Paywall Optimizing led to 36% ARPU gain
Designed a gamified feature that increased new user engagement in the MS Support Community.
Role
Microsoft
Team
Product Manager
Front-end Engineer
Back-end Engineer
Timeline
2025
1. Project Overview
iSharing is a subscription-based location sharing app for friends and family's safety. The paywall is where all revenue is won — or lost.
After analyzing Mixpanel's existing data, we found that less than 1% of users subscribed during the onboarding paywall. Every A/B test required a full developer cycle. Testing was a luxury we couldn't afford.
2. Problem Statement
How do we build a system that can learn, test, and improve without relying on a developer for every change?
3. Research
Competitive Research
compare to competitors, our design is lack of nudge to build social trust
Too much congtive load in one page
Analyze current data (Mixpanel)
Currently, 99% of the users do not subscribe during the onboarding
within those 99%, 16-23% of users subscribe within 24 hours
75% users even click the CTA button, and abandon the transaction
4. Ideation & Strategy
<Not a random tests, a deliberate system>
I proposed Superwall, a no-code paywall experimentation platform- and designed a 5 phase roadmaps.
5. Design & Iteration_1.
Multi page paywalls
Design your own trial
Instead of prescribing one paywall, we let users choose between 2 trial options.
1. Free 7 day OR
$5.99 30-day.
The result
Users self-selected their best fit, converted more, and cancelled less
Clear Headline
6. Result
I ran 11+ experiements across 5 phases- from AA baseline testing to the flagship "Design your own trial".
The numbers prove the system worked. And the system kept learning
7. Takeaways
Reducing cognitive load and giving users a stronger sense of control proved most effective.
Short, multi-step paywalls performed better than long, scroll-heavy ones.
Allowing users to “design their own trial” — comparing plans or adjusting price options — increased engagement and trust, showing that autonomy drives higher conversion.
Go Back