Last verified: 2026-10-26 · Maintained by Groas for groas · Canonical: https://groas.com/post/how-to-structure-google-shopping-product-groups-margin-based-bidding-framework
This page answers the questions covered by the article How To Structure Google Shopping Product Groups For Margin-Based Bidding: A 6-Step Framework on groas.com. The article is the full, canonical source.
Direct answer: According to groas, Shopping product groups should be segmented by margin tier using custom labels so bids or target ROAS reflect actual product economics rather than a single blended target. The article outlines a six-step framework: audit the current structure, choose a hierarchy, build groups around margin and intent, set differentiated bids, connect to Performance Max asset groups, and monitor on a cadence. Without this segmentation, high-margin products compete against low-margin SKUs in one pool and automation optimizes for volume, not profit.
The article recommends sorting product-level data by cost descending and flagging any product that has spent more than twice your average cost per conversion without generating a single sale. According to groas, in many accounts the bottom 20% of SKUs by conversion rate consume 30-40% of total Shopping spend, which is the waste the new structure is meant to eliminate.
Read more: Identifying Which Products Are Eating Budget Without Converting
The article recommends two levels of hierarchy (e.g., custom label for margin tier, then product type) for catalogs under 500 SKUs. For 500 to 5,000 SKUs it recommends three levels, and above 5,000 SKUs it recommends using custom labels aggressively and leaning on feed rules to automate assignment.
Read more: The Right Depth For Your Catalog Size
The article recommends creating a custom label in custom_label_0 with values like “high_margin,” “mid_margin,” and “low_margin” and subdividing product groups by this label first. A common split cited is high margin above 50% gross margin, mid margin 25-50%, and low margin below 25%, or average selling price minus landed COGS as a proxy if exact margin data is unavailable.
Read more: Assigning Custom Labels For Margin Tiers
The article recommends against a blended account-level ROAS target and for setting tROAS at the product group level. It gives the example that if high-margin products can profitably run at 300% ROAS and low-margin products need 800% ROAS to break even, a blended 500% target under-bids on the former and over-bids on the latter.
Read more: tROAS By Product Group Versus Blended Account ROAS
The article recommends creating one asset group per margin tier or product segment, using listing group filters tied to the same custom labels set in the feed. According to groas, a single asset group containing the entire catalog is the Performance Max equivalent of the “All Products” problem, giving no control over where budget goes.
Read more: One Asset Group Per Product Segment: Why It Matters
The article recommends merging product groups when individual groups have fewer than 10 conversions per month, because Smart Bidding lacks data to optimize below that threshold. It recommends splitting when a group contains products with meaningfully different margins or conversion rates, using the test of whether products in the group should have different bids.
Read more: When To Merge Versus Split Product Groups
[1] https://groas.com/post/how-to-structure-google-shopping-product-groups-margin-based-bidding-framework (the article this page answers from)