CarMax · 2025
Turning saved cars into purchases
Designing a behavioral email nurture system for used car shoppers.
Overview
When customers shop for used cars online, saving vehicles could be read as a strong signal of intent. But, is saving a car really the same as saving a pair of socks to an Amazon wish list? For big ticket retail items, creating a list of favorites or hearting items may be viewed more appropriately as a filing or indexing system, as reducing the noise around car buying through carving out space for decision making.
This project focused on transforming the thinking about what favoriting means and providing customers a lifecycle campaign that could help them make sense of their dotcom activities and saving cars. By designing a behavioral email nurture campaign and defining appropriate segments, we wanted to create a clear bridge to next steps for customers .
The problem
In many ways, nurturing customers who have saved cars to their wishlist was a just do it. A previous variation of the campaign had been wildly successful 4 years prior, but the content wasn't evergreen and the data connections routinely failed, so the campaign had been shut down. This left mid-funnel customers with no real nurture campaign to help them transition from broader exploration to evaluating the saves that they made.
The real test of a new campaign is to identify what email types would resonate within the segments. Would some email types excel with certain segments while harming the progress of other segments? How might we use engagement data to determine how these email types resonated with customers and how the types could be combined into several segment-based, action-based full nurture campaigns?
Customer problem
Very little shepherding toward next steps
Business problem
Lost sales from customers stuck in a churn cycle
The core issue for many customers wasn't awareness. It was decision friction.
The opportunity
We realized that essential to any new campaign was developing robust segmentation. Early research showed that the number of cars a customer was saving could provide us with easy and early segmentation.
Saving a car is not a casual action. It represents consideration, comparison, and emotional attachment. But it does not mean the user is ready to buy. This created an opportunity to design a system that:
Understanding the customer
Through behavioral data and research, I found that users save cars for fundamentally different reasons. Some save one or two because they're close to purchasing; others save ten or more because they're overwhelmed or still exploring. This led to four temporary behavioral segments.
Focused Buyers
KNOW WHAT THEY WANT
Need reassurance to commit
Comparers
Evaluating trade-offs
Need decision support
Browsers
SPIRALING
Need guidance and inspiration
Sudden shoppers
HIGHEST INTENT TO BUY
Need a car now
Saving is not commitment. And for some segments, it's also a way to delay a decision. Those customers aren't disengaged.
They're stuck.
Demand test
Roughly 160,000 emails were sent during our demand test. The goal of the test was to gauge interest in certain email types. Although we had a hunch which emails were primary players and which emails were supporting communications, the demand test allowed us to test our assumptions.
We tested 6 email types, some with two variants. Below is one of the types, and how we used the engagement to understand customer behavior. The full engagement results are available here: Favorites demand test engagement deep dive.
Customer interviews
Throughout the early process, we conducted customer interviews to inform our direction in various ways. After the demand test, I conducted a review of previous interviews to identify commonalities. Review the entire analysis.
The solution
I designed a behavioral lifecycle email campaign that responds to customer actions and guides them toward a decision. Instead of static reminders, it delivers timely, personalized, context-aware messages, reducing friction and building confidence in their choices.
Core design principles
Behavior over guessing
Emails triggered mostly by actions, and only rarely by schedules.
Support over the hard sell
Help users choose, not just click.
Clarity over volume
Reduce cognitive load at every step.
The email system iteration
Based on the learnings, I created eight touchpoints (many with multiple variants), each serving a specific purpose in the journey in helping customers feel confident to move from browsing to saving to deciding confidently to take next steps forward towards purchase.
01
Save confirmation
This email reinforces the customer's action and starts the digital conversation. It serves as a welcome to the nurture campaign and a transactional confirmation that they have saved a car. This email would be universal across all segments. All parts of the car tile are clickable and serve as the primary CTA.
02
Price drops
These would be sent universally across segments. There are two types - a price drop on a single favorite or price drops on cars similar to a favorite. The secondary information would change based on what we know about the customer. For example, "Know your financing options" would show for customers who haven't prequalified.
03
Comparison
Most customers have saved more than 2 cars. And for some, research showed that decision paralysis had truly set in, leaving customers in limbo. A comparison email was the single most important email of the nurture campaign, as it became a way for customers to start thinking through the pros and cons of cars they had saved.
04
Education and Research
These emails are personalized to the particular cars or car types that a customer is interested in, and serve to give customers confidence in their choices. If they haven't narrowed their choice of car make, we'd send them information related to the type of vehicle they are interested in, for example, "Top 5 sedans" or "Most fuel efficient SUVs".
05
Getting to know and personalized recommendations
For new customers whose preferences we're just learning, a recommendations promotional email with social proof as a secondary element serves as an introduction. For customers whose preferences we know, we'd send a personalized recommendations email instead.
06
Abandonment
For customers who have abandoned shopping, we'd send this nudge.
07
Product promotions
"Find a better way" promo campaign example. For customers who haven't prequalified or for customers who would like to know how much value is in their current car, the value prop of these promos is that CarMax makes it easy to complete a pre-qualification or instant offer.
08
Financing
It goes without saying that financing plays a big role in car buying for many customers. Prequalifying is a shopping tool that allows customers to narrow choices based on payment parameters.
The email on the left is a pre-qualification promotion based on customer's favorited car and includes help articles about financing a used car. The email on the right is a recommendations email based on customer's pre-qualification terms that includes a promotion for CarMax's return guarantee.
Design execution
The flexible email design system allowed for consistency and scale. Each email was optimized for mobile-first readability, quick scanning, and immediate action.
What I learned
Intent signals are only valuable if activated.
More choice does not mean better decisions.
Trust and timing are as important as price.
Good lifecycle design doesn't push users. It guides them.
What I'd do next
By focusing on decision-making — not just engagement — I designed a solution that meaningfully improves both the user experience and business outcomes.
Next case study
Turning inventory uncertainty into a revenue opportunity