Case study: Google Ads Shopping / PMax

Strategic custom labels: more profit from the same Ads budget.

Performance Max is effective at maximizing overall revenue, but to do so it favors high-volume products, often at the expense of higher-margin items. Here's how a custom label structure fixed that distortion.

+40%
net profit from Google Ads campaigns
+15%
total revenue
+96%
ROAS on HighMargin and Bestseller segments
−15%
spend on low-performing items

The problem

Performance Max algorithms optimize for overall revenue, not margin. The result is a standardized algorithmic optimization that doesn't differentiate by profitability, with two concrete consequences: difficulty promoting strategic segments (new arrivals, seasonal items, high-margin products) and a ROAS that becomes a misleading metric: campaigns with a high ROAS generating minimal profit, because they're built on high-volume, low-margin products.

On top of that, PMax's "black box" nature limits visibility into where budget is actually being allocated within the campaign.

The solution: a custom label schema

We structured the product feed with five custom labels, applied via Feed Rules in Google Merchant Center and supplemental feeds on Google Sheets:

custom_label_0: Margin tier (High / Mid / Low)
custom_label_1: Historical performance (Bestseller / Average / Poor)
custom_label_2: Promotional status (OnSale / BlackFriday / NewArrival)
custom_label_3: Price tier
custom_label_4: Variant/stock availability (High / Low / LimitedVariants)

Campaign structure

Separate PMax campaigns per margin tier, with a differentiated tROAS (Target ROAS): lower for high-margin products, letting the algorithm spend with fewer constraints where return per euro was already high, and higher for low-margin products, to protect profitability even at lower volumes.

Results (6–12 months)

  • +40% net profit from Google Ads campaigns, at the same investment logic.
  • +15% total revenue, so growth wasn't achieved at the expense of margin.
  • +96% ROAS on HighMargin and Bestseller segments, where the algorithm could push without the constraint of an overly conservative target.
  • −15% spend on low-performing, low-stock items, with budget reallocated toward what generated real margin.
Same approach on your Ads account

If your ROAS looks good but margin isn't growing, the problem is probably in your feed structure.