of your product in the first 70 characters, because shoppers often see only that portion depending on their screen.

For stores with large catalogs, use feed rules in Merchant Center to programmatically construct better titles from multiple data fields - concatenating brand, product type, color, and size attributes. This approach scales title optimization across thousands of SKUs without manual editing.

Images Now Have Downstream Consequences

This is a point many practitioners miss. With Google auto-generating video-like experiences from feed images, poor product titles, inconsistent image cropping, missing lifestyle imagery, and weak variant photography don't just hurt Shopping performance - they can now hurt motion asset quality too.

For Shopping ads specifically, clean white-background product images still outperform lifestyle shots. White-background images with multiple angles boost CTR by 25%. But consider adding the lifestyle_image_link attribute as well, because Google may pull those images for dynamic remarketing placements where context matters more.

Custom Labels: Your Segmentation Superpower

Custom labels are one of the most powerful - and underused - features in Google Shopping. While most advertisers rely on product type or brand for campaign structure, custom labels let you segment products based on your own business logic: profit margins, performance history, seasonality, or any attribute that matters to your strategy.

Google Merchant Center provides five custom label slots (custom_label_0 through custom_label_4). Practical implementations that move the needle:

• Margin tiers:

One e-commerce brand separated SKUs into "high-margin," "mid-margin," and "low margin" using custom_label_0 based on historical conversion values. They ran three Performance Max campaigns with tiered target ROAS goals, resulting in a 96% increase in ROAS for high-margin or bestselling products.

• Performance buckets: Tag products as Heroes, Sidekicks, Villains, or Zombies based on conversion data.

Segmenting out your top products ensures they get enough budget, and you can reduce your ROAS targets to give them particular pushes at key times of the year.

• Price competitiveness:

Isolate products that are cheaper than your competitors' products. Use price scraping tools to analyze your price vs. the market and label products accordingly.

Keep each label to 3-7 distinct values. Creating too many product groups can spread your data too thin, making it difficult for both you and Google's algorithms to make statistically significant decisions.

Bidding Strategy and Budget Calibration

Feed-only campaigns pair best with Maximize Conversion Value with a Target ROAS. Without the optional inputs like Target ROAS, you are giving Google's AI more freedom to bid aggressively, which can drive significant volume but at the cost of less control over ROAS.

Set your Target ROAS based on actual historical Shopping data, not an aspirational number. If your Shopping campaigns historically deliver 400% ROAS, start your feed-only PMax at 350-400% and adjust after two to four weeks of data collection. Budget matters more than you might expect. Wait 6 weeks after launching any PMax campaign - this time allows machine learning algorithms to gather sufficient data. For feed-only campaigns specifically, start with enough daily budget to generate at least one conversion per day. Too little budget starves the algorithm of the success signals it needs to stay within high-intent Shopping auctions. PMax requires a steady stream of "success" signals to stay within the Shopping auction.

For accounts with sufficient conversion volume, consider adding audience signals - but do so thoughtfully. A campaign that combined feed-only assets with audience signals consistently delivered the highest ROAS, confirming that this hybrid approach offers the best results. Upload customer match lists and website visitor segments. Campaigns using customer match see 50% faster optimization and 25% higher ROAS on average.

Monitoring: What to Watch and What Breaks

Google's new channel performance reporting is now available across all Performance Max campaigns. Channel performance reporting now extends to all campaigns, providing visibility into where ads run across Google's network. Before November 2025, only select beta accounts had this. Now it's standard.

For feed-only campaigns, this report validates your setup. If your Shopping channel accounts for 90%+ of spend, the configuration is holding. If you see significant Display or YouTube spend creeping in, something has gone wrong. Common failure points to audit: - If your campaign is suddenly generating text ads or spending heavily on video, check these failure points: verify that both Final URL Expansion and Automatically Created Assets are off.

• If your feed lacks depth, the system may overcompensate by pushing spend into the Display network to find volume.

• If purchase tracking is broken, the campaign will stall. Check your conversion tag and GA4 integration weekly. Use the "Views" metric as an early warning system. Watch the Views metric to detect video creep and use scripts to monitor network distribution. A feed-only campaign should show negligible view counts. If views spike, Google is auto-generating video assets from your feed images. Feed health checks that prevent performance decay: - Daily synchronization is ideal to keep prices and inventory accurate.

• Eliminate "zombie" SKUs - products bringing high cost per conversion and low conversion value - to reduce average CPA significantly without sacrificing scale.

• Run a weekly SKU-performance export.

The 80/20 rule applies: optimize your top 50 products. Stop trying to optimize 500.

Advanced Segmentation: Beyond the Single-Campaign Setup

Once your feed-only campaign is stable and generating data, segmentation unlocks the next tier of performance. The tiered campaign architecture works like this: create separate feed-only campaigns for different product segments using custom labels. Your hero products get their own campaign with aggressive ROAS targets and uncapped budgets. Mid-tier products run at standard targets. Long-tail and zombie products either get their own low-budget campaign or are excluded entirely.

Don't think about segmenting your campaigns if you don't have sufficient conversion volumes - less than 50 conversions per month makes segmentation premature. The algorithm needs data density to optimize within each segment. The full-funnel split is another approach gaining traction. Create a Performance Max campaign without a product feed (asset-only) focused on generating new demand and targeting upper-funnel users. In parallel, use a feed-only campaign for dynamic remarketing that targets users who have already shown interest. This way, you can better control how Google uses your budget for different customer journey phases.

This architecture gives you what a single PMax campaign cannot: true budget separation between acquisition and conversion. The asset-only campaign handles awareness on YouTube and Display. The feed-only campaign handles the close on Shopping and remarketing. Each campaign's ROAS target reflects its role in the funnel.

The Longevity Question: Will Feed-Only Keep Working?

Every practitioner using this approach should stay honest about one thing: this setup was not technically built to run this way, and its longevity depends on what Google allows in the coming years.

Google is already expanding what "feed-only" means. Google now has the ability to serve auto-generated ads based on images from your product feed on YouTube and potentially also on Display. The ad format will rely on images rather than videos, with slight interactive elements when displaying product tiles. The gap between feed-only and full-asset PMax is narrowing.

Google is already testing AI-generated video and image creation directly within Performance Max. By late 2026, campaigns are expected to automatically generate additional creative variations based on performance data. The direction is clear: Google wants every campaign to fill every placement. That doesn't make feed-only irrelevant. It makes it a deliberate choice rather than a permanent one. Use it to establish a Shopping performance baseline. Use it when your creative assets aren't strong enough to compete on YouTube or Display. Use it as the conversion-focused leg of a dual-campaign architecture. But monitor Google's updates quarterly, and be prepared to adapt. The stores that win aren't the ones who found a single setup and stopped thinking. They're the ones who treat campaign architecture as a living system - testing, validating, and adjusting as the platform evolves. Feed-only Performance Max, properly deployed and properly monitored, remains one of the most effective tools for forcing Google's automation to prioritize what e-commerce advertisers care about most: product-level, purchase-intent conversions that deliver real ROAS.