Product designer with a statistics background
1.2 Iphone X-credit.png

Optimizing credit card recommendations for debtors and travelers--dennis xing

Credit Sherpa

1. Hero v5.jpg



Consumers are not applying for the credit cards that fit their spending habits

Credit Sherpa is a fintech business that focuses on improving the financial health of its users. It differentiates itself from its competitors because it puts its customers wellbeing first when making credit card recommendations, rather than accepting payments.

My role as product designer

I led the design of the financial recommendation experience, planned and executed the user research and worked on the detailed visual design phase. I was the product design lead consultant for Credit Sherpa.

14. The Problem.jpg
4.1 The Goal.jpg

The goal

How do we offer financial advice to users while still making money off credit card recommendations?

1. sherpa needs.jpg

Final Screens

Personalizing Credit Card Recommendations

7.1 Final iphone screens.jpg

Research and Discovery

The Average American carries $4,789 in credit card balances

Looking at the Center of Microeconomic Data, credit card balances hit $830 billion at the end of 2017. 70% of Americans have at least one card. That’s 173 million cardholders over the age of 18. This means the average cardholder has approximately $4,789 in credit card balances.

2. credit card debt by age.png

Insights from Usability Testing

Users are crushed by interest fees

I tested our app with 8 users carrying credit card debt in San Francisco. My goals were to understand how people in their current financial state make credit card decisions and the workarounds and tactics they use.

  1. Users with credit card debt know exactly how much debt they have, but don’t know how much they are paying in interest

  2. Users with credit card debt who are already on a 0% intro APR credit card know exactly when their intro period ends

  3. Users with revolving balances don’t know how much their current APR is.

  4. Users don’t know how to utilize credit card points



Users in debt and users not in debt became our two target markets

1.2 debt persona.jpg
3. traveler persona.jpg


Users in debt and users not in debt have different recommended actions

From my interviews, the actions that users take differ drastically if they have debt or not. If users are in debt, the main thing that credit sherpa can do to help them is to let them know how much estimated debt they are in. If users are not in debt, the main thing is to suggest an alternative credit card that allows them to get more value out of their spending history.

2. Mind map.jpg
1.1 questionnaire.jpg


Crafting a Better Questionnaire to Bucket Users into Debt vs No Debt

The original questionnaire had high drop rates since there was too high of a cognitive load and not enough trust on behalf of credit sherpa. In addition, the previous site had no branding, so I came up with colors, typography, and a logo to make sure that the user had a consistent feel across the site.

I structured the questionnaire to be as simple as possible, so that users have no difficultly in inputting the correct values as well as getting to the results page as soon as possible. Usability testing this app with users, I found numerous issues that I then redesigned and iterated upon. Using the principle of Ockham’s razor, I minimized the amount of clicks for users .

2. iterations on high fi.jpg


The final questionnaire

I made sure to optimize both for happy paths as well as sad paths. The most important thing I had in mind when designing was to make sure that users were comfortable putting in personal financial data. From the avatar that I put in to the layout of the questions, I made sure to maximize for usability and trust.

1. Questionnaire Flow v2.jpg

Interactive Prototype

Seamless transitions to minimize questionnaire dropoff

v2 sherpa Kapture 2019-06-18 at 13.58.45.gif

I prototyped an interaction in Principle to showcase how seamless it would be for users filling out the questionnaire.

2.2 results.jpg


Finding the balance between information density and cognitive overload

Initially, I thought that showcasing estimated value per year for no debt users or interest per year for debt users would be vital information; however, upon user testing and feedback from key stakeholders, I realized that the key takeaway should be one single actionable insight.

10.2 iterations on cards.jpg


Personalized tips for current cards based on level of debt

From my user research, the most beneficial action for people in credit card debt to take would be to rid themselves of it. By highlighting front and center the estimated amount of credit card debt that they are paying, I nudge them into taking action.

Our algorithm buckets users into debt or no debt based on how they answered the questionnaire. As a result, the calculations for each type of card is different. The cards shown for the debt persona highlights the estimated interest paid and the no debt persona highlights the estimated value from their spending.

11.2 final cards.jpg
9.1 debtor final screens.jpg

Minimizing Performance Load
The current cards expanded view has the calculations of the estimated amount of interest per month to enhance user trust. One of the main doubts that users have with sherpa’s estimated numbers is “how are they calculated?” By showcasing the calculations, we can build trust with the user.

12. no debt final screens.jpg

Progressive Disclosure
For users who are interested in the calculations of how sherpa valued each credit card, they can click to see more. I designed for the majority and the few that needed an extra nudge had the option to.

Exercise in Simplifying
The high level financial summary has three key takeaways from the insights derived from the questionnaire. Consumer financial literacy is complex and jargony. By making it accessible to users from all backgrounds, sherpa can build trust.



Responsive Web

From my analytics, 50% of the users access the site on desktop. I focused our designs on mobile, but adapted it for web.

3. results page mobile and web.jpg
4.1 questionnaire page mobile and web.jpg



Currently we are A/B testing the new redesign of the site. From our initial validation testing I observed that users have more trust with credit sherpa and are willing to input more data. I optimized the designs for the two personas, users with debt and users with no debt.

Measures of Success

  • Questionnaire completion rate increase

  • Clickthrough rate on recommended cards increase


Project Learnings

User research is crucial

Only by talking to users in credit card debt and understanding how they pay their bills every month, I could design for them.

Financial regulations make designing for fintech hard

The financial industry has a bunch of constraints that I need to be mindful of. From the consumer protections enacted by the Federal Trade Commission to the area of credit cards, I made sure to put the consumer first and foremost.