Incrementality testing by platform: Overview · Meta · Google · TikTok · Amazon · Pinterest · Snapchat
Pinterest runs Conversion Lift Studies the same way other platforms do: a holdout group doesn't see your ads, and the gap in outcomes is the lift. Pinterest publishes little public detail on its studies, so this post draws on the Pinterest Academy course outline, published case studies and PPC Land.
The idea in one picture
This is the general pattern. Check the details with Pinterest.
What we know about how it runs
- Pinterest Academy has a course on when to use a study, how the study works and how lift is calculated, how to check feasibility, and how to plan for acting on results (Pinterest Academy).
- Pinterest lists conversion lift among its incrementality products (PPC Land).
- Minimum spend, minimum conversions and study length aren't published. Ask your Pinterest rep for the feasibility check before you plan a budget.
What published studies look like
| Advertiser | Design | Reported result |
|---|---|---|
| TUI (Belgium, Netherlands) | Study with a 20% holdout in both markets, measuring checkouts, sales, add-to-carts and site visits | Netherlands: 6.8% sales lift and 5.1x incremental ROAS. Belgium: 13.9x incremental ROI |
| Søstrene Grene | Seven-week study combining always-on and flighted campaigns, run across 14 European markets per PPC Land | 22% lift in checkout conversions, 21% sales lift, 1.3x incremental ROAS |
| Semaine Health | Geo-based pause-to-measure test run with WorkMagic | 12.4% incremental lift overall; 87% of the lift showed up off-site, on Amazon and Walmart |
These are success stories Pinterest chose to publish. Read them as examples of study design, not benchmarks.
Geo option
A pause-to-measure geo test keeps most ads running and pauses only a small, representative group of markets. Comparing exposed and held-out markets isolates the incremental effect without a large dip in overall performance. Semaine Health used it to see sales that happened off its own site, which user-level tracking would miss (WorkMagic).
What's new in 2026
- Self-serve A/B testing. At Pinterest Presents on September 17, 2026, Pinterest added self-serve A/B tests for creative, targeting and Performance+ setups (Mintec). These compare two versions of a campaign against each other. They don't hold anyone out, so they don't measure incrementality.
- A vendor read across tests. At the same event, Measured said its review of 70 recent Pinterest incrementality tests found Pinterest performing 2–4x better than attribution models credit it (Florescomedia). That's a measurement vendor speaking on Pinterest's stage, so treat it as a reason to test, not a result to assume.
Watch-outs
- Pinterest designs and scores its own study, and it sees Pinterest only.
- Published results are advertiser-selected case studies with no sample sizes.
- Requirements and metric definitions need confirming with Pinterest.
I design and read incrementality tests for brands, on the platforms and independently of them. See how a Measurement Audit works.
Sources
- Pinterest Academy: Conversion lift course
- Pinterest Business: TUI
- Pinterest Business: Søstrene Grene
- Pinterest Business: Semaine Health
- WorkMagic: Semaine Health and Pinterest
- Mintec: Pinterest Performance+ control and A/B testing
- Florescomedia: Pinterest Presents 2026 summary
- PPC Land: Explaining conversion lift