Why 75% of Your SKUs Get Ignored by Performance Max

When running Google Performance Max (PMax) across a large catalog, ecommerce brands often face a familiar pattern: account-level ROAS looks healthy, but revenue is driven by a tiny fraction of products.

For example, a Shopify seller running roughly 400 SKUs (translating to nearly 1,200 variant items in Google Merchant Center) managed a stable ROAS of 3.0 to 4.0 on a $200 daily budget. However, a six-month account audit revealed that nearly 75% of the SKUs received almost zero spend and generated no conversions.

The immediate instinct for many media buyers is to delete these inactive SKUs to concentrate budget on proven bestsellers. However, zero traffic is not the same as zero demand. PMax naturally allocates budget toward products with established conversion histories. When a SKU fails to get initial impressions, it enters a cold-start trap: without traffic, it cannot generate conversion data, and without data, the algorithm allocates even less traffic. Many products are not unsellable—they simply remain untested.

The ABC Framework: Segment Before You Purge

Before deleting SKUs from your product feed, audit your inventory using an ABC segmentation strategy:

  • Category A (Hero Products): High-performing SKUs with adequate spend, consistent conversions, and target ROAS. Maintain current budget allocations to protect core revenue.
  • Category B (Budget Wasters): SKUs that accumulate spend without converting, pushing cost per acquisition (CPA) well beyond targets. These products should be deprioritized or excluded.
  • Category C (Untested Long-Tail): SKUs with negligible spend (for example, less than $2 over six months). A lack of conversions here reflects insufficient algorithm testing, not weak product demand.

How to Restructure Campaigns for Long-Tail Testing

Instead of blanket deletion, use Google Merchant Center custom labels (you can configure up to five) to tag products by performance, margin, or seasonality, and manage them across distinct campaign structures.

  1. Main Campaign: Isolate Category A winners in your primary PMax campaign to safeguard overall ROAS.
  2. Test Campaign: Group Category C (untested) items into a separate testing campaign with dedicated budget control. This forces the algorithm to generate baseline impressions without competing against hero products.
  3. Exclusions: Exclude or lower priorities for Category B waste products to prevent unnecessary ad spend.
blockquote>

Important Feed Management Tip: Do not change product Item IDs when reorganizing your catalog. Google Ads relies on consistent Item IDs to preserve historical performance data and reporting context. Changing IDs resets learning history. Use custom labels or product groups within your existing feed structure instead.

Focus on Testing Opportunity, Not Just Conversion Status

PMax excels at scaling winning products, but its machine-learning model tends to starve long-tail items of budget. Rather than asking how many SKUs have zero conversions, evaluate whether those SKUs were actually given enough impressions to perform. Strategic feed segmentation ensures high-potential products get a fair evaluation while protecting your primary revenue drivers.