ebay motors
Taking the guesswork out of buying car parts and tires.
Overview
Context
Buying parts for your car online is basically an act of faith. You're trusting that the part you're looking at will actually work with the vehicle you drive, and if you've ever shopped for tires, you know that problem gets even bigger. Tires are heavy, they're complicated, and almost nobody wants to install them on their own. So really, the hard part was never getting someone to buy tires online. The hard part was everything that had to happen after they clicked buy.
0.02%
of tire sales on eBay included installation, against an industry benchmark of 10-15% online. eBay had technically built a way to buy installation, but almost nobody could actually find it or finish it.
My Role
What I designed
The fitment state model and single input layer for Guaranteed Fit
The scheduling calendar, window model, and flow architecture for Tire Installation
The badge system used across both
What I led
Cross-functional alignment with Search, the product page team, and checkout to secure page placement for fitment
A direct escalation to the VP and Head of Product to prioritize a critical taxonomy bug fix
Partner strategy across SimpleTire and CarAdvise
Mentorship of a newer designer
Sustained ownership of both initiatives in parallel
eBay's Parts & Accessories team needed two connected initiatives that turned out to be the same problem in different clothes: Guaranteed Fit, which solved for fitment trust, and Tire Installation, which solved for what happens after checkout. Different problems, same underlying thesis: close the trust gap, and the business results follow.
Guaranteed Fit
eBay's fitment tools existed, but they were buried behind price, reviews, and seller ratings, so customers were deciding on a part before ever learning if it fit their vehicle. A taxonomy bug compounded this: fitment filtering silently stopped working the moment a customer browsed into a parent category, with no warning, showing parts that didn't fit as if they did.
Entering accurate vehicle information was its own hurdle. A lot of people don't actually know their own vehicle's year, especially in the UK, and plenty will guess something close rather than admit they're not sure.
Tire Installation
The old experience broke down at nearly every checkpoint, Search, checkout, and after the sale. The only real way to know installation was even possible was to luck into a listing where the seller happened to offer it. Even a buyer who made it through checkout just got a vague "someone will contact you" instead of a confirmed time, and once the order was placed, there was no way to track or manage what they'd bought. The few info pages that did exist didn't even agree with each other on who was responsible for reaching out, the buyer, the installer, or eBay.
Design Problems
Awareness
Buyers should know eBay offers tire installation early in their consideration journey, not stumble into it by chance the way most of them currently do.
Confidence
Many people hesitate to shop for tires on eBay at all, not because of price or selection, but because figuring out installation logistics feels like its own project. Limited visibility into the process and timeframe only deepens that uncertainty.
Trust
Buyers want confidence that once they've paid, everything unfolds the way it's supposed to. When eBay's own pages couldn't even agree on who was responsible for what, that confidence never had a chance to form.
Challenges
Cross-functional alignment
Page placement at eBay's scale is genuinely high-stakes, a decision can swing millions, and every team has a reasonable case for wanting their own initiative seen first. Making the case for fitment meant working closely with Search, the product page team, and checkout, real, sustained justification across many conversations, not a quick sign-off. Tire Installation faced the same dynamic getting installation badging onto Search results, though by then real customer evidence from Guaranteed Fit made the case faster the second time around.
Executive escalation
The taxonomy bug needed a different kind of push than placement did, this was about winning a dev team's time, not page space, and engineering time was already spoken for. I built the case with real customer evidence, ran it as a live demo, and took that demo into meeting after meeting, Search, cross-category teams, anyone whose part of the experience the bug quietly touched. I worked directly with the VP overseeing that part of the org to make sure he understood exactly how broken it was, and once he saw it firsthand, dev got a clear mandate to prioritize the fix.
Technical ambiguity
Tire Installation's scheduling was split across two partner integrations, SimpleTire and CarAdvise, with an open question underneath both: build against their live availability API, or build independent logic and treat the partner as a black box. Nobody could answer that yet, not even SimpleTire could confirm how they'd communicate a drop-off time.
Constraints
Phase 1 of Tire Installation had to ship inside a two-week sprint, on top of legacy code that was never built to be fully integrated with the rest of the purchase experience.
1
MVP
An extremely tight timeline with delivering phase 1 within a 2 week sprint.
2
Scope
With legacy code and platform issues, design is typically not fully integrated.
3
Platform
Installation services were to be built on top of eBay's "Value added services" platform.
4
XFN partners
The search team is highly protective of any design solutions that affect search pages.
5
Scheduling
eBay's shop partners lack an API for appointment scheduling.
Solutions
For Guaranteed Fit, the core system was a state model, not a screen: a part either fits, doesn't fit, or might fit and needs more information. That third state carried real weight, the guarantee was literal, if a part carried the badge and turned out not to fit, eBay covered the cost, so a false positive wasn't just a bad experience, it was a bill. Underneath it, a single input layer for vehicle data, year, make, model, trim, was built to work for both first-time buyers and experts, with a VIN shortcut for anyone who didn't need the guided version. Once fitment status was reliable, the badge became a consistent visual system applied everywhere a customer might be deciding on a part.
For Tire Installation, I designed a fourteen-day scheduling calendar and a morning/evening window model borrowed from patterns customers already trusted from brick-and-mortar tire shops, built to work no matter which way the API question landed. I also mapped and tested two different flow architectures, installation offered upfront versus added at the end of the purchase, weighing the earlier-is-better instinct against a real abandonment risk in areas with no nearby installer coverage.
Screens
Finder modules now live at the top of search results and category browsing pages. They are more disruptive and call themselves to attention immediately. As you scroll to see more results, the module shrinks to a small bar at the top, it’s always there reminding you to find parts that fit. This pattern exists in both parts and accessories and tires.
Adding a vehicle is easy using a type-ahead or tap system. You can also filter and narrow down tappable results by typing. Once you've selected your vehicle, it's added to your "Garage." Your saved vehicles are stored with your account and can be edited, changed, or removed.
Once your vehicle is added, it replaces the Finder's Add Vehicle module. You can also see that every part that fits it has the "Fits Your Vehicle" badge. On product display pages, a much larger version of this lives at the top of the page. If you tap "Change," a list of your saved vehicles is shown, or you can enter a new vehicle at the end of the carousel.
In the case of tires, additional attributes are shown when you are browsing tires such as the size or sizes you have selected ass well as the rim size. you can also just switch the context to size.
How I got it built
Both systems depended on alignment beyond design. Fitment's badge placement across search, product pages, and checkout required building a shared case with Search and checkout leads, grounded in real customer evidence rather than a design opinion. For Tire Installation, I wrote design principles for every surface the work touched as its own short worksheet, what has to be simple, what we're not solving yet, what the empty state says, so engineering could start building before every technical question was settled.
Results
Guaranteed Fit
-9% Returns vs. Q1 baseline
+0.86% GMB growth
+12% Finder completion vs. Q1 baseline
Tire Installation
Shipped real, measurable improvement, badging, checkout clarity, a working calendar and confirmation flow, against a 0.02% starting baseline. eBay's three-year goal was to bring attach rate up to the industry standard of 10-15%. The program was always planned in phases well beyond this release, Phase 3 was a full migration to eBay's VAS platform, so the complete attach-rate story plays out over a longer timeline than any single phase can show.
Beyond the metrics
The fits/doesn't-fit/might-fit model built for Guaranteed Fit was never really only about car parts. Partway through the work, it became clear this was the shape of any compatibility problem: does this lens work on this camera body, does this blade fit this mower, does this cartridge work in this printer. I made that case at the time and pitched extending the pattern beyond Parts & Accessories and Tires. It wasn't part of the roadmap I was driving, but the idea got real traction with other category teams, and it's entirely possible some of that thinking lived on in categories I wasn't tracking closely after I moved on. Tire Installation's own roadmap carried the same instinct forward on its own terms, more installation partners, more regions, eventually other categories across Parts & Accessories. Two different projects, the same underlying lesson: a good system is built to outgrow the one problem that got it funded in the first place.
That year, I was leading both initiatives at full stretch, two major programs running in parallel, while also coaching a newer designer through her own ramp-up. eBay recognized that with two Spot Awards, one for stepping in to lead Finders through a leadership gap, one for the cross-project work itself.