Personalization
Deliver the right message at the right time to the right audience to create a uniquely personalized customer experience.
Problem Statement
User Problem
40% of users frequently self-select the wrong product.
A portion of users with high credit scores choose a card with no rewards, missing out on the benefits they could be receiving.
Others select a card they don’t qualify for, triggering a hard credit check.
Business Problem
When a user selects a credit card they don't qualify for and has their credit score run, it negatively impacts the card business.
Regulations are in place to protect and benefit users, such as:
Fair Lending Act
Fair Credit Billing Act
Fair & Accurate Credit Transactions Act
Improving product selection is beneficial for both the business and the user.
Key Performance Indicators
Personalization
Pilot a new capability to predict and personalize the Credit Card Home Page (CCHP) for UM credit segment visitors. By tailoring the experience, we aim to create a more relevant and engaging shopping journey, fostering a personal connection and reducing friction in finding the right card.
Data Science Discipline
Build upon machine learning models from design of experiments. Leverage models used in email and print campaigns, and apply them to optimize the website experience for enhanced personalization and targeting.
Find the Right Product
Streamline decision-making by connecting users with the right product at the right time.
My Role
As UX Lead for the credit card website, I played a key role in pitching to leadership the idea of utilizing decade-old machine learning models and expanding A/B testing capabilities. At the time, testing was a limited part of the site, with the enterprise team managing a maximum of 40 tests through an excel sheet. We pitched new capabilities while using technology we already had access to that need to be integrated into the tech stack. Winning this pitch to leadership help me build out a team of 10, moving into more of a coach/player role.
Fast Marketing Printed Board for Science Fair
Location: Greater Richmond Convention Center - Dart Science Fair
Attends: Few hundred Capital One employees
Socializing - DART All Hands Science Fair Booth
Printed Board for Science Fair
Location: Greater Richmond Convention Center - Dart Science Fair
Attends: Few hundred Capital One employees
Current vs Future Printed Board for Science Fair
Create A Test Design
To create a test design for a proof of concept for personalization, myself and our marketing team member began identifying key user segments and behaviors. Working closely with data scientist, we selected relevant machine learning models capable of predicting user preferences based on historical data. We then designed a controlled A/B test, integrating these models into the website’s personalization framework to deliver tailored product recommendations.
The test design focused learnings from previous moderated and unmoderated qualitative research with our key user segments, plus website traffic analysis. This proof of concept aimed to validate the effectiveness of machine learning in enhancing user experience through real-time personalization.
Test Design
Proof of Concept
This proof of concept focused on measuring the impact of personalization on key performance indicators such as engagement, conversion rates, and overall customer satisfaction, ensuring that our personalized recommendations not only aligned with user needs but also drove meaningful business results.
Results
20 days, using only 30% of the traffic population to the page.
Test A had an increase of 49% for prequalified.
Test B had a 21% increase in bookings for VentureOne and a 17% increase for bookings on Quicksilver.
Test A
Test B
Review My Other Case Studies
Feel free to take your time reviewing my previous case studies. I've included a diverse range of work, covering personalization, extended reality (XR), GPS solutions, and large-scale ecosystems. Each project showcases different aspects of my design, leadership and research expertise.