TheBlendedCampaignTrap
Google's algorithms do not understand commercial nuance. If you blend Profit, Clearance, and Learning SKUs into one PMax campaign with a single ROAS target, you corrupt the bidding signals for all of them.
The Blended Campaign Problem
When you mix SKU roles in one campaign, the algorithm does what it does best: find the easiest path to hitting your target. That path is almost never aligned with your commercial objectives.
The Algorithm's View
To the algorithm, a sale is a sale. It does not know that the clearance item has 5% margin while the hero product has 40%. It sees conversions and cost per conversion. You see margin destruction.
What Happens When You Blend
When Profit, Clearance, and Learning SKUs share a campaign with a single ROAS target:
- • Clearance stock dominates: Low prices convert easily, so budget flows there
- • New products starve: No conversion history means no algorithm confidence
- • Only "safe" bestsellers get budget: Proven performers crowd out everything else
- • Margin mix deteriorates: High-volume, low-margin SKUs eat the budget
"If you mix SKU roles in one campaign, Google will spend all budget on the clearance item because it converts easily, destroying your margin mix."
The Golden Rule
Never mix SKU roles under one strategy. The machine optimises for what you tell it, not what you secretly hope for.
If your clearance items and profit drivers share a tROAS campaign, the algorithm does not know one should be clearing stock at any margin while the other should be defended at high efficiency. It just hits the target the easiest way.
Proper Segmentation
Separate campaigns or asset groups by commercial role:
Profit Drivers
Separate campaign. Target: POAS. Budget: Protected.
Customer Acquisition
Separate campaign. Target: CAC/LTV. Budget: Ringfenced.
Demand Capture
Separate campaign. Target: Impression Share. Budget: Defensive.
Inventory Clearance
Separate campaign. Target: Stock velocity. Budget: Tactical.
Learning & Testing
Separate campaign. Target: Data density. Budget: Capped.
Naming Conventions
Structure names so any team member can identify the commercial intent immediately:
Naming Format
[Role]_[Category]_[Market]_[Network]
Example: PROFIT_Shoes_UK_PMAX
Example: CLEARANCE_SS24_UK_Shopping
Example: ACQUISITION_Gateway_UK_PMAX
Bidding by Role
Each role needs its own bidding logic:
- • Profit Drivers: tROAS aligned to POAS. Portfolio budgets to smooth volatility.
- • Acquisition: tCPA or very low tROAS to force volume. Strict daily caps.
- • Demand Capture: Target Impression Share on Search. tROAS on Shopping.
- • Clearance: Max Conversions or drop tROAS floors. Seasonal adjustments.
- • Learning: Max Clicks with strict budgets. Optimise for data, not ROAS.