V3 Valuation Tool Redesign

Fall 2023 | Industry Sponsored Project

A redesign of Carputty’s vehicle valuation tool to streamline data input, support better decision-making, and provide dynamic visualizations of vehicle value trends over time.

CONTEXT

Industry Sponsored, MS-HCI

My ROLE

UX Researcher


TOOLS

Qualtrics, MS Office, Figma, Figjam

TIMELINE

Aug - Dec 2023

(16 Weeks)

Tl;DR

I led the research initiative to uncover current pain points and opportunities in Carputty’s V3 Vehicle Valuation Tool. Through user-centered research, we identified key needs such as ensuring the tool’s reliability (e.g., data sources, market trends, supporting research) and improving users’ understanding of core concepts (e.g., VIN, valuation visualizations). Our research insights informed a complete design overhaul of the V3 tool.

  1. PROBLEM AND CONTEXT

What is Carputty?

Carputty is an auto financing startup based in Atlanta, encouraging individuals to view their vehicles as investments.

What makes Carputty special?

Carputty offers a unique vehicle valuation tool called "V3 Valuation Tool" that uses predictive AI and industry data to predict the future value of their vehicles over time.

COOL! SO… WHAT'S THE PROBLEM?

Although the V3 Valuation Tool is designed to empower customers with informed vehicle purchasing decisions, its current level of customer interaction is limited.

INITIAL PROBLEM STATEMENT

Through initial stakeholder elicitation, our initial focus was on how we could help customers make informed car buying decisions. This problem stamtent was later refined after the research process.

  1. SECONDARY RESEARCH

2-1. WEBSITE WALKTHROUGH

Given that the V3 Valuation Tool is a new feature, we began by exploring how users might discover it, learn about its purpose, and ultimately interact with it. Before conducting user research, we performed a secondary analysis through a website walkthrough to understand the current experience and uncover potential friction points Carputty customers might encounter

WEBSITE WALKTHROUGH: KEY INSIGHTS

Lack of Clarity about the Tool:


The V3 Valuation Tool is hidden without a clear call to action, making it unclear how or where to access it. This issue is compounded by an overall lack of clarity about the tool’s purpose and the site's core offerings.


Limited Interaction with the Tool:


Currently, users must enter both VIN and mileage to access a valuation, with no flexibility in input. The tool only returns one valuation at a time, and it lacks transparency—there’s no explanation of how the value is calculated or what factors are considered.

2-2. LITERATURE REVIEW

As the V3 Valuation Tool heaviliy relies on AI model, we wanted to further understand the trust in AI models with a particular focus in data visualization.

LITERATURE REVIEW: KEY INSIGHTS

Navigating the AI "Black Box"


AI tools often feel like a “black box” where users see the outputs but not how decisions are made. This lack of transparency makes it hard to build trust.



Visualizing AI Decisions


Researchers have explored visual ways to explain machine learning predictions

  • Counterfactual Explanation (i.e., What could have been different?)

  • Attribution Visualization (i.e., Why did a model made a certain decision?)


2-3. COMPETITIVE ANALYSIS

We conducted a comprehensive competitive analysis of vehicle valuation platforms currently operating within the market landscape. This involved an examination of leading websites and applications that offer vehicle appraisal services, scrutinizing their features, functionalities, user interfaces, and overall user experience.

COMPETITIVE ANALYSIS: KEY INSIGHTS

Unique Visualizations:


Carputty offers unique visualizations but falls behind competitors in information access, explainability, and data reliability

Accessibility and Ease of Use:


Competitors offers chat assistance and superior access features.


  1. PRIMARY RESEARCH

3-1. SURVEY

Next, our team deployed a Qualtrics survey and gathered responses from 25 participants. We asked about their familiarity with valuation tools, what vehicle aspects they consider most important, and what types of information they typically have on hand to input into such a tool.

SURVEY: KEY INSIGHTS

Difficulty understanding VIN

VIN and Mileage is not the most readily available information, even though they are needed to access the V3 Valuation Tool.

Unfamiliarity with Valuation Tools

Some users struggled to make sense of the results—especially understanding the symbols and knowing how to use the information to guide their decisions.

3-2. CONTEXTUAL INQUIRY

To watch users using the V3 Valuation Tool in real time, we conducted 3 contextual inquiry sessions with Carputty users who are somewhat familiar with the company and the tool.


The session compromised of 45 minutes walkthrough of finding the tool, and interacting with the V3 Valuation Tool.

CONTEXTUAL INQUIRY: KEY INSIGHTS

Questioning Reliability


While the V3 Tool provides a data source, users want further clarification on the rationale behind the tool's predictions to feel more confident in its reliability.

Interpreting the Results


Some users struggled to make sense of the results—especially understanding the symbols and knowing how to use the information to guide their decisions.

3-3. SEMI-STRUCTURED INTERVIEW

Finally, our team conducted 4 semi-structured online interviews with selected survey respondents to gather deeper insights. We explored their familiarity with valuation tools, preferences, and the context of buying/selling, while also examining the role of past, present, and future values in their purchasing decisions

SEMI-STRUCTURED INTERVIEW: KEY INSIGHTS

Researching about the car


Users conduct multiple ways of researching before buying a car, including getting quotes, comparing different cars, talking to others, and reading testimonials.

Market Trends


Users reported that understanding market trends is important to them when buying or selling vehicles.


  1. Research Artifacts

4-1 PERSONA

Taking account of our primary and secondary research, we developed persona that captures our customers’ needs, frustrations, and goals while using the Valuation tool.

4-2 USER JOURNEY MAP

We mapped our personas onto a journey map to track their experiences with the Valuation Tool, from discovering it to interpreting its results.

  1. REFINED FOCUS

5-1. HOW MIGHT WE?

We revisited our initial "How might we" statement to ensure it accurately reflects our research findings. This revised statement incorporates key themes identified during our data synthesis.

5-2. DESIGN REQUIREMENTS

  1. DESIGN

6-1. GROUP BRAINSTORMING

As a group, we engaged in brain dump & SCAMPER to generate 4 different concepts

6-2. USER FEEDBACK SESSION

We held a think-aloud feedback session to understand how reliable and intuitive each concept is, uncover any unnecessary or missing elements, and address accessibility concerns

6-3. WIREFRAMING

Based on our feedback session, we combined features from each concept to create a low-fi wireframe

6-4. HI-FI PROTOTYPE

Access Full Prototype

  1. EVALUATION

7-1. HEURISTIC EVALUATION

We conducted expert evaluation with 3 UX experts, where experts evaluated our prototype against 10 Nielsen Heuristics. Evaluators completed a series of tasks in the prototype while filling out Heuristic table and identifying usability issues.

7-2. USER-BASED TESTING

We conducted 4 user-based testing sessions in which users completed a set of 10 task scenarios using our prototype. Afterwards, participants completed a SUS questionnaire. While conducting the sessions, researchers kept track of task completion rate, time it took to complete each task, and number of clicks

  1. REFLECTION

8-1. FUTURE STEPS

8-2. REFLECTION

1. Navigating Project Direction with Clients

Working with clients throughout this project made me more comfortable with ambiguous answers and addressing feasibility concerns.


2. Carefully Selecting Research Methods

Because of the nature of our problem space, we realized contextual inquiry was not as useful as other methods we engaged in. Looking back, we could have uitilized our time better by distributing surveys early and recruiting as many people as possible for semi-structured interviews.

© S.Park

| Last Updated May 2025 🍀

© S.Park

| Last Updated May 2025 🍀

© S.Park

| Last Updated May 2025 🍀