Cox Automotive
Digital / UX Designer 2018-2021
During my time at Cox Automotive working on Autotrader.com in the Client Operations division, I led UX both customer facing and enterprise design initiatives that furthered the mission of being the #1 Marketplace for buying and selling cars.
Below, I’ll walk through my UX design process.
Creative Asset Library | 2018-2019
Easy to use graphic advertising library for sales and auto dealers.
Role
I was the lead UX Designer throughout the project cycle from planning and requirements gathering to strategy, wireframing, and content management. I collaborated with the design and development teams to brainstorm ideas and problem solve, as well as marketing and strategy/implementation teams to ensure that the rollout was supported nationally.
TL;DR: Dealers need a way to see the ads they’re buying, and a better way to quickly make changes. What if we could help sales and dealers at the same time?
Intro
Creative Asset Library (CAL) is a web tool that uses a step by step process to guide sales agents and their dealer clients through the process of choosing the right advertising graphics. CAL replaced the Graphic Sample Site, created in 2014 (5 years prior to this project), which functioned as a directory for sales agents but which suffered from a number of usability issues and was ultimately abandoned prior to the start of the CAL project.
Problem
With this new redesign, Autotrader wanted to bring more focus to the integration of existing technologies, by allowing users to select graphics, add them to a shopping cart, and send them directly to the design team for fulfillment. This time saver would ultimately reduce the back-and-forth between the design team and the sales agents because it removed the need for sales agents to send graphics to their dealer clients for approval prior to setting them live on Autotrader.com, which historically resulted in change requests in the graphics. The scope of the tool changed from internal only to client facing, so it was important that the quality and tone of the tool reflected the brand experience customers expected from Autotrader. It was important that the tool feel friendly and personalized.
Target
CAL targeted two specific personas, the sales agent (internal) and the dealer client (external).
Universal problems:
unsure of which graphics apply based on their subscription
current process takes a long time
The dealer client’s major problems were:
weren’t able to view the graphics they’re purchasing prior to paying
time from purchase to seeing the graphics on the site was too high
The sales agent’s major problems were:
material was often outdated, system was broken/glitchy
navigating the system was confusing
The goal of the redesign was to:
Reduce the number of pre-approval requests.
Reduce the number of requests that took longer than 48 hours to complete (because of back-and-forth).
Less pre-approvals means less time wasted for both designers and sales agents, and reducing average request time means higher customer satisfaction.
Product Strategy
The first challenge was executive stakeholder buy-in, so we started by analyzing the current user flow and talked to our sales users. (we didn’t get approval to talk to our dealer clients).
The three concerns we heard were that the system was: too confusing, not responsive, and outdated.
My team built a new IA that included integration with our newer systems, which would dynamically generate the templates and add the dealer’s information on them. This accomplished two things. It solved the problem of outdated graphics, because instead of a human manually updating the templates, the process was automated. Second, dealers were able to view the graphic with their dealership info on it, so it would look exactly as it would look on the live site. No placeholder text.
We pitched this new IA to our stakeholders and got approval to move forward.
Ideation
Next we brainstormed solutions. How to display two different graphic sizes, for 20+ OEMs, at 3 different levels of service (each level has different graphic options, and dealers on one subscription level can’t have graphics from a different subscription level. Confusing, right?)
This tool also needed to work for dealer clients, inexperienced sales agents, and experienced sales agents, which covered a wide range of different experience levels. The tool needed to be easy to use, regardless of the user’s depth of familiarity with the subject matter.
What we came up with was a step-by-step call and response based on a decision tree, with each question narrowing the number of available options. The three key decision points were:
Level of service
Make (Ford, Chevrolet, etc.)
Ad Size (728x90, 160x600)
Another important consideration as we designed the tool - It needed to be totally responsive - often our sales agents only have a phone or a tablet to access the tool, but our dealer clients may be in their offices accessing the tool from the web, so having the tool scale easily was an important consideration.
Execution
We created wireframes based on this and showed it to our sales agents, and the feedback was positive, but we learned that it would be difficult when switching between different dealers or different makes, to remember the specific templates they needed to request. They’d have to write down the template names as they went, and keep track of it.
User story: A sales agent is with a dealer that sells three different makes, and has two different subscription levels between the three makes. Sales agent has to track up to 6 different graphics before submission, requiring the user to break the flow of the tool six different times before completion.
We needed a way to keep the process within the tool to reduce dropoff. First, we thought of a clipboard, the user could copy and paste the different template names from CAL into a graphic request submission, and the designers would know which graphic was needed. However, this still meant leaving the tool. Then, we came up with the idea of a shopping cart widget that would sit on the side of the screen, and users could “add to cart” different graphics, which would track across different subscription levels and makes. This experience was familiar to the dealers as it mimicked the purchase experience.
The ultimate goal of the tool was to improve efficiency, so we designed a flow that allowed the user to add all of the desired graphics for a session to the cart, then click a button similar to “check out” and be able to submit those graphics in a request directly to the design team.
Technical constraint:
Due to technical limitations at the time, integrating the tool with the third party API for our case management system proved to be too high a level of effort, so we decided that for the MVP, we would build the button, but instead of having it send the data automatically, it would copy the template information to the user’s clipboard, and then the user would enter the data into the third party system manually. Not ideal, but it kept the user flow from exiting the tool until the session ends.
We also added a promotion calendar and a frequently asked questions page because these were determined to be low effort, high impact.
Results
Throughout the implementation period, I met with the dev team weekly and tracked progress, challenges, and when issues arose, bubbled them up to the appropriate leader. As we got close to launch, I worked with our sales training team to develop training guides, and our internal marketing team to develop engagement and retention campaigns across multiple fiscal quarters. After launch, I spent weeks meeting with every team of sales agents (20k+) from every part of the country, demoing the tool and answering questions. We also set up workflows for the design team to ingest requests coming from CAL.
To track the effectiveness of the tool, we used the following:
Salesforce data to track number of cases submitted, and how long those cases took to complete
We interviewed sales users to get feedback
Technical constraint: Because of the way the tool was built, we weren’t able to integrate site visitation tracking. 😕
Outcome
The number of cases we saw were insignificant in number.
The duration of cases did seem to go down, however.
To determine why we weren’t seeing the results we expected, we met with sales agents. When asked their thoughts on the tool, most users said they liked it, and that it was helpful. The case data didn’t support this, so we asked for them to clarify with us how they use the tool. What we realized was that the user was using the tool, just not in the way we expected. Instead of using the tool as a way to submit graphics, the user was using it as a reference, a way to stay up to date on the available graphics, and we found that users would screenshot the graphics they wanted, then email the screenshots to dealers.
We discovered that as it was built, the tool required sales agents to visit with dealers, in person, in order to use the tool. This made the sales agents’ jobs harder, because they couldn’t visit the dealer as frequently as was needed to stay up to date on graphic requests.
So:
Sales agents were more informed = less back and forth on the graphics = case durations lower
Some Numbers:
Over the following year, total time saved equated to 24.9%. 🥳
Graphics requests that required pre-approval were reduced by over 2,500. 🤩
The time taken to complete graphic work by agents was reduced from a total of ~1560 hours for approvals, to just 225 hours. That’s a reduction of ~40%. 🤯
Next steps:
Business goals and priorities have meant that additional iterations on CAL haven’t been able to be added to upcoming design sprints.
I have made recommendations for new features and improvements that will scale with the business.
Short term improvement includes the ability to create a link of the graphics in a user’s cart that sales can send to dealers in lieu of in person visits.
Below is a second design iteration with added flexibility. I left the role before further design refinements could happen.