Umba
Understanding User Motivations To Reduce Churn
Context
Umba is a digital bank focusing on the user segment of underserved & new banking customers in Africa. We were struggling with approving users with limited or no banking history into our loan program.
Impact
40% Increase in the # of user sticking with the app after loan decline
Team Structure
Product Manager
2 Engineers
1 Data Analyst
1 QA Analyst
Product Designer (Me)
Timeline
Week 1: Design
Week 2: Develop
Week 3: QA & Launch
I got declined...
Now what?
Umba was on a mission to change industry norms in African banking. Traditional banks created a predatory environment that broke any trust with consumers. Our challenge was to build back trust and create a credit product that empowers financial well being.
Creating alignment between the business, users, & product
Goals
We want to curb risky loans, tracked by % of loans defaulted
We want a way to become credit eligible
We want to build back trust with users
Metrics
% of loans defaulted
% of users retained after loan decline
% of customer reviews about loan decline
So if I don't get a loan, why should I stay?
The loan decline process was very stark, not offering users opportunity to do anything else. This led to over 90% of users churning after being declined for a loan. Most users are applying to dozens of loan apps so they dont care to stick around with any that dont provide money.
Without knowledge of or incentive to use other features why would anyone stick around?
Testing user motivations
By offering a way to grow eligibility we were able to test if users would use other features after being declined for a loan. This gave us insight into how many users we could leverage, and what features they were interested in. Using a fake door test gave us answers faster and in a real world scenario, instead of a lengthy process gathering research users or survey answers.
~20% of users clicked through to interact with other features.
Looking at the data
Diving into data analytics on Mixpanel, I learned that 70% of users were clicking into the rewards tab. This is likely due to the fact customers want money (they applied for a loan), and Rewards is a second attempt at achieving this goal. But users still churn given they cannot get any money from Rewards.
Rewards was not providing value to the user, it was only analytics.
Leveraging organic movement with incentives
Using the natural movement of users to the rewards tab, I implemented an incentive based FTUX. People are coming to Umba to get money, so lets give them some. I worked with the finance team to budget with respect to the Cost of Acquisition so it was a win win. Umba gets to retain customers, and Users get money to grow their finances.
Users were rewarded for completing a loan application regardless of their eligibility, and encouraged engagement with other features through monetary incentives.
Halved losses for the month & Set Up long term growth
We rolled out the feature gradually, first to 10% of users, then 30%. We retained 40% of users that were declined for a loan and provided real value to users building product trust. What excited me most, is creating an opportunity for users that would otherwise never be offered a loan to build their banking history, growing the financial well being of those individuals and for their future generations.
We want to curb risky loans, tracked by % of loans defaulted
Goals
Metrics
<5% loan defaults
We want a way to become credit eligible
Engaged with 40% more users after loan decline
We want to build back trust with users
Reached a 4.1 star rating in App Store
Bonus Content Below
Design Nitty Gritty
Creating this required a strong balance between utilizing as much of our existing design system and providing clear interaction and their impact for the user. There were a lot of small decisions made that allowed for fast development and consistent UI. Here are a few of them.
The screen layout focused on making sure users understand which actions are recurring so they don’t expect rewards for everything.
Providing more information through our button component offered a way to correlate information behind a click to the action you are about to take. This is something we then adopted into the design system.
To curb fraud, we implement a selfie check before any actions that can be exploited. By offering a preview of the rewards screen, users can see there are monetary incentives to completing the selfie check.
