Research
Shorter Forms, Better UX Without Increasing Risk
When UX Research Meets Data Science
The Challenge
A B2C lending platform was stuck in a classic business tension: they needed to optimize their application journey to get more customers, but couldn't compromise their ability to screen for creditworthy borrowers. Through user interviews, I found that form completion was the biggest drop-off point - people found the lengthy application tedious and irrelevant. However, this is not a simple cutting out questions to make it easy, it also involves the risk assessment. So what we really needed was to figure out which questions actually mattered for lending decisions.
What I Did & The Outcome
I went beyond traditional usability research and dove into the platform's historical database to analyze what actually predicted loan repayment. Using predictive analytics and user segmentation, I identified which form fields really mattered for lending decisions. This proof of concept led to close collaboration with the data team who further validated and refined my approach. As a result, we collaboratively cut the application form in half while maintaining decision-making quality, increasing acquisition by 10% with no increase in default rates within the monitored period.
Approach
Going Beyond Traditional UX Research
Instead of just doing usability testing, I accessed the platform's historical database containing customer information correlated with loan repayment outcomes. This let me approach the problem from both user experience and business risk perspectives.
Predictive Analytics for Form Optimization
I analyzed the data to identify what actually drove loan default likelihood, moving beyond assumptions to evidence-based understanding of repayment behavior. Built a predictive model using half the dataset and validated it against the remaining data to find variables with the highest predictive power.
User Segmentation and Risk Profiling
While analyzing the data, I conducted segmentation analysis to identify distinct user groups and their risk profiles. This revealed different behavioral patterns across customer segments, letting me develop targeted approaches for each group rather than one-size-fits-all solutions.
Cross-Functional Hypothesis Testing
I worked with stakeholders to test whether we could significantly reduce form complexity while maintaining decision-making quality. This created a compelling business case for both improved UX and sustained risk management.
Impact & Results
Key Research Insights
40% form reduction potential
Linked emotional experience goals to measurable technical parameters
Predictive variable identification
Found specific data points with highest correlation to repayment behavior across different user segments
UX-risk balance
Proved that user experience improvements didn't require compromising screening quality
User Segmentation Breakthroughs
Distinct customer segments
Linked emotional experience goals to measurable technical parameters
Targeted optimization
Found specific data points with highest correlation to repayment behavior across different user segments
Personalized approaches
Proved that user experience improvements didn't require compromising screening quality