CarMax Auctions · B2B workflow
Vehicle Preview Enhancements
Restructuring how dealers move through vehicle photos, so inspecting damage stops interrupting the evaluation it exists to support.
- Role
- Senior Product Designer, Design Lead, UX Researcher
- Team
- 1 PM, 3 Engineers
- Timeline
- About 6 months, 2024
- Platform
- Desktop web, dealer actions tooling
The short version
Problem
Dealers clicked through 15 to 20 photos in a row and repeatedly opened and closed an enlarged view to check one vehicle for damage.
What I did
Led research, reframed the problem around keeping dealers in place during inspection, and redesigned the preview so thumbnails and zoom on hover replace the open, close and resume loop.
Result
Measurable lifts in engagement, dealer confidence and time spent per vehicle. Definitions and baselines are in Outcomes.
- +16%Pre-auction engagement: more vehicles saved and more notes created before auction day
- +25%Dealer confidence in photo quality and likelihood to bid, via dealer survey
- 50%Less time per vehicle for average dealers
01 · Context
Vehicle preview is where a dealer decides to bid.
CarMax Auctions is CarMax’s wholesale business, where licensed dealers browse and evaluate inventory before bidding. The preview page is where most of that evaluation happens, especially for exterior condition and damage.
- Business stakes
- 20,000–25,000 active dealers and 10,300+ auction vehicles listed per week.
- User stakes
- Dealers risk losing $1,500–$5,000 per vehicle when it’s assessed incorrectly.
02 · The problem
Dealers didn’t need fewer photos. They needed to inspect damage without losing their place.
- 15–20Consecutive clicks to move through a single image gallery
- Open, close, repeatDealers opened and closed the enlarged view again and again to check one vehicle
- 1–2 vs 4–8 minPer vehicle for power users compared with everyone else
- Both groupsShowed the same friction: a structural problem, not a skill gap
03 · My role
What I owned, and who I had to bring along.
I owned
- Framing the problem space
- Running and synthesizing research
- Defining the solution direction
- The interaction model and UI, refined with engineering through implementation
I influenced
- Keeping the expand photo button, so we could track engagement after release and validate the design
04 · Constraints and success criteria
What we couldn’t change, and what we agreed to move.
Constraints
- Delivery within a roughly six-month window
- Existing gallery infrastructure limited how much could be rebuilt
- Changes couldn’t disrupt behaviors dealers already relied on
Success criteria
- Pre-auction engagement: vehicles saved and notes created
- Dealer confidence in photo quality and likelihood to bid
- Time spent evaluating each vehicle
05 · Key decisions
Three calls that shaped the outcome.
-
1
Restructure the existing gallery instead of rebuilding it
Why this one
It fit the six-month window and the existing gallery infrastructure without disrupting behaviors dealers relied on.
What we gave up
A deeper restructure of the gallery model.
-
2
Show thumbnails when a dealer selects a photo group
Why this one
It replaced 15–20 sequential clicks with a direct view of the whole group, so dealers jump straight to the photo they need.
What we gave up
Some image load speed.
-
3
Add zoom on hover, against a mobile-first design methodology
Why this one
73% of Auctions dealers use desktop, so desktop got priority for close damage inspection.
What we gave up
Strict mobile-first parity. Mobile and tablet users kept close inspection through touch gestures.
06 · The solution
Before and after, tied to each decision.
07 · Validation
How we knew it worked before it shipped.
Usability testing
Tested the redesign with 3–5 dealers. It confirmed they could preview vehicle photos faster.
What changed
Dealers asked to keep the arrow buttons, so we kept them alongside thumbnails and zoom.
Launch approach
Phased rollout to 200 dealers to reduce risk and test system performance.
08 · Outcomes
What moved, measured against what.
| Metric | Result | Baseline | Measured with |
|---|---|---|---|
| Pre-auction engagement (vehicles saved, notes created) | +16% | Same-length period before launch | FullStory, over a few weeks after launch |
| Dealer confidence in photo quality and likelihood to bid | +25% | Survey before launch | Same survey after launch; 53 respondents completed both |
| Time per vehicle, average dealers | 50% less | [ADD BASELINE] | [ADD SOURCE] |
What I learned. Dealers didn’t need fewer photos. They needed to inspect damage without losing their place.
