TL;DR
- Client: a luxury furniture brand (category: Luxury Home Goods; client not named)
- Challenge: Meta fatigue; needed a scalable upper-funnel channel
- Result: 68% CPA reduction, 4.2x incremental ROAS, +127% brand search lift, with incremental lift measured through holdout testing
The challenge
A luxury furniture brand struggling with Meta fatigue needed a scalable upper-funnel channel. Paid social had stopped finding new buyers at an acceptable cost, and the team needed a channel it could measure on incremental sales rather than platform-reported conversions.
What we did
We deployed premium CTV with strict frequency caps and measured incremental lift through holdout testing.
- Audience: affluent households modelled from first-party customer dataCONFIRM before publishing
- Inventory: premium streaming publishers bought through private marketplace dealsCONFIRM before publishing
- Holdout design: matched markets held back from CTV and compared with exposed marketsCONFIRM before publishing
- Test length: [CONFIRM: number of weeks]CONFIRM before publishing
Results
Incremental lift was measured through holdout testing, so the 4.2x counts only sales above what the held-out markets produced.
What this means if you are considering CTV
- Test CTV beside your existing paid social and search rather than replacing them.
- Agree the holdout design before launch.
- Judge the test on incremental sales and CPA, not view-through conversions.
Frequently asked questions
What did premium CTV do for this luxury furniture brand?
It cut cost per acquisition by 68% and delivered 4.2x incremental return on ad spend, with a 127% lift in brand search.
How was the incremental lift measured?
Through holdout testing: some audiences or markets did not see the CTV ads, and results were compared with those that did.
Can we run the same kind of test alongside our current agency?
Yes. A CTV test with a holdout can run beside existing search and social programs without changing them.
