Improving UX of financial tools by identifying consumer needs and tool redesign
Organization
Financial Consumer Agency of Canada
Role
UX Researcher
Team
Manager, 1 UX designer, 2 UX researchers, and IT
Date
Oct. 2023 - July 2024
UXR Skill
-
Competitive Analysis
-
Content Analysis
-
Sentiment Analysis
-
Low-fi Prototyping
-
Text Mining
Tools
-
Figma
-
Excel, R, Python
-
Power BI & Automate
Key Results
01
Implemented content analysis on two years of consumer feedback, creating Excel and Power BI dashboards that now guide tool improvement efforts.
02
Set up the analysis environment and introduced a new reporting system with Power BI, which has become the team standard.
03
Increased team's productivity by 45% by automatizing sentiment analysis via Power Automate.
04
Implemented 4 UI changes in a bank account comparison tool.
05
Influenced research directions for tool development by identifying 3 popular finance topics through text mining.
BUSINESS GOALS
Enhancing User Experience of Financial Tools
The Financial Consumer Agency of Canada (FCAC) offers 8 financial tools, but they have remained unchanged for a decade. ​
-
Mortgage Calculator
-
Mortgage Qualifier Tool
-
Budget Planner
-
Account Comparison Tool
-
Financial Goal Calculator
-
Financial Literacy Test
-
​Credit Card Comparison Tool
-
Credit Card Payment Calculator
​I led 3 initiatives to enhance the UX of these tools and ideate for new tools that can fulfill consumer needs in current economic climate.
Exercising My Research Expertise
I defined how to approach each initiative.
For some initiatives, I didn't conduct primary data collection due to resource allocation to high-impact projects and a tight deadline.
​
Thus, I leveraged existing and public resources, applying diverse mixed-methods research techniques. ​
01.
Developing Qualitative Data Analysis Framework
-
Identified the appropriate analysis for 2,000+ consumer tools feedback.
-
Set up an analysis environment using Excel and Power Automate.
-
Developed optimal visualization for cross-functional reporting using Power BI and Excel.
02.
Modifying UI of an Account Comparison Tool
-
Leveraged prior research and competitive analysis to deliver data-driven UI modifications.
-
Implemented 3 UI changes to address the consumer needs on bank fee prioritization.
-
Implemented 1 UI change to boost consumer awareness of low-cost and no-cost accounts.
03.
Discovering Consumer Interests via Text Mining
-
Identified 3 popular finance topics on social media through web scrapping and topic modelling.
-
Produced interactive data visualization for cross-team reporting.
-
Influenced research directions for tool development.
PROJECT 1: BUILD DATA ANALYSIS FRAMEWORK
The first step towards improving tools’ UX was to analyze a wealth of unexamined consumer feedback. Without it, we wouldn't know key consumer suggestions for tool improvement.
I deemed content analysis appropriate for two reasons:
-
It groups a large volume of text into meaningful categories.
-
It allows for the prioritization of categories based on frequency.
​
I prepared the dataset through cleaning and manipulation and then engaged in iterative coding process in content analysis.
Data Cleaning
Data Manipulation
Content Analysis
-
Checked for irrelevant text.
-
Verified correct import of numeric values from a survey platform.
-
Created columns with focal variables of interests in Excel.
-
Developed and assigned categories to raw consumer feedback
-
Grouped categories into topics.
This process resulted in an analysis environment, allowing the team to carry on the analysis in future. I also created Excel and Power BI dashboards, making them a standard reporting system for the team.
Analysis setup in Excel
Reporting System in Excel and Power BI
Here is a snippet of the dashboard showing key consumer suggestions, grouped by topic and prioritized by their frequency.
Suggestion Topic | Consumer Suggestion | Frequency |
|---|---|---|
Want to apply Excel formula | 7 | |
Calculation | Show values in 2-decimal place | 18 |
Use real dates to communicate payment dates | 5 | |
Result Presentation | Want to specify the year to view in a summary table | 19 |
Want to enter the start of mortgage | 5 | |
Want to enter the name of a bank | 20 | |
Data Fields | Want to enter bank credit protection information | 34 |
Calculation
Result Output
Data Fields
35
21
7
26
12
23
10
These dashboards are crucial for collaboration with IT, allowing for iterative tool improvement on prioritized needs.
I used these dashboards to share key consumer suggestions with senior leadership and cross-functional teams.
Automatizing Sentiment Analysis
I had a secondary goal of finding out how consumers feel about each tool through sentiment analysis. To support team's need in reducing data analysis time, I automatized sentiment analysis via Power Automate.
Before Implementation:
Manually conducting sentiment analysis for one tool required an entire week.
After implementation:
The same sentiment analysis now takes 30 minutes for one tool. Team members can simply click a button to initiate the analysis. 🕒
PROJECT 2: IMPROVE UX OF ACCOUNT COMPARISON TOOL
The first specific tool that I modified was the account comparison tool, which lets users to search for back accounts with 8 search filters, and they can compare the attributes of different bank accounts.​
​
I learned about the business requirement through meetings with my manager, followed by finding out IT and user requirements.
Business Requirement
1. UI changes should require minimal IT effort.
2. UI changes should let users know of low-cost and no-cost accounts.
3. UI changes must enhance overall UX. ​
IT Requirement
My goal was to learn about what constitutes as minimal IT effort.
I learned that any UI changes happening at HTML level require minimal effort​ and will help me to deliver UI modifications on time.
User Requirement
The deadline was tight, and there was not enough time to conduct a full UX research study.
I leveraged team's prior UX research findings and found users prioritize account-related fees when searching for bank accounts.
FACILITATING STAKEHOLDER ALIGNMENT
I regularly shared different requirements with the team through weekly team meetings and Teams. Aligning my team on the requirements was critical to the success of the project because it allowed for creative and faster problem-solving:
They informed me about existing UX research on consumer needs in the tool, helping me identify available resources.
They significantly improved my design solutions through critical feedback and iterative design cycle.
They guided me through tool improvement cycle, from how to submit IT requests for design solutions to conduct regression testing before tool re-launch.
DATA-DRIVEN DESIGN
The next step in redesigning the tool was translating user needs into actionable design options.
Through gap analysis, I found the tool did not satisfy the user needs on account fee prioritization: its search filters related to account fees were hidden under an expandable button on a landing page. Based on the results, I proposed the following:
-
Proposed Change 1. Have filters about account fees shown at the top for better visibility.​
To improve overall UX, I drew upon UX design principle on cognitive overload and 3-click rule and proposed the following changes:
-
Proposed Change 2. Get rid of unnecessary phrases.
-
Proposed Change 3. Get rid of an expandable button to present all filters simultaneously.
To ideate design options for increasing consumer awareness of low-cost and no-cost accounts. I conducted competitive analysis of government agency websites and proposed to apply a typical government's information tag:​
-
Proposed Change 4. Included the label 'Low-Cost and No-cost Accounts' to accounts that are low-cost and no-cost.
Here is an overview of my approach to tool improvement, enabling data-driven decision-making.
Prior UXR findings on User Needs
Gap Analysis
UX design principles
Competitive Analysis of Websites
Initial design
Second iteration
Final design
The final design is now available to the public!
PROJECT 3: DISCOVER CONSUMER INTERESTS VIA TEXT MINING
I had another goal of discovering consumer's recent finance interests and ideate for new tools.
​​
I evaluated text mining appropriate for two reasons:
-
I had resource constraints and couldn't conduct interviews or focus groups to discover consumers' interests.
-
I can make use of social media about personal finance and find out most prevalent topics among users.
I followed these two steps, web scraping and topic modelling.
CAVEATS TO TOPIC MODELLING
Model Exploration
The analysis does not tell you a number of topics "hidden" in the data, and one needs to explore with different number of topics until they find a model that shows the clear distinction between each topic.
​
For example, I explored models with 7, 8, 20 topics and decided to go with a 8-topic model, as each topic clearly differentiated from one another.
Result Interpretation
​
Interpretation happens at a surface level. The analysis tells you general topics that are most prevalent across the entire corpus, but one cannot truly understand the detailed nature of each topic.
​​​
You read representative posts (i.e., posts where a topic is most talked about) to get a sense of the nature of a topic, but you cannot generalize too much of its nature to other posts.
RESULTS
Topic 1.
This was the most prevalent topic and accounts for 47% of the total content in the data.
General & Real Estate Investment
Topic 2.
This was the second prevalent topic and accounts for 31% of the total content in the data.
Housing and Leasing challenges
Topic 3.
This was the third prevalent topic and accounts for 28% of the total content in the data.
Inflation and Saving Tips
"
People asked about investing wisely when dealing with significant money (e.g., receiving an inheritance) and shared tips on real estate investment.
What this topic is about
"
People talked about the downsides of being a homeowner, lease issues with landlords.
What this topic is about
"
People expressed concerns over inflation on daily purchases, grocery prices, and gas and shared how to save money (which stores for grocery shops).
What this topic is about
COMMUNICATING RESEARCH TO STAKEHOLDERS
Text mining was new to my team, and I used several strategies to get them on board throughout my process.
I showed the expected results from text mining. This makes it clear to everyone that text mining was appropriate to answer the research question.
I shared small wins with the process through Teams, from fixing a bug in Python on Friday to successfully setting up R for topic modelling.
I prepared a research report in three formats: Word, PPT, and interactive visualization using Flourish. This flexibility increased team's motivation to read results.
RESEARCH DIRECTION
DIRECTION 1
-
To understand what tools can be developed on the three topics, additional research is required to unpack the nature of consumer needs on the topics. ​
-
I recommend conducting competitive analysis on major banks in Canada and identify whether they offer any tools on one of the topics.
-
​Conducting focus groups is also ideal to probe consumers on how they envision a tool to help them deal with the topics.
COMPETITOR ANALYSIS
FOCUS GROUPS
DIRECTION 2
-
I recommend the agency to apply the findings for website improvement. Presently, FCAC does excellent in offering information on general & real estate investment, housing and leasing challenges, and inflation and saving tips. ​​​
-
It would be important to conduct research to understand what subtopics within each topic consumers want to learn more about. Providing subtopics that are upmost interest to consumers will ultimately result in continuous engagement with the website.
Subtopic 1
Subtopic 2
Subtopic 3
Inflation and Saving Tips
REFLECTION
Leading this multi-layer project reinforced my appreciation for being flexible due to having expertise in mixed methods. Flexibility means you understand that ​researchers do not need to conduct primary data collection, and they can still lead high impact research by identifying techniques that leverage existing resources and evaluating their appropriateness to answer business questions. ​
​​My research had high impact. I received the Maestro award, given to a team member per year, for introducing innovative analytical approaches to consumer tools feedback, and 2) my team actively uses the dashboards that I created for too improvement consultation with the IT team.
​
My advice to other researchers is explore research techniques that do not involve primary data collection :).
-
Text mining​​
-
Content and sentiment analysis​
-
Secondary research (e.g., competitive analysis)
Quantitative
Qualitative