A study conducted by RepairPal revealed a troubling reality in the auto-repair industry: 50% of women calling repair shops in New York received price quotes that were 33-94% higher than those given to men. In response to this, RepairPal’s CEO, Art Shaw, set out to create an experience that would ensure fair and transparent pricing for everyone, eliminating gender-based discrimination in car care.
Our work focused on improving the design and content architecture of the tool to address customer feedback that it was confusing and overly complicated. By simplifying the experience, we aimed to enhance funnel completion rates and increase shop conversions, ensuring a more equitable and user-friendly experience for all customers.
The redesigned tool delivered significant results: a 27% increase in funnel completion rates, a 31% increase in conversions upon receiving an estimate, and 62% more customer contacts for shop leads compared to the original version.
The kickoff for this project began with reviewing existing market research collected over time. While no specific personas were established—given that car repair spans most demographics—the research highlighted critical insights into user empathy and confidence levels within the auto-repair industry.
A 2012-2013 study conducted by Northwestern University revealed troubling disparities in car repair quotes. Men and women calling the same group of repair shops for a $365 Toyota repair experienced vastly different treatment:
This study illuminated a key problem: the lack of transparency in car repair pricing disproportionately affects those perceived as less informed. RepairPal’s mission is to combat this issue through tools like the estimator, designed to bring fairness and confidence back to consumers navigating the car repair process.
Existing pain points with the estimator tool were significant. Only 50% of users completed all steps in the tool, with the largest drop-offs occurring at the ZIP code (step 1) and service input stages. Customer satisfaction was alarmingly low, with only 33% of users expressing confidence in the results. Conversion rates were even more concerning—just 0.1% of users booked an appointment after receiving an estimate.
To better understand the challenges faced by car owners, we focused on exploring product and market pains from multiple perspectives. Research revealed that car problems often evoke negative emotions such as nervousness, frustration, and overwhelm, especially when tied to cost fears. Additionally, most common car repair issues fall outside of warranty coverage, adding financial stress to an already frustrating experience.
We conducted interviews with individuals aged 21–50 who had undergone car repairs within the last 3–6 months. Our goal was to uncover users’ emotional states, concerns, and decision-making approaches during their repair journeys. Conversations explored how users sought guidance—whether through friends, family, online searches, or shop visits—and the factors that influenced their choices, such as proximity, word-of-mouth recommendations, or search engine results.
Our interviews focused on walking users through their end-to-end car repair journey, starting from the moment they recognized a potential issue. Before conducting the interviews, we analyzed common car repair searches and the typical processes users follow when choosing mechanics. This provided us with a baseline "skeleton" framework to compare against the real-world insights shared by participants.
We began usability testing by having participants navigate RepairPal’s existing site to identify pain points, struggles, and moments of delight. Insights from this testing helped us refine our hypotheses and compare results against new design proposals. Scenario-based testing with interactive prototypes allowed us to evaluate hypothesized features, weighing the pros and cons of various design iterations to identify the most effective solutions.
Through testing, we discovered that users formed emotional connections with car images based on their make and model. This led to delays, as users searched for images that perfectly matched their vehicles, distracting them from completing their selections. To address this, we introduced a neutral "skeleton" model for vehicle types. This visual representation avoided emotional triggers while remaining intuitive for users to identify their car's body type, simplifying the selection process.
We asked participants to apply their previous repair experiences to RepairPal’s estimator flow to identify pain points and opportunities for improvement. Insights from these tests guided the final designs, helping us scope down the MVP to focus on the features that delivered the most value to users. This process also pinpointed the areas of greatest friction, allowing us to prioritize fixes that improved usability and overall engagement.
We conducted three rounds of user testing using both moderated and unmoderated formats to uncover pain points and validate proposed design changes.
Through this process, customer feedback revealed critical insights that highlighted areas requiring immediate attention:
Key Metrics
The existing estimator design was a barrier to engagement, with users feeling overwhelmed and underprepared. Many participants noted that their lack of car knowledge was amplified by the estimator’s complexity. One participant shared, “I feel like the repair calculator on RepairPal is too specific.”
Our key takeaway was clear: the redesign needed to inspire trust, confidence, and clarity to keep users engaged. By simplifying the design, addressing organizational issues, and providing accessible tools, we aimed to create an experience that empowered users regardless of their automotive expertise.
In the redesigned experience, we tested displaying car imagery for every make and model in our database during the "select year, make, and model" step. While the intention was to make the process more intuitive, user testing revealed that images caused delays. Users, driven by their emotional connection to their vehicles, often searched for an image that precisely matched their car’s color and details, rather than selecting their intended option. This resulted in confusion and overlooked selections compared to the efficiency of plain text lists.
For diagnosis, users appreciated having visual aids but found overly specific representations of cars—such as exact colors or detailed models—distracting. Instead, a neutral, skeleton-style vehicle image allowed users to identify their car’s body type without forming an emotional attachment to the visual. This change helped streamline the process and improve focus.
Most users lacked technical proficiency or a clear understanding of what was wrong with their car but could identify the general area of a problem based on sensory cues like touch, sight, or smell. To accommodate this, we introduced an interactive vehicle image that allowed users to select problem areas intuitively, bridging the gap for less mechanically inclined users.
For those more experienced with mechanics, we maintained an alternative path: a faster option to search through a database or select from dropdowns featuring common issues based on their car’s make and model. By offering two tailored approaches, the design balanced accessibility for novice users while ensuring experienced users didn’t feel "dumbed down" or restricted, addressing feedback directly from our testing.
This dual-path approach reinforced our commitment to "humanity in repair," ensuring all users—regardless of expertise—felt supported and confident navigating the repair process.
To enhance user engagement, I structured the process with a progressive disclosure model. This approach simplified the user’s view by only presenting relevant information at each step, reducing cognitive overload. A key goal was to provide clear visual and contextual cues about how long the process might take, ensuring users felt informed and prepared. Research showed that when users lacked visibility into the time or effort required, they were more likely to abandon the page. The design also oriented users with a clear path forward after completing each step, including what to do next once they received their estimate.
To improve the estimator’s flow, I replaced the legacy dropdown menus with a step-based interface for selecting make, year, and model. This approach gave users a high-level view of their options, enabling quicker, more confident decisions with fewer clicks. I also introduced the ability to “clear” input fields, allowing users to reset their selections or request a different quote without restarting the entire process. In the original design, changing input redirected users back to the beginning, creating unnecessary friction. These updates streamlined the experience, reduced frustration, and kept users within the flow.
