PERSONALIZING THE PARKING LOT
Creating spreadsheets, writing the details of a car on a napkin and taking pictures of cars on laptops are just a few of the ineffective methods that users use to keep track of what they’re interested in. So, if those methods are ineffective, what should users be doing to help them keep track of what’s important to them when it comes to purchasing a car? The obvious answer is to save or add the car to their favorites. Right? Unfortunately, it’s not that simple.
“Keeping track of cars is hard and plenty of users weren’t even doing it. It’s like with dieting – you know that you should be doing it but you’re not.”
Saving or adding a car to their favorites typically requires a user committing to registering on a particular dealer’s website which most users are not comfortable doing, especially when they’re just beginning their car shopping journey. In addition, account creation requires trust and brand loyalty. Users are not willing to register or commit to creating an account without first trusting the company. So, if users are not willing to make that commitment, how can we help them keep track of cars that are of interest to them without forcing them to register for an account? That’s where the Watchlist (aka Parking Lot) comes in.
THE CHALLENGE
ENABLE NON-REGISTERED USERS TO SAVE CARS TO THEIR FAVORITES
Our goal for this project was to figure out a way to allow non-registered users to save cars to their favorites. The hope was that our users would be delighted by this new behavior, save more cars and ultimately this would help them to narrow down their choices to make a purchase.
Our high level goals were to:
Make it fast and simple for users to save cars to their favorites
Encourage users to save more cars
Create a space for users to keep track of their favorites without an account
MY ROLE
I led the research and design of the Parking Lot from June 2019 through March 2020 and collaborated with two other designers on the product. In addition, I worked alongside a Product Manager, Lead Developer, Analyst, Quality Assurance Engineer and 2 Software Developers. I stopped working on the product after completing the final designs and specks as the MVP was being built.
The product is set to launch on April 1st, 2020.
THE KICKOFF
UNDERSTANDING THE OPPORTUNITY
Before delving too deep into any type of research, our team wanted to better understand the data and see just how large of a market share opportunity this would be. I partnered with our team’s analyst and product manager to better understand if the quantitative data matched our assumptions.
Early insights told us that within a 90 day period, 41% of all attempts to save a car were unsuccessful — users were not signing in or completing the registration process.
RESEARCH
MAPPING THE CURRENT EXPERIENCE
One of the first research methods I completed was mapping out the current flow of the “save a car” experience. The UX flow provided the rest of the core team with a visual reference to quickly and easily identify potential problems within the current experience.
We discussed the flow as a core team and came up with the following assumptions::
Redirecting users to the sign in page after they click on a heart to save the car is harmful to their experience.
The sign in page provides no context or visual reference to the action users are completing.
Users are being redirected to different pages depending on what part of the site they’re saving a car from which causes confusion and inconsistency across the end-to-end experience.
Forcing users to sign in or register for an account is causing them to feel commitment anxiety and abandon the experience.
TALKING TO USERS
Quantitative data only told half the story. We wanted to better understand how our users felt about the current experience and attempt to validate or invalidate our assumptions as quickly and cheaply as possible.
I led a series of live, moderated user interviews with our Product Manager. One of our software developers also joined when time permitted.
KEEPING TRACK OF CARS IS HARD
Most of the users that we interviewed were not even keeping track of cars they liked, and those that were used ineffective methods like: creating spreadsheets, taking pictures of cars on their laptops and even writing notes on napkins.
PURCHASING TIMELINE MAKES A DIFFERENCE
Being 6 months out vs. 3 months out makes a huge difference in terms of users being more flexible to account registration or even needing to use a save-a-car feature.
USERS EXPECT TO HAVE TO SIGN IN TO SAVE CARS — BUT THAT’S NOT WHAT THEY WANT
Users have been trained to expect that they’ll need to sign in to save a car but the majority of them would never do it for fear of dealerships bombarding them with phone calls and emails.
ANOTHER PASSWORD TO REMEMBER — NO THANK YOU
Asking users to have to remember another password makes them feel anxious and frustrated.
DISCOVERY
ESTABLISHING A PATTERN
Confident in the results and early insights from our user interviews, I then did some independent research on other e-commerce sites to identify if a pattern existed for this behavior that we could test with users before investing our time and effort into creating something from scratch.
It turns out that there weren’t many sites that allowed their users to save any products or favorites without registering for an account. However, I was able to find a couple: Best Buy and Autotrader. Both sites allowed their users to save some products (or cars in Autotrader’s case) without registering for an account – but not without their limitations.
SETTING A BENCHMARK
The best part of establishing a pattern early on was that we were able to run benchmark testing on Autotrader’s experience and it wouldn’t cost us anything but time. We started off by launching 5 unmoderated tests via UserTesting which provided us with benchmark data to better understand how useful this tool would be for users and how they felt about the overall experience.
The learnings we captured were invaluable. We heard users say things such as:
“I like it. I’d still like to know how long these cars will be temporarily saved for.”
“I don’t want to have to sign in to save cars, but I do like that it gave me the option to temporarily save without having to create an account or sign in.”
“I like the idea that I can see my saved cars on the spot without having to create an account, I will probably create an account, but I like that I don’t need to.”
DETERMINING OUR DESIGN TARGET
Before I could start designing anything, our team wanted to be sure who exactly we were designing this experience for. Would users who were six months out use it differently than three months out? What about users who needed a car this week? Could they benefit from this new experience and how?
We ran additional tests with 15 users total: 5 that were 6-12 months out, 5 that were 3-6 months out and finally, 5 that were less than 1 month away from having to make a car purchase. The results helped us to identify that this type of experience could, in fact, help users in all timeframes. However, this experience would be most beneficial for users that at CarMax we call ‘mid-funnel’ and ‘lower-funnel’ users as they generally had some idea of what they were looking for and now needed help to keep track of the vehicles that they were interested in buying.
THE DESIGN
BRAINSTORMING & SKETCHING
To start figuring out the design of this new experience, I led a couple of quick brainstorming sessions with another designer and our core team. Quick sketching helped the team to get all of our ideas on paper quickly to talk through what we liked, disliked and what was realistic from a development perspective.
WIREFRAMING & RAPID PROTOTYPING
Once we landed on the sketches that we all voted on in terms of wanting to test those solutions, I created low fidelity wireframes to show the team a more polished version of the solutions that we had come up with.
We decided to only focus on solutions that kept the user on the same page but would allow them to view their saved cars. From the wireframes, I was able to put together a rapid prototype to demonstrate the flow of this new experience. Check out a video demo below.
After getting the prototype in front of a handful of users, it became apparent that allowing users to remain on the same page was ideal and that this experience didn’t have any major usability risks so far.
CREATING CONCEPTS
Based on the feedback that came out of the prototype testing and discussions, I created a series of concepts in a higher fidelity to better illustrate for myself and the team how this product could look at an almost pixel-perfect fashion.
CONSTRAINTS AND MAKING TRADE-OFFS
LIMITED RESOURCES
Although our team eventually grew to consist of multiple development resources, it started out with only a single developer which made running any larger in-production experiments or tests nearly impossible. Dealing with those constraints along with getting a lot of pushback from the core team about our solutions being “too complex” for an A/B test caused me to have to make trade-offs in the design and ultimately, the end-to-end experience.
We pivoted from allowing users to save up to 6 vehicles without signing in/register for an account to only allowing them 4. This change was due to limited resources and technical complexity. (We chose 6 initially just because CarMax allows users to compare up to 6 cars at a time.) This was one of the trade-offs that we had to make before launching a test in production.
TEST, LEARN, RINSE, REPEAT
CONCEPT VS. CURRENT
We ultimately decided to test the mini drop-down version of the parking lot against the current experience. We wanted to see if the qualitative data that we collected would match the quantitative data from this A/B test.
The results showed us that:
98% confidence in total lead submissions was directionally positive
10% more people engaged with saved cars in the test group
9% fewer visits signed in to MyCarMax accounts in the test group, however, lead submission rate went up as did quality leads
While the first test was in flight, we wanted to maximize the learning opportunity by also intercepting users that were in the test group. We interviewed 5 users and documented their live feedback.
“I wouldn’t expect there to be a limit, but more so have something come up every 5 or 10 favorites telling me to sign in or create an account.”
Users were overall delighted with new experience; however, they did call out some opportunities for us to iterate on the design. The empty state created some confusion; users were unsure if they needed to register for an account to save cars. The max limit dialog felt very sales-y and users were disappointed when they discovered they had reached their limit. Some even believed that there shouldn’t be a limit at all.
TESTING A NEW PAGE
As a team, we had always discussed the possibility of trying a brand-new parking lot page and seeing how users would feel about that experience vs. this initial drop-down. Perhaps somewhat selfishly, a new page would allow us to quickly run tests without having to partner with any other cross-functioning teams or delay based on the tests that they were already running.
A new page would also give us the opportunity to more easily explore testing a larger limit, recommended vehicles when one or more when one of their saved cars were no longer available and other enhancements that we documented early on. I put together a quick design for a new beta page and we tested it against current state and v1.
“Surprisingly enough, users didn’t seem to mind being taken to a new page.”
The data was overall flat, users engaged with the new beta page and v1 equally, but both were still significantly greater than the current experience. Our data proved the concept and the numbers weren’t lying – so it was time to build and productionize this new experience and change the business logic for CarMax.
GOING TO PRODUCTION
CUSTOMER PROGRESSION
In order to demonstrate the value of our new experience and product to the Product Leadership Team (PLT), I created a slide deck to show the path of customer progression.
TODAY’S CUSTOMER PROGRESSION:
Anonymous Users:
Basic Capabilities
Light geolocation to inform recommendations of vehicles nearby
Car page recommendations
Compare on SRP (Search Results Page)
Account User:
All Previous Capabilities +
Save unlimited cars
Save search
Email inventory alerts
Recommendations based on saved cars
Appointment setting
Lead submission
Prospective Customer:
All previous capabilities +
Ability to harden financing
Get their vehicle trade-in value
Select a MaxCare plan
Upload driver’s license
Upload finance documents
Sign e-docs
Lots more
As seen above, CarMax was providing varying levels of capabilities to Anonymous Users, Account Users and Prospective Customers.
Anonymous Users – Those that have not made the commitment to register for an account.
Account Users – Those that have made the commitment and registered for an account.
Prospective Users – Those that have submitted a lead on a car and have completed some of the transactional process: uploaded their driver license, finance documents or signed e-docs.
“We just weren’t providing any capabilities to what we defined as engaged users.”
FUTURE CUSTOMER PROGRESSION:
Anonymous Users:
Basic Capabilities
Light geolocation to inform recommendations of vehicles nearby
Car page recommendations
Compare on SRP (Search Results Page)
Engaged Users:
All previous capabilities +
Save a limited amount of cars
Share your list with a co-shopper
Compare your favorites
Account User:
All Previous Capabilities +
Save unlimited cars
Save search
Email inventory alerts
Recommendations based on saved cars
Appointment setting
Lead submission
Prospective Customer:
All previous capabilities +
Ability to harden financing
Get their vehicle trade-in value
Select a MaxCare plan
Upload driver’s license
Upload finance documents
Sign e-docs
Lots more
FINALIZING DESIGN
In an effort to reward engaged users and give customers the benefits of the parking lot as quickly as possible, we made an executive decision to go with a new page and look into the possibility of exploring more drawer-like solutions later on once we released the MVP. Since the design of the page was somewhat rushed for testing purposes, I went back and designed a more refined version and after going back and forth with my team, we decided on a final design. I specked out all the different states and screen sizes before handing that off to our developers.
As of March 2nd, 2020 I switched product teams recently and am no longer working on this product. It’s set to launch at the end of the month and is currently being built.
IF I HAD MORE TIME…
Before I left the team, I did some additional exploration in collaboration with another designer from the Checkout team. We wanted to see how this experience could fit into the future of buying a car completely online and that led us to trying a more cart-like UI, a pattern that the Checkout team was already leveraging.
With more time, I would continue to collaborate with other cross-functional teams like Checkout and build out a rapid prototype of this UI and have users walk through the entire end-to-end experience from browsing all the way through to Checkout.