Contextual advertising for an online store selling shoes

The architecture of Google Merchant Center: how a product feed determines profitability in the footwear niche

Google Merchant Center is the foundation of Shopping and Performance Max. The XML feed must clearly communicate to the algorithm exactly what is being sold: model, brand, color, material, gender, size, availability, and price. Errors in microdata, GTIN, prices, or shipping details can lead to Merchant Center being blocked, which instantly halts the main flow of e-commerce traffic.

Grouping variations and detailing attributes

• Grouping via item_group_id combines sizes and colors for a single model. Without this, Google performs poorly on specific queries like «red sneakers size 42» and may display irrelevant variations.
• The attributes size, color, material, and gender are required for filters in paid search results and organic presence in Shopping. The more accurate the data, the higher the relevance, CTR, and chance of conversion.

We link PPC with SEO data from the product listing: the title, description, microdata, and feed should all speak the same language. This helps algorithms match products more accurately with search queries and reduces the risk of discrepancies in Merchant Center.

Margin, Seasonality, and Inventory Synchronization

Custom Labels allow you to segment the catalog by margin, season, ABC group, and size range completeness. High-margin winter boots cannot be optimized with the same Target ROAS as summer sandals on sale. For the Ukrainian, US, and EU markets, this segmentation saves significant budgets, as the bid depends on the product’s business value.

A separate risk is wasting the budget on unavailable sizes. If popular sizes 38–39 or 42–43 are sold out, traffic to that product listing will not convert. Therefore, we set up rules to exclude products with an «incomplete» size range and monitor the Out-of-Stock Click Rate.

Performance Max strategy for footwear e-commerce

You shouldn’t «feed» the entire catalog to a single Performance Max campaign. Machine learning optimizes for the easiest conversions: low-cost products, remarketing, or brand-specific queries. The Senior Team’s task is not to interfere with automation, but to properly limit it within the bounds of profitability.

Campaign segmentation, target ROAS, and brand exclusions

It’s advisable to segment campaigns by brand, category, margin, and average order value. Otherwise, PMax might sell shoelaces or insoles, showing a high ROAS but failing to drive business growth. Brand Exclusions are necessary to ensure the campaign doesn’t capture cheap branded traffic where the store already has a strong SEO position but instead focuses on acquiring new customers through New Customer Acquisition.

What we check during the audit

During the audit, we look for budget «leaks»: brand cannibalization, excessive remarketing, low-quality platforms, and spending imbalances between Search, Display, and Video. For PMax, we use technical reports and scripts to reduce the «black box” effect.

Visual assets based on UGC and lifestyle

Catalog photos on a white background perform poorly in Display, Discover, and YouTube. Footwear requires lifestyle shots: shoes on feet, in motion, in real-life settings. UGC content, reviews, and short videos often yield a higher CTR, so we immediately establish requirements for the Assets Group based on a specific budget and target ROAS.

Search advertising and Demand Gen: capturing existing demand and generating new demand

Search Ads targets established demand: «buy leather boots», «women’s Nike sneakers size 39». Demand Gen operates higher up in the funnel generating interest in collections, brands, and seasonal categories. It’s important not to pause these campaigns due to Last Click bias: they often create the first touchpoint before a purchase.

DSA, negative keyword matching, and dynamic remarketing

1. DSA captures low-frequency traffic based on models, SKUs, colors, and sizes that can’t be fully covered manually.
2. Cross-exclusion filters out queries like «sole repair», «how to clean suede», and «wholesale» so the budget isn’t wasted on informational traffic.
3. RLSA increases bids for users who abandoned their cart and are searching for similar shoes on Google again.

Dynamic remarketing for abandoned carts brings back the most engaged audience, while transparent PPC analytics show exactly where every hryvnia is spent.

Demand Gen: LTV Audiences and Shorts

First-party data is becoming the foundation of targeting. We upload lists of high-LTV customers such as those who have purchased two or more pairs and build lookalike audiences to attract new, high-spending buyers. Vertical videos for YouTube Shorts featuring unboxing, reviews, or model styling align with mobile consumption patterns.

Analytics: Consent Mode v2 and Smart Bidding

Without accurate analytics, Smart Bidding loses its bearings. Cookie restrictions, iOS 14+, cross-device paths like «viewed on iPhone purchased on PC», and incomplete attribution make standard tracking insufficient. Therefore, Consent Mode v2 and Enhanced Conversions must be implemented before scaling the budget.

Hashed data and micro-conversions for training algorithms

Enhanced Conversions send hashed SHA-256 customer data, such as email or phone number, to Google Ads and help recover some of the lost signals. If a store has fewer than 30 sales per month, we add micro-conversions: viewing the size chart, adding to cart, and starting checkout. This supports algorithm training without false optimization.

We measure profit, not revenue

For mature accounts, we set up server-side tracking and Profit Bidding: ads receive not only revenue but also margin after cost of goods sold. In reports, we separate New and Returning customers, calculate NC-ROAS and ROMI accounting for returns after 14–30 days, and track Search Impression Share in the most profitable categories.

UX/UI requirements for a website to convert paid traffic

Google Ads brings in users, but the website makes the sale. Before launching, we conduct a landing page audit: our specialists take the business’s preferences into account while ensuring compliance with CRO requirements. In the footwear industry, even high-quality traffic won’t pay off if the user doesn’t understand sizing, shipping, returns, or alternatives.

Checklist for a high-converting footwear product page

• An interactive size chart on the first screen with instructions in centimeters next to the «Buy» button.
• Transparent return and exchange policies that alleviate the fear of making a mistake for cold traffic from PMax.
• Cross-sell and Up-sell blocks: «What to wear with», «Similar models», «Other colors».
• A fast shipping widget, such as «Ship via Nova Poshta today».
• Core Web Vitals: heavy images should load quickly, preferably within 2.5 seconds on mobile internet.

Once the feed, analytics, and landing pages are ready, Google Shopping, Merchant Center, and Ads can be launched in as little as 3 business days within clear, fixed-price packages.

Why delegating Google Ads to a senior team is an investment

Advertising for an online shoe store is a complex system: product feeds, Custom Labels, PMax, Search, Demand Gen, Consent Mode v2, enhanced conversions, server-side tracking, and CRO. A mistake in just one component can undermine the profitability of the entire account.

WEDEX works with large budgets in the Ukrainian, US, and EU markets, providing 24/7 access to accounts, direct communication with a specialist, a personal account manager, and monthly reporting. We start with a free audit of your current campaigns and evaluate not just a «pretty ROAS», but CPA, ROI, POAS, and actual sales scaling.