E-commerce Google Ads

E-commerce advertising stands on two things: the quality of your product feed and how honestly you measure ROAS (return on ad spend). If both are right, Google’s algorithm has clean data to work with and can push budget toward products that actually generate revenue. If either one is wrong, the same algorithm can quietly spend money across hundreds or thousands of products without a clear path to profit.

That is what makes e-commerce Google Ads different from lead generation. In a lead generation account, the main questions are usually keywords, landing pages, and lead quality. In an online store, the account is connected to a moving catalogue: products go in and out of stock, prices change, margins differ, variants multiply, promotions start and end, and some products sell well while others only attract clicks.

This page is the e-commerce-specific deep-dive of my Google Ads management. The goal is not just to “run Shopping ads.” The goal is to build a system where the feed, campaigns, bidding, segmentation, and tracking all support profitable growth.

The product feed: the raw material of the ads

Shopping and Performance Max ads are fueled not by keywords but by the product feed in Google Merchant Center. Google reads the feed, matches products to searches, builds product ads, and decides which items to show. That means the quality of the ad is limited by the quality of the feed.

A weak feed makes strong advertising difficult. A strong feed gives the system more chances to match the right product to the right search. In many e-commerce accounts, improving the feed returns more than adjusting bids, because the feed controls eligibility, relevance, and how the product appears before someone clicks.

The most important fields are simple, but they need careful attention:

• Title — the most important field. It should contain searched terms, usually including brand, model, product type, and key attributes. For example, a title like “Running Shoes” is much weaker than a title that includes the brand, model, gender, size range, material, or defining feature where relevant. The title should not be spammy or stuffed with repeated keywords. It should read like a clear product description that also reflects how customers search.
• Category (Google product category) — the category helps Google understand what the item is. A wrong category can lead to showing on wrong searches, poor matching, or reduced performance. A product placed in a broad or incorrect category may compete in the wrong auction, attract irrelevant clicks, or fail to appear where it should.
• Image, price, stock, GTIN — these fields affect both approval and performance. Missing or wrong data can get products disapproved, limited, or held back. The image must match the product. The price in the feed must match the landing page. Stock status must be accurate. GTIN and other identifiers help Google understand the exact product and compare it properly.

Feed optimization is advertising work, not just a technical cleanup. If a product title does not include the terms buyers use, the product may never enter the right auctions. If prices are inconsistent, products may be disapproved. If stock data is outdated, the budget can go to products that cannot be purchased. If variants are unclear, users may click expecting one item and land on another.

A clean feed also makes later campaign decisions more reliable. When products are accurately described and categorized, performance data becomes easier to trust. You can see which products sell, which categories have demand, and where the budget is being wasted.

Campaign structure: Shopping, Performance Max, Search

E-commerce campaigns need structure. Without structure, Google can still spend, but you lose control over what the budget is doing. The right setup depends on the size of the store, margins, product range, brand demand, and available conversion data, but the core channels usually work together.

• Performance Max — today the backbone of e-commerce; it combines Shopping inventory with Search, Display, YouTube and Gmail. Powerful, but must be set up correctly.
• Separate the brand campaign — to stop PMax from eating branded traffic you’d already win, run a separate brand campaign and use brand exclusions in PMax.
• Search — complementary for high-intent, non-category searches.

Performance Max can be very effective because it gives Google access to multiple placements and formats. It can show product ads in Shopping results, reach users across other Google inventory, and use conversion data to find likely buyers. But that flexibility is also the risk. If it is not configured properly, it can lean heavily on easy branded traffic, push products with low profitability, or hide where the value is really coming from.

That is why brand traffic needs special handling. A user searching for your store name is different from a user searching for a generic product. Branded searches often convert well because the person already knows you. If Performance Max receives credit for too much branded demand, the campaign may look more profitable than it really is. Separating the brand campaign helps show the difference between demand you already have and demand the ads are creating or capturing competitively.

Brand exclusions in Performance Max are part of that control. They reduce the chance that PMax simply takes credit for branded searches and makes prospecting performance look stronger than it is. This does not mean brand advertising is unimportant. It means brand and non-brand should be understood separately, because they answer different business questions.

Search campaigns still have a place. They are useful for high-intent searches that do not fit neatly into category Shopping traffic, for specific product problems, competitor comparisons where appropriate, or queries where text ads give more control over the message. Search can also support products or categories where Shopping alone does not explain the offer clearly enough.

The structure should make performance easier to read, not harder. A good account shows where money is going, what kind of demand is being captured, and which parts of the catalogue deserve more budget.

Value-based bidding (tROAS)

For an e-commerce store, a conversion is not just “a sale.” One order may be small and low-margin. Another may be large and profitable. If the account only tracks conversion counts, Google sees both as equal. That is not enough for serious e-commerce advertising.

Once enough conversion-value data accumulates — real sale amounts, not just conversion counts — we can move toward Target ROAS bidding. Target ROAS tells Google to optimize for revenue value relative to ad spend, rather than simply chasing the highest number of purchases.

But the target ROAS should not be chosen out of thin air. It should be calculated backwards from profit margin.

That means looking at:

• the gross margin of the products or categories;
• return rate, where relevant;
• shipping or fulfilment considerations if they materially affect profitability;
• the target profit the business needs;
• the difference between break-even ROAS and the ROAS that makes the account worth scaling.

For example, a store with high-margin accessories can often afford a different ROAS target from a store selling low-margin electronics. A best-selling product may deserve aggressive bidding if it produces repeat customers or has strong margin. A product with thin margin may need a stricter target, even if revenue looks attractive on the surface.

This is where the “honest” part of ROAS matters. Revenue is not profit. A campaign can show strong revenue and still fail commercially if margin is low, returns are high, or ad spend is too aggressive. The target has to reflect the business model, not just the advertising platform.

My math background applies directly here. Rather than guessing a number that “feels right,” I connect the advertising target to commercial reality: margin, return rate and target profit → break-even ROAS → target ROAS.

This only works if conversion-value tracking is correctly set up. The account must receive accurate order values. It must avoid double-counting. It must use the right attribution and conversion actions. If tracking is wrong, automated bidding learns from wrong data. Before trusting Target ROAS, the measurement foundation has to be checked carefully: conversion tracking.

Segmentation: where the budget goes

Not all products are equal. Treating the whole catalogue as one block is one of the easiest ways to waste e-commerce ad spend.

Some products are best-sellers. Some have high margins. Some attract clicks but rarely convert. Some sell only in season. Some are useful as entry products but not profitable as direct ad targets. Some should not be advertised at all because they are out of stock, priced poorly, or unable to compete.

Segmentation means separating products so budget can be directed intelligently. I separate high-margin and best-selling products from underperformers, then shift budget toward the profitable side. This can be done by product groups, custom labels, categories, campaigns, asset groups, or listing groups depending on the account structure.

The practical goal is simple: profitable products should not be held back by weak products, and weak products should not drain budget from the winners.

Examples of useful segmentation include:

• grouping products by margin level;
• separating best-sellers from long-tail catalogue items;
• isolating seasonal products during the period when demand is active;
• excluding out-of-stock items so users do not click into dead ends;
• pushing promotional products when price or offer strength is temporarily better;
• separating categories with different buying behaviour;
• identifying products that get many clicks but no meaningful revenue.

Out-of-stock handling is especially important. If the feed says a product is available when it is not, the store pays for traffic that cannot convert. Even when a similar product is available, the user experience is weaker. Accurate stock data protects budget and trust.

Seasonality also matters. A product that performs badly for most of the year may become important during a specific buying period. A campaign structure should allow those products to be pushed when demand is present and reduced when demand fades. The same applies to promotions: if a category has a temporary price advantage, the account should be able to react.

Segmentation also makes reporting more useful. Instead of saying “Shopping is up” or “PMax is down,” we can ask better questions: Which categories are profitable? Which products deserve more exposure? Which items bring revenue but not profit? Which products should be removed from paid promotion?

Together with SEO

In the same store, e-commerce SEO runs alongside Google Ads. They are different channels, but they should not operate in isolation.

Google Ads gives fast feedback. It can show which products people search for, which terms lead to sales, which categories convert, and which offers are ignored. That data is useful beyond the ads account. It can guide SEO priorities for category pages, product content, internal linking, and content planning.

For example, if paid search data shows that a certain product attribute consistently appears in converting queries, that attribute may deserve clearer placement in page titles, category copy, filters, or product descriptions. If a category converts well in ads but has weak organic visibility, it may be a good SEO priority. If a product receives many clicks but poor conversion, the issue may be pricing, page content, availability, reviews, or product-market fit.

SEO works in the other direction too. Strong organic pages can improve the overall store experience, clarify product categories, and reduce reliance on paid traffic over time. Organic traffic builds a longer-term asset, while Ads brings sales quickly and provides faster testing.

Running both in one hand is the real advantage: the product and keyword data that Ads surfaces feeds directly into SEO category and content priority, with nothing lost between channels.

Measurement first, then scaling

E-commerce advertising should not be scaled just because an account has traffic or revenue. It should be scaled when the feed is clean, the tracking is reliable, the campaign structure separates important traffic types, and bidding is connected to profit.

The usual order is:

1. Fix the measurement foundation.
2. Clean and improve the product feed.
3. Structure campaigns so performance can be understood.
4. Segment products by commercial value.
5. Use conversion value and ROAS targets based on margin.
6. Scale what is profitable and reduce what is not.

This approach is slower than simply launching a Performance Max campaign and increasing the budget, but it is safer. It gives Google better inputs and gives the business better decisions.

For the measurement foundation see conversion tracking; for overall strategy, Google Ads. If your store’s advertising is not profitable, get in touch — a quick look at the feed and account and I’ll tell you where to start.