Key takeaways
- Why platforms overstate by design
- The triangulation stack
- Building the dashboard that ends arguments
Why platforms overstate by design
Every ad platform attributes generously toward itself: last-click models starve upper funnel, view-through windows claim credit for existing demand, and iOS signal loss pushed platforms toward modeled conversions that flatter their own optimization. Summing platform-reported revenue across channels routinely exceeds actual revenue by 30-60%.
None of this means platform data is useless. It means platform data answers inside-platform questions - creative iteration, bid efficiency - while blended numbers must answer portfolio questions.
The triangulation stack
Three measurement layers, each with jurisdiction: platform-reported ROAS for directional optimization inside channels; blended MER (total revenue over total spend) as the allocation truth; periodic incrementality tests calibrating how much discount to apply to platform claims.
- Platform data: creative and bidding decisions, weekly
- Blended MER: scaling and cutting decisions, weekly
- Geo-lift tests: calibration events, quarterly
Building the dashboard that ends arguments
One page, updated weekly: blended MER trend, channel spend shifts, branded search volume, cohort LTV movement. When every stakeholder reads the same truth serum, attribution debates become portfolio decisions - and portfolio decisions compound.
