Project 03 — 2025
Kite — Checkout redesign for a marketplace with a leaky funnel
A research-led rebuild of checkout that treated drop-off as a design brief, not a metric to accept.
- Role
- Product Designer — research to delivery
- Timeline
- 3 months
- Team
- 3 engineers, 1 PM, me
- Scope
- User researchUXUIA/B testing
Context
A mid-size marketplace had healthy discovery metrics and a leaking checkout: seven of ten completed carts vanished between “checkout” and “pay.” Leadership wanted a visual refresh; the data pointed somewhere deeper.
I led the research and redesign of the entire purchase flow, from cart to confirmation.
The problem
Session recordings and exit surveys showed three distinct leaks: unexpected shipping costs surfacing late, a mandatory account wall, and a form that punished mobile users with keyboard-switching and error resets.
Each leak had an owner who could explain it — but no one owned the flow. That, more than any single screen, was the design problem.
Framing
We reframed success from “a prettier checkout” to “protect the buyer's momentum.” Every decision was tested against one question: does this keep the buyer moving, or give them a reason to stop?
That framing gave us permission to kill beloved features — like the account wall — that momentum analysis showed were pure friction.
Key decisions
Decision 01
Guest-first checkout
Account creation moved to after purchase, framed as “track your order.” The wall had blocked 22% of buyers; converting them post-purchase cost nothing and offended no one.
Decision 02
Total cost up front
Shipping and fees estimate from the cart screen onward. Sticker-shock exits at the payment step fell sharply once the number stopped moving.
Decision 03
One column, one decision per screen
Mobile checkout became a strict single column with progress expressed as steps remaining. Errors became inline and forgiving instead of page-resetting.
The solution
Checkout that keeps its promises
The final flow shows the real total from the start, asks for exactly one decision per screen, and never forces an account. The visual refresh came last — clarity first, style second.
Outcome
checkout completion rate
drop-off at the payment step
guest checkouts, week one after release
Reflection
The redesign validated a habit I've kept since: follow the flow, not the ticket. The biggest win wasn't on anyone's roadmap — it was deleting a wall between two teams.
It also taught me to design the experiment, not just the screen. Clear hypotheses made the A/B results legible to the whole company.
More curated work
01 · 2026 · Research
Nara
A study companion that turns AI output into something learners can actually trust and follow.
02 · 2025 · Design system
Sentra
Turning five inconsistent tools into one dashboard — and one system the team can build on.
04 · 2024 · Branding
Atlas
One identity system stretched honestly across brand, app, and clinic-facing tools.
Interested in working together?
The block below is clickable
Contact
If this way of thinking fits your product, let's talk.
I'm open to full-time roles and collaborations — bring me a problem worth solving.