Kaoun is a FinTech/RegTech company that builds reliable infrastructure for payments and credits in Tunisia.
Kaoun’s mission is to enable all individuals and businesses to access financial and government services using any phone, anywhere, anytime. Kaoun created an array of digital products that connect users to financial and government services, through their mobile devices.
Current market trends in Tunisia 2019
My role
Creating a bank for the 64% unbanked population
My role is to conduct user research that yields actionable design decisions that I can in-turn transform into interfaces and interactions that are based on scientific information and logical rationale.
Deliverables
Conclusive user research
User interfaces for mobile
Controlled experiment data
Visuals/Animations for onboarding KYC tasks
Design Toolkit
Process
Design Process
Given the low knowledge of our user persona, we zeroed in on our users through a hybrid research method (qualitative & quantitative):
Problem Statement
How can you build a financial inclusion instrument to an unbanked, unfamiliar population in the onboarding experience?
Flouci is the first wallet that is designed to innovate mobile payment in Tunisia. It serves as a quick, easy, and convenient way to open a bank account, send and receive money, and pay different merchants in-person or online, all from within the app.
The problem with the software industry in Tunisia is its reliance on outsourced product development and a one-size-fits-all customer success management, thwarting customer journeys and leading to churn. Software and technology companies, more specifically startups are putting experimentation to use, hence, the fast-paced product launches.
Research & Insights
A mixed research approach to understand what and why
User personas
We delved into a few personas, given the wide array of the unbanked population
From previous background research, we found that the bigger problem for the users was understanding e-KYC and how to get started with a digital bank. Onboarding conversion was the weakest link of the phases of our product's user journey.
Qualitative research: investigation and hypothesis development
Firstly, the team will try to formulate a hypothesis using a variety of methods, mainly focus groups, usability studies (i.e. artificial environment), surveys, and secondary data trends.
Radar plot - Keyword and Theme Occurrences in the Focus Groups
Quantitative research: controlled experimentation
The experiments are designed around two hypotheses developed and compares two versions of the same product (i.e. one of which is already live and seen by customers as the “norm”)
Variation A: Control group (the unaltered, original version)
Variation B: Treatment (the optimized version that is expected to perform better).
First step of a sequence of controlled experiments
Highlights & coded qualitative research
🔍 4 out of 7 people skipped the KYC steps list (where we inform them which info we're retrieving from them for the identification)
🔍 User #3 said "I didn't read how this is legitimately affiliated"
Insights
ℹ️ The low level of interactivity in the KYC steps screens downplayed the importance of users reading and learning the needed steps and their legitimacy before starting the process, and led them to speculate the lengthiness, “unusual” information in the KYC process.
ℹ️ Problems with trust and accessibility require more interactivity in the process (more information pieces and transitions between them)
Research method choice
We based our project on a series of experiments that inform a better design for the user experience.
Statistician George Box's iterative learning process
Moreover, Kaoun’s product team has prioritized sign up experience as well as initial trust as the main issues to tackle through the study, thus, endeavoring to solve the bank account creation usability problems and attempt to improve its user-friendliness. The processes to tackle are the ID scan as well as the proof-of-life video recording.
Testing and feedback
Going beyond UX heuristics
The research team perused through the transcripts of the different interviews and focus groups to find similarities and patterns consistent with problematic interactions, and more specifically mobile application screens. For instance, the researchers remarked that tutorials are very hard to follow, and that a lot of respondents tend to drop off, or complain that the process is lengthy, incomprehensible, and inconvenient to use.
The series of experiments on the hypotheses expounded different ideas about the thinking, behavior, and experience of the Tunisian users of mobile applications, namely Flouci.
Experiment I
Our first design involved a step we internally called "KYC List" where we tell users what they're in for, in terms of steps and information to provide.
Screens tested in the experiment I
Variant B converted 37.9% better than Variant A. We are 95% certain that the changes in Variant B will improve the conversion rate.
Experiment II
Our approach to tutorials in our user performing the main identification tasks was a pre-action tutorial. Through the research (i.e. Dovetail insights), we learned that a pre-action tutorial that does not involve active user interaction may yield a bad experience and failure to perform the task. Our exploratory research also emphasized user's ability to stay on screens less 4 seconds on average while going through this "lengthy" process.
Building variants with low fidelity wireframes
We deduced the following experiment to be run and executed:
Interactions tested in the experiment II
Variant B converted 64.7% better than Variant A. We are 98% certain that the changes in Variant B will improve the conversion rate.
Experiment III
We also saw that respondents in our research had problems with giving a company personal information they do not usually provide. We looked at different solutions in our variant and hypothesis development phase. We came up with the following solution: We give them a "why" screen where we explain why we need such documents (e.g. ID) to verify their identity and create a legal bank account for them.
Interactions tested in the experiment III
The test was not statistically significant to be conclusive. Even though Variant A converted 22% better than Variant B, the test demonstrates that Variant B will not necessarily improve the conversion rate.
Solution
Starting a digital banking experience for the non-tech savvy
We compiled the flows from each experiment and found a more data-driven design of our Know-your-customer process. ↗ See the prototype
More interactivity for less skimming and more readability
Interactive cards that show the eKYC’s (i.e. remote identification process) steps, their status, and demand action (e.g. button tap) to be initiated are better for users to complete opening a digital bank account.
Introducing the new way of verification, step-by-step
Using a step-by-step tutorial that requires people to click through a demonstrated task (such as recording a proof-of-life video) will push more users to complete the eKYC process.
How did Flouci do with Tunisians?
After launching the app with the new KYC, here are the echoes from the underbanked Tunisian users. 4.5 on Google Play & 4.9 on the AppStore.
Impact
Outcomes
15.6%
Increase in bank account creation
11%
Decrease in the average lostness rate
4.2%
Lower time taken to finish the KYC process
Learnings
Designing for users who learn mobile banking for the first time
Putting forward access to all process-related information: access to information that adequately describes the user’s process and aims to engage was key for the first two experiments to stimulate more conversion.
The dilemma between interactivity and lengthiness: it is imperative to draw a balance between interactivity and lengthiness (i.e. high number of buttons, screens, and transitions).
Controlled experimentation for strategic product marketing: The hybrid research model aims to improve the user experience in the product development in a systematic manner.
Summary
I designed a product as financial inclusion instrument to the unbanked, unfamiliar population through UX.
Background & context
Kaoun is a FinTech/RegTech company that builds reliable infrastructure for payments and credits in Tunisia. Kaoun’s mission is to enable all individuals and businesses to access financial and government services using any phone, anywhere, anytime. Kaoun created an array of digital products that connect users to financial and government services, through their mobile devices.
Current market trends in Tunisia 2019
Deliverables
Conclusive user research
User interfaces for mobile
Controlled experiment data
Lottie animations for KYC processes
My role is to conduct user research that yields actionable design decisions that I can in-turn transform into interfaces and interactions that are based on scientific information and logical rationale.
Problem Statement
Flouci is the first wallet that is designed to innovate mobile payment in Tunisia. It serves as a quick, easy, and convenient way to open a bank account, send and receive money, and pay different merchants in-person or online, all from within the app.
The problem with the software industry in Tunisia is its reliance on outsourced product development and a one-size-fits-all customer success management, thwarting customer journeys and leading to churn. Software and technology companies, more specifically startups are putting experimentation to use, hence, the fast-paced product launches.
How can our product be a better financial inclusion instrument to the unbanked, unfamiliar population through UX?
Research & Insights
Qualitative research: investigation and hypothesis development
Firstly, the team will try to formulate a hypothesis using a variety of methods, mainly focus groups, usability studies (i.e. artificial environment), surveys, and secondary data trends.
Quantitative research: controlled experimentation
The experiments are designed around two hypotheses developed and compares two versions of the same product (i.e. one of which is already live and seen by customers as the “norm”)
Variation A: Control group (the unaltered, original version)
Variation B: Treatment (the optimized version that is expected to perform better).
Research method choice
We based our project on a series of experiments that inform a better design for the user experience.
Moreover, Kaoun’s product team has prioritized sign up experience as well as initial trust as the main issues to tackle through the study, thus, endeavoring to solve the bank account creation usability problems and attempt to improve its user-friendliness. The processes to tackle are the ID scan as well as the proof-of-life video recording.
Highlights & coded qualitative research
🔍 4 out of 7 people skipped the KYC steps list (where we inform them which info we're retrieving from them for the identification)
🔍 User #3 said "I didn't read how this is legitimately affiliated"
Insights
ℹ️ The low level of interactivity in the KYC steps screens downplayed the importance of users reading and learning the needed steps and their legitimacy before starting the process, and led them to speculate the lengthiness, “unusual” information in the KYC process.
ℹ️ Problems with trust and accessibility require more interactivity in the process (more information pieces and transitions between them)
Testing and feedback
The research team perused through the transcripts of the different interviews and focus groups to find similarities and patterns consistent with problematic interactions, and more specifically mobile application screens. For instance, the researchers remarked that tutorials are very hard to follow, and that a lot of respondents tend to drop off, or complain that the process is lengthy, incomprehensible, and inconvenient to use.
The series of experiments on the hypotheses expounded different ideas about the thinking, behavior, and experience of the Tunisian users of mobile applications, namely Flouci
Experiment I
Our first design involved a step we internally called "KYC List" where we tell users what they're in for, in terms of steps and information to provide.
Results
Variant B converted 37.9% better than Variant A. We are 95% certain that the changes in Variant B will improve the conversion rate.
Experiment II
Our approach to tutorials in our user performing the main identification tasks was a pre-action tutorial. Through the research (i.e. Dovetail insights), we learned that a pre-action tutorial that does not involve active user interaction may yield a bad experience and failure to perform the task. Our exploratory research also emphasized user's ability to stay on screens less 4 seconds on average while going through this "lengthy" process.
We deduced the following experiment to be run and executed:
Results
Variant B converted 64.7% better than Variant A. We are 98% certain that the changes in Variant B will improve the conversion rate.
Experiment III
We also saw that respondents in our research had problems with giving a company personal information they do not usually provide. We looked at different solutions in our variant and hypothesis development phase. We came up with the following solution: We give them a "why" screen where we explain why we need such documents (e.g. ID) to verify their identity and create a legal bank account for them.
Results
The test was not statistically significant to be conclusive. Even though Variant A converted 22% better than Variant B, the test demonstrates that Variant B will not necessarily improve the conversion rate.
Solution
We compiled the flows from each experiment and found a more data-driven design of our Know-your-customer process. ↗ See the prototype
More interactivity for less skimming and more readability
Interactive cards that show the eKYC’s (i.e. remote identification process) steps, their status, and demand action (e.g. button tap) to be initiated are better for users to complete opening a digital bank account.
Introducing the new way of verification, step-by-step
Using a step-by-step tutorial that requires people to click through a demonstrated task (such as recording a proof-of-life video) will push more users to complete the eKYC process.
Outcomes
15.6%
Increase in bank account creation
11%
Decrease in the average lostness rate
4.2%
Lower time taken to finish the KYC process
Learning points
Putting forward access to all process-related information: access to information that adequately describes the user’s process and aims to engage was key for the first two experiments to stimulate more conversion.
The dilemma between interactivity and lengthiness: it is imperative to draw a balance between interactivity and lengthiness (i.e. high number of buttons, screens, and transitions).
Controlled experimentation for strategic product marketing: The hybrid research model aims to improve the user experience in the product development in a systematic manner.