UX Case Study: Optimizing Loan Applications
Context:
The fintech platform Celsius offered retail users the ability to take loans backed by their own cryptocurrency as collateral. This was a high-priority product for the business but our analytics showed that, of the users who started the loan flow, only 12% completed it.
Understanding the problem:
The main drivers for offering a loan product were to maintain feature parity with our competitors and to create a revenue stream, but there wasn’t much user research to vet the idea.
We conducted user interviews and surveys to better understand what users wanted out of this feature and how they might use it. Our results were overwhelming: most users we spoke to did not, fundamentally, understand the product.
The majority of users were only familiar with loans that were secured by their credit history, like education loans or traditional mortgages. Even users who were familiar with the idea of repossession didn’t understand how collateralized loans worked.
Users didn’t have a clear understanding that our product was a collateral loan backed by the value of their crypto. Further, those that did understand the basics of collateral loans were confused about the mechanics of important concepts like margin calls and liquidations.
Additionally, we saw that users were consistently interacting with our loan calculator far more than the actual loan application.
My hypothesis:
I believed that users were curious about taking loans but didn’t like the pressure associated with starting the loan application. There was no easy way for users to explore their loan options without starting a formal application, and users were being scared off.
In past efforts to make the product accessible, we had removed the common financial terms and jargon associated with loans.
Terms like margin call and liquidation were de-emphasized, but these were precise financial terms with no obvious replacements. Hiding them caused confusion about how the product actually worked.
It was also a potential landmine for users who might be liquidated (read: lose their entire collateral) if they didn’t understand these concepts.
The result was a product that was actually less accessible despite our efforts.
The fix:
Embrace the calculator concept early.
If users felt pressured by the application, we could give them a space that was explicitly for experimenting. A calculator could combine multiple steps from the old flow into a single, more intuitive screen.
Our squad’s designer created a calculator concept that accommodated all use-cases including interest and collateral options. The loan calculator was integrated into the loan application flow, allowing users to start the flow without feeling the pressure of a formal application.
Our loan calculator concept.
Rather than eliminate jargon, we had to embrace and explain it.
We reintroduced the necessary financial terms like margin call and liquidation, and provided tooltips to educate users who needed help. We explained these concepts in plain, actionable language.
The old loan summary design which did not include key concepts like margin call.
Our new design explaining collateral with a tooltip.
Our new design explaining margin call with clear, actionable language before the user requests their loan, and including a tooltip with additional information.
The result:
Conversion for the loan application flow increased from 12% to 21%.
My role:
Collaborated with user researchers to identify the friction points and KPIs
Collaborated with product designers on all design iterations
Developed copy for all points of interaction such a titles, buttons, copy, and tooltips
My lessons from the project:
I knew that fintech, and particularly the cryptocurrency space, had inherent UX challenges due to the combination of complex financial concepts and emerging technology. This project emphasized that we needed to be wary of shortcuts: you can only simplify a complex product so much before you pave over the parts that users need for product mastery.
I suspect that our product’s success was due in part to experienced users overcoming our UI. Less sophisticated users were getting stuck where these experienced users had succeeded. As our customer base expanded from the crypto-literate to the crypto-neophytes, we needed to do a better job of researching and understanding them. We had to be better at meeting customers where they are, not where they used to be.