How do DTC brands run incrementality tests on paid social?
Quick answer
Run a geo holdout test: pause or reduce Meta spend 20 to 50% in matched regions for 2 to 4 weeks while holding other regions constant. Compare blended new-customer revenue in holdout vs control. True incremental ROAS is often 30 to 50% lower than platform-reported ROAS after iOS changes. Use results to set scale ceilings, not just creative tweaks.
Why platform ROAS lies
After ATT and modeled conversions, platforms attribute view-through and organic demand to ads. Sustainable DTC with strong brand search and community validation (Reddit, influencers) sees wider gaps. Fix tracking first, then test incrementality.
How do you design a geo holdout?
- Select matched regions (similar historical CAC, population, seasonality)
- Reduce or pause Meta spend in holdout geos for 2 to 4 weeks
- Measure Shopify new-customer orders and revenue by region (not platform columns)
- Calculate incremental revenue = control growth − holdout growth, adjusted for seasonality
What about Meta conversion lift studies?
Meta lift tests require sufficient spend and population. Useful at scale ($50k+/mo). Geo holdouts work earlier and give channel-level answers you own.
How do you act on results?
If incremental ROAS is below your benchmark floor, cut prospecting before retargeting, refresh offer, or shift budget per the Meta vs Google split guide. Retest quarterly as creative and audience mature.
Frequently asked questions
What is the difference between ROAS and incrementality?
ROAS is attributed revenue divided by spend inside the ad platform. Incrementality is the revenue you would not have received without the ads, measured via holdouts or lift studies.
How long should a geo holdout run?
2 weeks minimum for high-volume brands; 3 to 4 weeks for most sustainable DTC with under 500 weekly new customers from paid.
Can you run incrementality tests on Google?
Yes. Use geo experiments in Google Ads or pause brand Search in holdout regions to measure organic capture. Pair with Meta holdouts for channel-level decisions.
What if holdout regions still grow?
Some growth is organic, influencer, or email. Document all marketing activity in holdout periods. Smaller holdouts (20% spend reduction) reduce revenue risk.