Design Internship at FairVentures
Sep - Dec 2019 | Waterloo, Ontario
My experience working on a data visualization application as a Product Design Intern at FairVentures.
|
THE TEAM
|
Vanessa Seto - Product Design Intern
Rafael Sologaistoa - Product Management Intern Samantha Rempel - Software Developer Intern Kevin Wang - Software Developer Intern |
Adam Euerby - Senior Product Designer
Bridget Levac - Innovation Lead Ly Lam - Technology Lead Kirstin O'Brien - Data Scientist |
MY ROLE
I was one of two designers working on this project at FairVentures, the other being a senior product designer (my amazing boss Adam Euerby!). My team of eight for this project include the developers, product managers and data scientists mentioned above. I took design leadership in building early wireframes, visual mockups, making iterations from weekly feedback and designing final screens to be pushed for development.
RESULT
By leveraging the Fairfax insurance platform, my team successfully developed and launched the FairVentures Churn Analyzer Application for over 10 major insurance companies with thousands of users in 4 months. My team was able to design for insurance profitability, more specifically booking and retaining profitable customers, which was found to be a largely impactful problem for Fairfax insurance companies.
Note: For the purpose of FairVentures confidentiality, specific product details are classified or protected under NDA. Therefore, for this portfolio, I placed filler text for the designs and did not discuss all artifacts related to this project.
|
|
The Process
From research, ideation, wire-framing, designing, to prototyping, here's the end-to-end design process my team followed.
Problem Discovery
Customer retention is a large factor for insurance companies, especially when it comes to renewing or cancelling accounts. Companies depend on if a customer renews or not, if they cancel midterm, or remain active. Fairfax insurance companies are eager for data solutions in order to predict these customer decisions and increase customer retention through data applications.
On approaching this problem, it was important to gain a better grasp of specifically how data analysts currently analyze churn rates. We conducted interviews with individuals at Fairfax insurance companies and we discovered:
On approaching this problem, it was important to gain a better grasp of specifically how data analysts currently analyze churn rates. We conducted interviews with individuals at Fairfax insurance companies and we discovered:
- What users are trying to achieve - gather information on customers, create campaigns in area
- What users actually do to achieve those goals - use current ML tools, not specific to insurance
- What experiences (personal, social, and cultural) users bring to the tasks - highly skilled professionals
- How users’ professional knowledge and experience influence the product - prior experience in data field
PROBLEM STATEMENT
The team gathered observations from user interviews in order to pinpoint the underlying themes and trends from research, and we defined the problem statement as:
" How might we use data model predictions to assist Fairfax insurance companies to make informed and profitable business decisions? "
|
PRODUCT GOALS
- Useful & Impactful - Allow companies make informed and profitable business decisions
- Advanced - Make predictions and recommendations
- Relevant - Take into account customer insurance data & segments
- Scalable - Use real data from Fairfax insurance companies
Ideation
In order to create the initial design of the application, my team started with prioritizing the core features of the apps through a detailed feature description and behaviour-driven development approach. This required the aggregation of complex data visualizations, and the understanding of user flows and needs.
WIREFRAMING
The initial wireframes were created with the intent for scalability into a series of applications. The design of this application followed a blueprint framework in order to be expandable throughout the other Data Lab apps. I experimented on a variety of different structural frameworks and filter options to capture the overall visual approach.
|
|
|
|
|
|
STYLING
Working within the FairVentures design system, the visual design for this application drives the framework for all FairVentures Data Lab applications. The colour template reflects the neutral tones of the FairVentures branding, with the addition of cool tones to create a simplistic yet powerful feeling. For typography, we chose to stick with the simple System UI Font in order to reflect the stable and systematic feel we wanted to have consistently throughout the application.
Let's Explore the Visual Design
The FairVentures design system guided our visual treatments to have a consistent colour palette and minimalist, trendy accents, combined with a clean, classic typeface and layout. Keeping in mind the problem statement, the main design challenge for us was to represent this data in the most user-centered way, in order to abstract all of the complex customer data into a clear design.
DASHBOARD
For the Dashboard, we used coloured horizontal progress bars which provided a visual and trendy way of displaying the company’s overall progress. It gives a summary of the company’s progress in meeting their goals, as well as compares historic customer data with current and predicted data, in order for users to compare data between months, quarters and years.
|
|
GLOBAL FILTERS
Global filters are customizable and are constantly displaying active filters, which I chose to display at the top the screen. This way, the user can always see the active filters even when switching between pages, indicating that the filters are constantly applied for all of the displayed information.
|
MODEL SUMMARY
Designing for the model summary page involved a clear hierarchical heading structure to indicate sections and to group similar ideas. This heading structure can be found in the typography section. The summary page also includes percentage graphs which are designed in this colour-coded circular manner due to their high importance on the page.
Through cycles of iterations with the model summary page, we determined the best way to architect the information that was needed as well as leveraging the data models that power this application. This was done through circular graphs that displayed the accuracy for each data model within the application.
Through cycles of iterations with the model summary page, we determined the best way to architect the information that was needed as well as leveraging the data models that power this application. This was done through circular graphs that displayed the accuracy for each data model within the application.
|
|
|
DATA BY REGION
Part of the data visualization challenge for this app was finding a way to put data on a map, in order for users to focus on impactful regions. This map page displays region density through opaque circles and outlines the top regions from the map through an overview section at the top of the page.
|
|
DATA TABLES
Designing data tables and graphs involved understanding the purpose behind each piece of information and what courses of action the user and the company will take. The user can customize the information displayed by selecting global filters or sorting columns in tables. There is also an option to export the charts as a CSV file for analysts to review with their teams.
|
|
SETTING & NOTIFICATIONS
The settings feature will allow the user to change the language, company, username and password. Additionally, notifications act as a log for the user to be aware of data changes and updates happening in the backend of the app. Such notifications include data upload dates, data refreshing, files exported, company changes pushed, and any logistical notifications such as user password changes.
|
|
The Solution
The FairVentures Churn Analyzer App evaluates customer insurance data through data model prediction and forecasting. This outcome allows insurance companies to make informed business decisions that increase overall profitability. The apps predict customer decisions and produces recommendations for data analysts and line of business representatives to make meaningful customer decisions by evaluating customer churn behaviour.
Conclusion
This was my first project at FairVentures and it taught me the importance of collaboration with development and data teams, especially making iterations with development to assess constraints, requirements, objectives. All of my teammates worked on independent aspects of the project, while meeting up to review progress, critique and give feedback to each other. The journey and the final outcome of the project was truly a rewarding experience.
At the end of the four months, this application was developed and shipped to over 10 FairFax insurance companies, with a notable uptake in Latin American insurance companies. This suite of data applications is now an ongoing project that is still being worked on at FairVentures Lab.
At the end of the four months, this application was developed and shipped to over 10 FairFax insurance companies, with a notable uptake in Latin American insurance companies. This suite of data applications is now an ongoing project that is still being worked on at FairVentures Lab.
|
|