How a Leading Fintech App Proved True Incremental Growth with Holdout Testing

Outcomes

Objective

A leading mobile fintech app, with a suite of services spanning banking, peer-to-peer money transfers, and credit cards, ran cost-per-engagement (CPE) campaigns with Perform^[cb] built around meaningful financial events, including first money movement (P2P transfers) and card orders. As the partnership scaled, the client wanted to answer the hardest question in mobile user acquisition: were these users genuinely incremental, or would they have converted anyway?

Approach

Since the start of the partnership, Perform^[cb] has driven over 1.6 million app installs and new banking customers for the app. When the client set out to run a geo-holdout incrementality test, Perform[cb] recommended its proprietary PSA-style incrementality methodology instead, and the client chose to move forward with that approach.

For every campaign, a matched holdout of real users from the same vertical and partner mix, with zero ad exposure, was tagged as its own source inside the client’s MMP and verifiable at any time. With all campaign and control attribution running through a single source of truth, lift could be measured cleanly rather than inferred.

Results

The test demonstrated strong, verifiable levels of incrementality: 89% on iOS and 97% on Android, in an industry where many channels struggle to clear 50%. On iOS, campaign conversion rate reached 0.36% against a PSA control of just 0.04%; on Android, 0.26% against a 0.01% control. The results proved that the overwhelming majority of conversions the program drove were genuinely net-new, giving the client evidence-based confidence to allocate budget toward proven lift.