Google Ads for Fashion: The Feed and Returns Problem No One Tells You About
Fashion ecommerce returns a significant portion of units sold online, which means most brands running Google Ads optimize toward products that come back. You hit a 4x ROAS on the dashboard and lose money once fulfillment and return shipping eat the margin. The fix is not creative or bidding. It is a feed problem and a conversion value problem.
This guide walks through the feed structure, campaign separation, and returns-adjusted tracking that let apparel brands scale Shopping and Performance Max without training Google to bid high on their worst SKUs. No fluff. Just the mechanics that decide whether Search and Shopping are profitable or just busy.
Why Google Ads works for fashion ecommerce
Search and Shopping capture bottom-funnel intent when someone types "black midi dress" or "men's linen blazer". That intent converts because the searcher already decided to buy. Your job is to show it to them before a competitor does.
Performance Max and Shopping put product images, pricing, and availability directly in the SERP before the click. Someone searching for "navy linen shirt" can see five options, compare prices, and click the one that matches their aesthetic without reading a single word of ad copy.
Google can represent 30 to 40 percent of online revenue for scaled apparel brands because it captures existing demand without creative production. No photoshoots. No hooks. No testing ad copy for six weeks. You upload a feed, Google matches queries to products, and the traffic converts.
The problem is profitability once you account for returns. Fashion ecommerce often returns 18 to 25 percent of units sold online. If you optimize to gross order value, Google learns to bid higher on products that generate revenue but come back. You hit a 4x ROAS on the dashboard and lose money once fulfillment and return shipping are deducted.
The product feed decides everything
Shopping and Performance Max read your catalog through the feed, not campaign settings. Your feed is your campaign. If the feed is broken, no amount of bid optimization will fix it.
Fashion and apparel require six mandatory attributes: color, age_group, gender, size, item_group_id, and brand. Missing any of these and Google either rejects the product or shows it for the wrong queries. Footwear additionally requires GTIN or MPN. Google will reject the entire feed if you skip this for shoes.
Variants must be grouped under a single item_group_id so Google shows one product card with selectable options. If you upload each size and color as a separate product without grouping them, you waste impression share competing against yourself.
Poor feed hygiene means Google shows your navy shirt for "black shirt" searches. Someone clicks, sees the wrong color, bounces. You pay for the click. Google's algorithm learns nothing useful.
Title structure matters. Brand plus type plus attributes works better than vague marketing copy. "Everlane Cotton Crew Neck T-Shirt Navy Medium" outperforms "The Perfect Tee for Every Occasion". Descriptive titles match more relevant queries. Marketing fluff matches nothing specific.
Free PDF · the crash course
AI Agents for Marketing Crash Course
Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.
Campaign structure that separates signal from noise
Brand and non-brand must be split into separate campaigns because they have wildly different ROAS and margin profiles. Brand searches often convert at 8 to 12x ROAS. Non-brand Shopping typically sits at 3 to 4x. Blending them means Google allocates budget based on blended performance, and you lose visibility into what actually drives profit.
Standard Shopping campaigns give you control over which products get budget. Performance Max does not. If you need to prioritize hero products or suppress clearance inventory, Standard Shopping lets you set product group bids manually. Performance Max decides for you based on predicted conversion probability.
Best-sellers and new drops deserve dedicated Performance Max campaigns with higher ROAS targets. Long-tail SKUs and clearance inventory go into separate low-priority Shopping campaigns. A clearance item that converts at 1.5x ROAS will train Performance Max to find more low-margin traffic if you let it live in the same campaign as your hero products.
Dynamic Search Ads work only if your site has clean, descriptive product pages. If your product page title is "SS26 Collection Drop" instead of "Women's Linen Wide-Leg Pants Olive Green", Dynamic Search Ads will match the wrong searches.
A typical structure for google ads for fashion ecommerce: one brand Search campaign, one non-brand Shopping campaign using standard Shopping, one Performance Max for hero products, one low-priority Shopping for clearance. That separation keeps signal clean and lets you set different ROAS targets for different inventory tiers.
Returns-adjusted conversion values or you optimize toward your worst SKUs
Fashion ecommerce often returns 18 to 25 percent of units sold online, which means raw revenue ROAS lies. If you optimize to gross order value, Google learns to bid higher on products that generate revenue but come back. You train the algorithm to prioritize high-return SKUs.
Returns-adjusted conversion values pass net revenue after returns back to Google so the algorithm learns true profitability. This requires integrating return data into Google Ads conversion tracking, not just order confirmation. Most brands only send the order value at checkout. Google never sees the return two weeks later.
Brands that implement this see Performance Max shift spend toward lower-return categories without manual exclusions. The algorithm figures out which products stick. You stop bidding high on items that come back at 30 percent.
Without it, you hit a 4x ROAS on the dashboard and lose money once returns and fulfillment costs are deducted. A $100 order at 25 percent return rate is actually $75 in kept revenue. If your blended cost per acquisition is $25, your real ROAS is 3x, not 4x. The difference is whether the campaign is profitable.
Implementing returns-adjusted values means tagging each conversion with order ID, tracking returns in your backend, and sending a negative conversion value back to Google when the return processes. Most analytics platforms do not do this by default. You have to build it.
Performance Max for fashion: what it actually does
Performance Max combines Shopping, Display, YouTube, and Discovery into one campaign driven by your feed and creative assets. For fashion, it works best when given a clean feed, high-quality lifestyle images, and a tight item_group_id structure. Feed quality matters more in PMax than in Standard Shopping because you have less manual control.
Google's algorithm decides where to show your products based on predicted conversion probability. If YouTube converts better than Display for your dresses, PMax shifts budget there automatically. You do not pick placements.
Asset groups let you create different creative combinations for different product categories. Dresses get lifestyle imagery with models. Outerwear gets flat lays and detail shots. You can segment by product type and control the creative strategy per segment without building separate campaigns.
The lack of search term visibility means you cannot exclude bad queries the way you can in Standard Shopping. You rely entirely on feed quality and negative keywords at account level. If PMax is spending on irrelevant traffic, the fix is almost always in the feed titles or the product_type attribute, not in campaign settings.
PMax scales fast but requires strong ROAS guardrails and product-level profit data or it will spend on high-return, low-margin SKUs. The algorithm optimizes to whatever conversion value you send it. If you send gross revenue, it finds more gross revenue. If you send net revenue after returns, it finds more profitable orders.
Search campaigns and when to bother
Standard Search campaigns make sense for branded queries and high-intent product-specific terms like "buy Everlane denim jacket". Someone typing your brand name plus a product already knows what they want. Search lets you control the message and own that traffic before a competitor does.
Non-brand Search in fashion is expensive and often loses to Shopping because shoppers want to see the product, not read ad copy. A text ad that says "Shop women's dresses" performs worse than a Shopping ad showing six dresses with prices.
Exact match brand terms should always run in Search to prevent competitors from buying your traffic. If you do not bid on your own brand, someone else will. Branded Search is defensive spend. The ROAS is high because the traffic already wants you, but you still have to pay to keep it.
Dynamic Search Ads can fill gaps for long-tail product queries if your site structure is clean, but they require aggressive negative keyword lists. Without tight negatives, DSAs will match queries like "free shipping on dresses" or "dress size chart" and waste budget on informational searches that do not convert.
Search works best as a defensive brand layer and a high-intent product supplement, not as the primary acquisition channel for fashion. Shopping and Performance Max do the heavy lifting.
Free PDF · the crash course
AI Agents for Marketing Crash Course
Learn how to deploy AI marketing agents across your go-to-market — the best tools, prompts, and workflows to turn your data into autonomous execution without writing code.
Benchmarks and what profitable looks like
A blended ROAS of 3.5 to 5x can be typical for scaled fashion brands after accounting for returns and discounting. Brand campaigns often hit 8 to 12x ROAS. Non-brand Shopping typically sits at 3 to 4x. Performance Max often lands in the middle at 4 to 6x. If your blended ROAS is below 3x, either your feed is broken or your conversion values are not adjusted for returns.
Click-through rates on Shopping ads for apparel often run 1 to 2 percent, higher for hero products with strong imagery. Low CTR means your product images or titles are weak compared to competitors in the same auction. The fix is better photography or more descriptive titles, not higher bids.
Cost per click varies wildly by category. Basics and activewear are cheaper than dresses and outerwear. CPCs for "black t-shirt" might be $0.40. CPCs for "wool coat women's" might be $1.80. The competition and margin profile differ by category, so acceptable CPA differs too.
Return rates often in the 18 to 25 percent range mean your acceptable ROAS must be at least 1.3x higher than a category with 5 percent returns to break even. If a home goods brand can be profitable at 3x ROAS, a fashion brand needs 4x to 4.5x just to hit the same margin after returns.
The brands that scale profitably treat Google Ads as a margin and merchandising lever, not a traffic volume game. They know which SKUs return at what rate. They segment inventory into profitability tiers. They set different ROAS targets per tier. They feed returns data back into the conversion tracking.
What breaks and how to fix it
Feed errors are the most common campaign killer. Missing required attributes, duplicate item_group_ids, or inconsistent color names. Google rejects products or shows them for the wrong queries. You see low impression share and high CPCs because your feed is competing in auctions it should not be in.
Performance Max going rogue usually means poor feed hygiene or no returns-adjusted conversion values guiding the algorithm. PMax optimizes to whatever signal you give it. If the signal is noisy, the spending is noisy. Fix the feed first, then check your conversion values.
ROAS dropping during a sale or promo period is expected. Discounts compress margin. The fix is to lower targets temporarily or exclude sale inventory from hero campaigns. Sales periods are high-volume, low-margin windows. Adjust the target, let it run, then reset when the sale ends.
High impression share with low CTR means your product images or titles are weak compared to competitors. Google is showing your products. People are not clicking. That is a creative problem, not a bidding problem. Better images, clearer titles, or tighter feed attributes fix this.
Shopping campaigns spending on irrelevant queries trace back to vague product titles or missing negative keywords at account level. If your title is "Floral Dress" instead of "Women's Midi Floral Dress Cotton Summer", Google will match it to queries like "floral tablecloth". More descriptive titles reduce irrelevant traffic.
The fix is almost always in the feed or the conversion value definition, not in bid adjustments or campaign settings. Campaign structure and bids matter, but they cannot overcome a broken feed or a conversion tracking setup that optimizes to the wrong outcome. Start with the data. The campaigns follow.
---
See Graphed in action.
More Google Ads guides
Google Ads for Bail Bondsman: What Changed and What Works Now
Google permanently banned bail bond advertising across all platforms in 2018. Paid search is dead for bondsmen, but organic channels work better than ads ever did.
Google Ads for Balloon Decor: What Works When Organic Reach Runs Out
Balloon decor businesses fail at Google Ads because they treat it like social media. Search intent is different, the customer is different, and the funnel collapses if you send clicks to a gallery instead of a booking form.
Google Ads for Bankruptcy Attorneys: Chapter Targeting, Budget Math, and Urgency Play
Bankruptcy is a volume practice with thin margins. Your Google Ads for bankruptcy attorney campaigns need chapter splits, urgency triggers, and ruthless negative lists to hit cost-per-signed-case targets.
Google Ads for Banks: Campaign Structure and Compliance Requirements
Financial services CPCs run $50 to $150. Banks convert at 5.9% when campaigns target the right geography and products. Wrong structure burns budget on broad keywords that violate compliance rules.
Google Ads for Barbershop: A Step-by-Step Guide to Booking More Chairs
Most barbershop Google Ads campaigns burn money on people buying clippers, not booking cuts. This guide shows you how to target proximity searches, write copy that converts, and track real bookings.
Google Ads for Basement Waterproofing: What Actually Converts
Most waterproofing contractors burn 40-60% of their Google Ads budget on DIY searchers and commercial queries. Here's how to isolate homeowner intent and make the unit economics work.