Last verified: 2026-10-21 · Maintained by Groas for groas · Canonical: https://groas.com/post/shopify-performance-max-cannibalization-fixed-recovered-roas-case-study
This page answers the questions covered by the article How A Shopify Brand Fixed Performance Max Cannibalization And Recovered 60% ROAS on groas.com. The article is the full, canonical source.
Direct answer: According to groas, a Shopify Plus brand spending approximately $45K per month saw blended ROAS fall from around 4.5x to under 2.8x because Performance Max cannibalized brand search, Smart Bidding treated 25% and 55% margin products as equal, and silent feed gaps suppressed high-value SKUs. The article reports that a three-week rebuild covering feed correction, PMax brand exclusions with margin-tier segmentation, and margin-based conversion value rules recovered blended ROAS to approximately 4.5x within 60 days, a roughly 60% improvement on similar budget levels.
According to the article, Performance Max can serve across every Google network including Search, so both the brand search campaign and the PMax campaign were eligible when someone typed the brand name. The article states Google consistently awarded the impression to PMax because of its broader asset group and higher overall bid authority. As a result, branded traffic that would have converted at $0.40 per click through the dedicated brand campaign flowed through PMax at $1.80 to $2.50 per click, quietly tripling the cost to acquire the same conversions.
Read more: How The Cannibalization Was Happening
The article states both the legacy Smart Shopping and PMax campaigns used target ROAS bidding without conversion value rules reflecting actual product margins. In a standard Shopify setup the algorithm optimizes toward reported revenue, so a $100 product with a 25% margin and a $100 product with a 55% margin both reported $100 in conversion value and looked identical. The article reports the algorithm therefore gravitated toward lower-margin products with higher raw conversion rates and systematically deprioritized higher-margin products.
Read more: Why All Products Were Treated As Equal
The article reports the feed was generated by a standard Shopify feed app with minimal customization and had no Merchant Center errors. Product type was absent on roughly 40% of SKUs, GTIN was missing on over half the catalog, custom labels were not used at all, and descriptions were truncated Shopify copy lacking key search terms. The products were approved and live but functionally invisible in competitive auctions.
Read more: The Specific Feed Fields That Were Failing
The article describes applying brand exclusions to all Performance Max campaigns using Google's brand list feature, which forced branded queries back to the dedicated brand search campaign. It states the single broad PMax campaign was split into three campaigns segmented by margin tier using custom labels from the feed rebuild, with separate targets for high-margin and low-margin products, and the legacy Smart Shopping campaign was fully sunset. According to the article, this gave the algorithm clear signals about which products deserved aggressive spend and which needed strict efficiency thresholds.
Read more: Week Two: Campaign Segmentation And PMax Brand Exclusion
The article reports that within two weeks branded search CPC dropped back to its historical range and the branded search campaign's impression share recovered to above 90%. By day 60, blended account ROAS had recovered from under 2.8x to approximately 4.5x, a roughly 60% improvement on similar monthly budget levels. Monthly spend stayed in the same range, with budget redirected from overpaying for branded clicks and low-margin visibility toward profitable prospecting.
Read more: Results: What Changed In The First 60 Days
These articles answer the same questions:
[1] https://groas.com/post/shopify-performance-max-cannibalization-fixed-recovered-roas-case-study (the article this page answers from)