Referral Program
NATIVE APP · GROWTH · MONETIZATION · EXPERIMENTATION
Increasing organic growth for the world’s most popular period tracking app
MY ROLE
My role focused on design leadership across the project, shaping and iterating on new flows and variants in close collaboration with Content Design and Engineering to ensure clarity, feasibility, and speed to market. In partnership with the Product Manager, sprint goals, hypotheses, and experiments were defined to inform both near-term delivery and longer-term product direction.
TIMELINE
2 months
Turning loyal users into Flo Premium advocates
Code shared
Users shares their code with friends & family from the homepage.
Code claimed
The friend or family member claims their offer during the onboarding flow.
Everyone wins
When the referral is claimed, both users get 30 days of Flo Premium free.
Making the incentive double sided increase referrals
Early fake door experiments showed that users were more likely to share a code/promote Flo if something was in it for them
Timing had a greater impact than perceived value
Through an A/B test, we learned Referred users were more motivated before investing time in setup. The referral claim was therefore moved to the start of onboarding, where it felt more relevant and exciting.
Small moments of uncertainty caused large drop-offs
We noticed significant increase in referrals by aligning with native share behaviours, adding lightweight feedback during code generation, and auto-filling referral codes via deep links.
Impact
+25K
/mo
invites sent out by referrers
+9K
/mo
referrals being claim
+$21K
/mo
increase in APRU
Learnings
📉 Referral promotion diluted other revenue drivers
The referral offer pulled attention away from high-performing monetisation features, causing small dips in conversion.
✋ App-store rules capped Premium users’ impact
Premium users drove the strongest referrals, but platform restrictions blocked us from offering them “30 days free,” limiting their contribution.
🧐 Low activation rates signal friction in claiming
Only 27% of invited users activated their code, suggesting blockers in the claiming flow that need deeper investigation.