Role
Project Overview
Problem Statement
End goal
Senior Product Designer
Create flagship fraud detection software for Experian that allows clients to detect and action fraudulent activity quickly.
This product is shipped and in use.
How might we increase the speed in which an end user actions a case as fraud and can identify personal information and connections that help determine fraudulent activity.
Create streamlined and visually simple UI that encourages user to quickly dive into large amounts of data without feeling intimidated.
1. Research
Empathize
We utilized user research to empathize with our clients and determine how to solve their problems.
We ran virtual user testing sessions through our UX research department and helped guide user interviews.
There were dozens of tests run and we were able to conclude that users needed a clear way to action a case quickly and visual separation between different types of data.
We utilized Pendo to capture data in real-time and enhance the product continuously.
2. Design Approach
Define
We defined the problem again post-user research. We needed to reconsider how we looked at approving a case and what information was most important. We also decided was information could be hidden behand tabs at first glance and which info needed to be prioritized.
Using user personas, we dove deep into who we are defining for and what their unique needs were.
Iterate
We pursued multiple iterative concepts based on the personas and feedback we received through user research both from clients and from user testing.
We brainstormed several concepts for actioning as well as different ways to display data. We concluded Tableau gave us the most customization while also appearing the most current.
3. Solution
Data-density
After multiple iterative versions and a re-assessment of user tests, we decided to move forward with the version that displayed data in varying formats and broke the data up depending on what had been flagged as fraud and what passed the background check. This helped the data not feel so overwhelming.
Case Actioning
We also moved the actioning feature to the right hand side so the case actioner could still see all the data to the left in an unobstructed way.
4. What we learned
Solving our problems
THIS WAS A BIG ONE! Catching fraudsters requires speed. Time to action a fraud case went from 3 minutes to 30 SECONDS 🎉 This encouraged the user to visit the app regularly and keep their wellbeing levels highly on average.
Client rating went up
Average rating for the software was up from 3.8 to 4.6 for ease-of use, up from 3.5 to 4.4 for efficiency, and up 3.2 to 4.5 for satisfaction
