Feed-Driven Campaign Segmentation: Beyond Custom Labels

Most Google Shopping guides cover custom labels for campaign segmentation. Custom labels are valuable, but they're one of five feed attributes you can use to structure Shopping campaigns. The other four — product_type, brand, google_product_category, and price — are native feed fields that create segmentation opportunities without requiring any additional label configuration.

The key insight is that your campaign structure should reflect your catalogue's commercial reality: which products have different margins, different conversion rates, different competitive landscapes, and different seasonal patterns. Feed-driven segmentation is the mechanism that connects your product data to your campaign structure.

Segmentation dimension 1: product_type

The product_type field is your own category hierarchy — you define the values. In Google Shopping campaigns, product groups can be split by product_type, which means your category structure directly enables campaign segmentation at whatever granularity you set in your feed.

The practical implication: a well-structured product_type hierarchy is an asset beyond catalogue management. A product_type of "Mens > Jackets > Wax Jackets" creates a Shopping product group that's segmentable at three levels: Mens (top), Jackets (mid), Wax Jackets (specific). Each level can have its own bid.

Segmentation dimension 2: brand

Brand is one of the most commercially meaningful segmentation dimensions. Different brands have different conversion rates, margin profiles, competitive landscapes, and search demand patterns. Bidding all brands identically treats a premium heritage brand and a budget own-label brand as equivalent — which they're not in any commercial sense.

In Shopping campaigns, product group splits by brand let you set brand-specific bids or ROAS targets. Brands with higher average order values and strong brand search demand can sustain higher CPCs. Brands where you compete primarily on price may need tighter ROAS targets.

Segmentation dimension 3: price tier

Price correlates with multiple performance dimensions: conversion rate (higher price = lower conversion rate), average order value (obviously), and margin potential (varies by category). Price-tier segmentation is useful even without product-level margin data — you know that your £250+ products need to convert at a lower rate than your £25 products to achieve the same ROAS.

Segmentation dimension 4: google_product_category

GPC-based segmentation is the most structurally clean approach for catalogues spanning multiple product categories. A campaign segmented by top-level GPC (Apparel, Electronics, Home & Garden) immediately creates product groups with fundamentally different economics that should be bid differently.

The advantage of GPC over product_type for this use case is standardisation: GPC IDs are consistent across your catalogue (you're using Google's taxonomy), whereas product_type values are your own creation and may be inconsistent if entered manually. Rule-assigned GPC values in Trajekt are guaranteed consistent.

Combining dimensions: the segmentation matrix

The most sophisticated Shopping campaign structures combine multiple segmentation dimensions. A product that's High-Margin (custom_label_0) + Current Season (custom_label_1) + Premium Price (custom_label_2) + a specific Brand + a specific Category represents a very precise commercial profile — and can be bid at a very different level from a Low-Margin + Clearance + Entry-Price product.

You don't need to create a campaign product group for every possible combination. The practical approach is to identify the 5–10 segments that account for the majority of your revenue and build explicit product groups for those. Let everything else fall through to a catch-all "Other" product group at a conservative bid.

The feed quality requirement behind all of this

Feed-driven segmentation only works if your feed attributes are consistently and correctly populated. A brand field that's empty for 40% of products creates a "brand unknown" product group that's unmanageable. A product_type hierarchy that's inconsistent across different product ranges means your Shopping product groups don't cleanly represent commercial categories.

The Trajekt rules engine is particularly valuable here: it applies consistent values to attributes across your full catalogue, corrects inconsistencies from your ecommerce platform's data entry, and runs on every feed refresh so that new products automatically inherit the correct segmentation values. The rules encode your segmentation logic once; the feed executes it continuously.