CarGurus - Car Valuations · Accent Portfolio Template

CarGurus - Car Valuations

UX Design · Strategy & Research

Car valuations sit at the intersection of trust and opportunity.
For CarGurus, they’re one of the first steps in a seller’s journey, a moment where curiosity can turn into action: selling, trading, or financing.

But the old valuation experience was breaking that moment.

It was form heavy, slow, and confusing, with zero context and a single unexplained number at the end. Sellers didn’t trust it, and dealers didn’t value the leads that came out of it.

Meanwhile, competitors like KBB and TrueCar were using valuations as an engine for conversion, using transparent ranges, comps, and instant offers.

CarGurus was losing out on both trust and conversions.
Our goal was to fix that, not just by redesigning a form, but by redesigning the entire relationship between consumers, dealers, and the value itself.

“I just want to know what my car’s worth and why.”
— User interview, seller

Process

My Role & Objectives

I led UX design and research for this initiative on the AI Innovation team, partnering with Product, Engineering, and Data Science.
The objective was clear:

Build a valuation experience that people trust, and that drives real business outcomes on both sides of the marketplace.

For consumers: reduce friction and explain the “why” behind the number.
For dealers: improve lead quality and realism so every interaction starts on solid ground.

Collaborators: Product Manager, 3 Engineers, Data Scientist, Marketing Analyst
Timeline: 12 weeks (Research → Concept → Testing → Launch)
Constraints: Complex data models, tight dependencies with backend and dealer systems.

Understanding the Problem

We started by digging deep into how and why the flow was breaking down — combining qualitative research with quantitative funnel data.

Methods

User interviews (buyers & sellers)
Dealer partner interviews
Competitive analysis (KBB, Edmunds, TrueCar)
Funnel analysis with Data Science
Heuristic review of existing flow

What we found

Selling a car is stressful.
Sellers worried about getting ripped off. A single, unexplained number only made that worse.
Condition labels caused confusion.
People didn’t know if their car was “Good” or “Fair,” which made them question the entire valuation.
Trust was collapsing.
Without transparency, users felt like the process was random, a black box.
Dealers needed grounded expectations.
Misaligned valuations led to poor-quality leads and wasted time.

Data insight:
73% of users started the form, but only 24% finished.
Less than 1% submitted their contact info.
We were losing people before they even saw a value.

“Trust is the currency of car valuations. Without it, the entire funnel collapses.”

Defining the Opportunity

After mapping user pain points, dealer frustrations, and funnel data, one insight stood out clearly:

Trust was the currency of the entire valuation experience.

Without it, users dropped out, and dealers lost confidence in the leads that came through.
To fix this, we needed to redesign not just the interface, but the relationship between user input, system logic, and output clarity.

We synthesized everything into one guiding question:

How might we build a car valuation experience that feels transparent and trustworthy for consumers, while delivering higher-quality, more realistic leads for dealers?

Through collaborative workshops with Product, Data Science, and Dealer Relations, we aligned on three design principles:

Build trust through transparency.
Show where the value comes from and why it matters. Turn a single “mystery number” into an informed, data-backed range.
Simplify, don’t overwhelm.
Replace long forms with a conversational, one-question-per-screen flow that keeps users moving and reduces anxiety.
Balance both sides of the marketplace.
Design for both consumers and dealers, helping sellers enter confidently and realistically, while helping dealers receive more qualified, ready-to-sell leads.

We also mapped user and business needs to ensure alignment:

User Needs	Business Needs
Clear, believable valuation with context and comps	Higher completion rates and quality leads
Simple, low-effort flow	Reduce friction, increase engagement
Understanding of how value is determined	Strengthen brand trust and credibility
Realistic expectations for selling or trading	Improve dealer satisfaction and ROI

Ideation & Design Exploration

Our early explorations focused on clarity, confidence, and storytelling.

Key directions we explored:

One-question-per-screen progressive flow (reduce cognitive load).
Condition selector with visual examples and definitions.
Output screen showing a valuation range, not a single number.
Comps and side-by-side trade-in vs. private sale context.
Clear trust signals: source attribution, progress bar, transparent copy.

What tested best:
The progressive disclosure model, one step at a time, one question per screen, kept users moving.
It felt conversational, transparent, and simple.

Testing & Iteration

We ran moderated usability sessions with sellers and follow-up tests with dealer partners.

Findings and refinements:

Condition clarity: Adding examples (“Small scratches, 60K miles = Good”) boosted confidence and reduced confusion.
Trust signals worked: Showing comps and ranges made users feel the valuation was earned, not random.
Instant Offer reframed: By repositioning it as a natural next step, continuation rates nearly doubled.
Dealer dashboard improvements: Richer context (mileage, condition, VIN) helped dealers instantly qualify leads.

Final Design

The final experience replaced the long, opaque form with a guided, transparent flow that built trust step-by-step.

Core features:

Progressive, one-question-per-screen flow.
Clear progress bar and contextual help.
Transparent range output with comps and market data.
Side-by-side trade-in vs. private sale comparison.
“Why this value?” drawer explaining data sources.
Instant Offer positioned as a natural continuation.

Dealer-side upgrades:

Lead attribution for valuations.
Richer vehicle context in dashboard.
Alignment between valuation data and dealer offers.

Visually, the design was clean, open, and data-driven; showing confidence without complexity.