#Launched
#UserGrowth
#E-Commerce
I redesigned Shopee's sign-up authentication system to unlock the "Silver Economy" (Age 40-65), driving a 24% lift in high-value user acquisition and 70% improvement in first-purchase conversion through a trust-first strategy.
1 PM, 2 Engineers,
1 Researcher, 1 Data Analyst
2020.08-2020.09
3-week sprint
Lead Product Designer
IMPACT AT A GLANCE
This project redefined how I think about growth — trust beats speed every time for high-value demographics
+24%
lift in high-value users
70%
first-purchase Conversion
10M+
Users Impacted
CONTRIBUTION
E-commerce moves fast. Working across 6 teams in 3 weeks taught me to align stakeholders quickly, solve problems creatively, and ship with confidence
Technical Problem-Solving
FE background bridged design-eng gaps
Data-Driven
Partnered with Data & UX teams to analyze 50K+ signups
Cross-National Team Alignment
Taiwan + Singapore research & design systems
3
New Feature Designs
21
Prototypes
6
UX Improvements
CHALLENGE
Q2 2020. Shopee Taiwan's growth engine hit a wall
For years, our playbook was simple: bigger discounts = more users. By Q2 2020, that stopped working. Pouring more money into promotions wasn't moving the needle anymore.
We needed net new users, not bigger discount
PROBLEM
The Sign-Up Bottleneck. Here's where things got tricky
Unlike Amazon or eBay, Shopee operates on a vertical browsing model. Users must sign up before they can view product details, check prices, or add anything to cart
Wait... Were we completely wrong?
We added OAuth. Signups dropped.
Pause. Breathe. What’s next?
Time to dig deeper and find the real problem
REFRAMING THE PROBLEM
Why were high-value users leaving? Digging Deeper with Data Analyst
I partnered with our data analyst to segment the funnel by age and analyze behavioral patterns. For a B2C product at Shopee's scale, we relied on quantitative data analysis — session recordings, drop-off rates, and user behavior metrics — to uncover the real barriers
AUTHENTICATION EVOLUTION
We kept iterating, but each version left our highest-value users behind
Version 1.0 (2018)
Email + Password only
Version 2.0 (March 2020)
Email + Password
➕ Third-party OAuth (Google)
➕ Social media login (Facebook)
Version 3.0(???)
The answer?
👇👇 Scroll down to find out👇👇
SOLUTION
From "Speed-First" to "Trust-First"
FEATURE 01
Phone-First Authentication
Aligning with what users already trust
92% completion rate (vs. 58% with OAuth-first)
We deprecated the Email-centric flow and enforced Mobile Number + OTP. For this demographic, a phone number is their digital identity — "memorable and secure"
FEATURE 02
Progressive Completion Feedback
Reducing anxiety with immediate visual confirmation
Average time on form reduced by 35% (78s → 51s)
Even with optional fields, users still showed hesitation. We added subtle UI feedback to reassure users they were on the right track.
FEATURE 03
Privacy-Preserving Forms
Making everything optional gave us more accurate data
41% dropout rate, 34% fake data
Button remained Disabled until all fields were filled. Created anxiety and felt like a pass/fail exam
18% dropout rate, 8% fake data
We redesigned the profile form to align with the user's mental model of privacy, switching from Exact DOB to Age Ranges (e.g., 40-50)
NEW FEATURE
Interest Selection Onboarding
Turning "out of scope" into a 3-day win
The Roadblock
Engineering said the Interest Selection feature needed database schema changes. "3 weeks of work. Out of scope."
My Proposal
I sat with the backend engineer and looked at the actual database structure. Found an unused JSON field we could repurpose
The Result
Those who selected interests had 2.1x higher Day-1 engagement
Complete Flow
Phone → OTP → Profile → Interests
Reflection
This project redefined my approach to growth, proving that for high-value demographics, Trust > Speed is the true driver of conversion. By treating Code as Design Material, I was able to map database schemas to negotiate scope with engineers, transforming a "No-Go" feature into a launchable reality. Ultimately, this experience shifted my focus from simply building screens to designing for the Algo, ensuring that every user interaction serves as the foundational data layer that powers long-term personalization.
