For ecommerce on Google, your product feed is the campaign

Last week's piece on Meta ended with an uncomfortable point. This week's piece on Google ends in a related but different one.

On Meta, the algorithm has become so good at finding buyers that audience targeting barely moves the needle, and creative is the only lever left that still scales. On Google, the algorithm has become so good at matching intent to product that your account structure barely matters, and the product feed is the only lever left that still scales.

Most brands, and most agencies, still act like the campaign is the game. It is not. The campaign is the frame the feed sits in. The feed does the actual work.

What Google actually shows people now

For most ecommerce brands, the majority of Google Ads spend flows through Shopping Ads and Performance Max, not text Search ads (Store Growers). Both campaign types are, at their core, product-feed campaigns. What Google shows the searcher, whether that is a Shopping unit at the top of the SERP or a product listing inside a Performance Max placement, is data pulled directly from your Merchant Center feed. The image is from the feed. The title is from the feed. The price is from the feed. The brand line is from the feed.

Which means the ad is the feed. There is no separate creative asset being written or designed for each product. What Google sees, and what the buyer sees, is what you wrote in the feed a month ago and forgot about.

This has been true for years. It has become the whole game since Performance Max started swallowing Shopping and Google started matching search queries directly to product titles inside Shopping campaigns (Google).

Why the feed does most of the work

There are three mechanical reasons for this, and understanding them changes how you spend your time.

Google matches queries to product titles. When someone searches "waterproof hiking boots size 11," Google's Shopping algorithm looks first at your product titles to decide whether to show the product. If the title reads "Boots, Merlin Series, Item 2043" you will not appear. If it reads "Merlin Waterproof Hiking Boots, Size 11 US, Men's Trail" you will. This is not just relevance for the buyer. It is literally the mechanism Google uses to decide whether to enter you into the auction at all.

The word order in a title has a cost. Product titles get truncated in Shopping placements. What is at the front of your title is what the buyer sees. What is at the back may as well not exist. Brands that put the SKU code at the front and the actual product description at the back leak money on every single impression.

The feed is the source of truth for everything downstream (Google). Merchant Center is where product data lives, where errors get flagged, and where every campaign type pulls from. A single incorrect GTIN, or a mismatched price between the feed and the site, and Google either disapproves the product or quietly lowers its visibility. The feed sets the ceiling for what the account can do.

The parts of the feed that matter most

Three feed attributes carry most of the weight, and each of them is worth an afternoon of proper attention.

Product titles. The most important attribute in the entire feed, and the one most commonly generated automatically from a poorly-structured product database. Rewrite them the way a customer would search. Brand at the front, product name, then the specific attributes a buyer types when they are ready to buy: size, colour, material, use case.

Product identifiers. GTIN, MPN, and brand. When your GTIN matches other sellers of the same product, Google has more data about which queries convert for that product and is more likely to show yours (Store Growers). Missing or wrong identifiers is one of the fastest ways to make your entire catalogue invisible.

Custom labels. These are the feed attributes that let you segment products by anything you choose: margin, bestseller status, seasonality, new arrival, on sale, stock level. Custom labels are how you move from "spend the same on every product" to "spend more on the products that make money," and they are largely unused in most accounts.

Get these three right and the account has a completely different shape. Skip them and every campaign structure decision downstream is applied to a broken foundation.

The account structure conversation that actually matters

Once the feed is in shape, there is a real decision to make about how to run it.

Standard Shopping gives you visibility and control. You can see which search queries triggered which products, exclude the bad ones, prioritise specific product segments, and use campaign priority as a query-sculpting tool. It is more work, and for accounts under about 30 conversions a month, it is often the right call.

Performance Max is the automation option. It runs Shopping plus Search plus YouTube plus Discover plus Display plus Gmail plus Maps all at once, driven by a mix of your feed and any asset groups you upload. When it works, it works impressively. When it does not, it burns budget on the placements you cannot see. The middle path most experienced accounts settle on is a feed-only Performance Max campaign, where the only assets provided are the product feed. That keeps Performance Max operating close to a Shopping campaign with the extra intent signals it has access to, without opening the door to random Display and Gmail impressions.

What people consistently get wrong

The same patterns show up across nearly every underperforming ecommerce Google account we audit.

They obsess over campaign structure and ignore the feed. Hours spent debating Standard Shopping versus Performance Max. Five minutes spent on the actual product titles. The account is a real conversation. The feed is where the money is left on the table.

They accept Google's auto-recommendations. Google reps have revenue targets, and the account optimisation score is a metric Google invented to pressure advertisers into accepting changes (Store Growers). A low score does not mean the account is performing badly. It means the account has not done what Google wants it to do. Turn off auto-apply. Review recommendations manually. Accept only the ones that make sense.

They chase ROAS in Ads Manager and ignore the actual economics. The same trap as Meta. A 4x ROAS on a 20% margin product is not a win. Contribution margin, blended CAC and LTV are the scoreboard. Google's dashboard tells a story that may or may not match the bank account.

They don't use custom labels. Every ecommerce account has product-level economics that vary wildly. Some products are 40% margin. Some are 8%. Some clear stock. Some anchor the brand. Treating them all identically in the feed is treating them all identically in the bid strategy, which means the account never spends where it should.

They forget that first-party data is the last real advantage. Customer Match, Enhanced Conversions and a clean CRM feed of qualified purchase data are increasingly what separates the accounts that hold ROAS through cookie deprecation from the ones that quietly bleed. Foundation Series piece four covered this, and everything we said there applies here.

The feed does the work most brands leave undone

The Google Ads game for ecommerce has moved upstream. The interesting decisions used to sit inside the ads manager. Now they sit inside Merchant Center. Product titles, identifiers, images, custom labels, feed hygiene, first-party data integration. All boring, all unglamorous, and all where the actual leverage lives. The brands that scale profitably on Google in 2026 are the ones who stopped treating the feed as a technical afterthought and started treating it as the campaign itself.