A/B testing microcopy: how tiny copy changes raised conversions for PetalPay

Design · 3 min read

PetalPay, a fintech wallet startup, saw a large drop-off on the identity verification step. Heatmaps and session replays suggested users were uncertain about required documents and feared data misuse. The design team prioritized microcopy as the fastest lever to reduce friction without backend changes.

They ran sequential A/B tests on four copy elements: the verification headline, one-line privacy reassurance, inline examples of acceptable documents, and error messages with actionable next steps. Tests were instrumented in Optimizely with goals around completion rate and time-to-complete, while retention and fraud flags were monitored to ensure signal quality wasn't sacrificed.

Small wins added up: a clearer headline and explicit privacy reassurance improved completion by 9%, and actionable error messages cut repeat attempts by 15%. PetalPay's case underlines that startups should treat copy as a product lever—fast to iterate, measurable, and often cheaper than engineering-heavy solutions.