Shipt
Self-serve data tool
Redesigned C3's mapping step to give users real-time feedback, cutting task time 20% and hitting 4.5/5 efficiency with technical users.
Ops tooling
Launched
Business context
C3: A tool to scale retailer onboarding
Shipt is an online delivery platform that partners with retailers, each sending product data in their own formats. The C3 is an internal tool that lets ops team members manage data pipelines.
The items you see on Shipt.com? Those are managed via the C3!
Problem
Users had no way to know if their work was right
Users take a long time to complete the mapping step of the data pipeline. Without feedback from the UI, users don't know if they've written the mappings correctly. This forces them to troubleshoot manually and escalate to engineering — adding hours of delay to every new retailer onboarding.
How might we close the feedback loop so that users can write code faster and more accurately?
Understanding what users do
As a Catalog Operations Specialist...
"I map and load retailer product data into Shipt's systems so members see accurate selections and pricing."
Their problem wasn't skill — it was feedback. These were technically capable users who could write code and manage complex pipelines, but had no real-time signal to tell them if their work was correct. I inherited discovery findings from my design manager and synthesized them into specific problem statements.
Old UI
Current state + user flow
The current UI space is not helping users in their workflow
Catalog specialists write the mapping code and need to validate and potentially refine their code. The data they need to preview is not accessible.
Users spend most of their time iterating on the mapping specs.
Design Deep Dive
Moving the preview where the work happens
The right-hand panel had always existed, but it was occupied by reference content most users ignored. The problem became clear when I watched users work in the old UI: they were constantly collapsing and reopening the bottom preview drawer to check their code — a repeated context switch that broke their flow.
The new layout prioritizes the preview panel, the area that users really need to see.
The new preview panel is right next to user's work area so as they write the mapping specs, they can verify the data in the preview.
Design Deep Dive
Surfacing the data users actually needed
Expanding the preview panel beyond basic product info gave users richer data points to validate their work which addresses the core design challenge.
Old preview didn't show enough information to help users validate.
Preview panel shows more data fields to give users more information to validate their mappings.
Design Deep Dive
Giving users context and control
Decisions:
Transparent Logic Descriptions: Translated backend code functions (
clean_html,set_display_name) into clear, plain-language summaries so users understand the impact before toggling.Hierarchical Organization: Grouped rules under collapsible feed accordions (Product feed, Inventory feed).
Data transformations are described in plain language so users understand their impact before toggling on/off.
Design Deep Dive
Helping users write mapping specs
While writing mapping specs, users sometimes need to refer to common functions to manipulate data. The library tab serves as an additional resource to help users with their workflow.
Caption
User testing
Validating the preview panel with users
Before finalizing the design, I ran a moderated test with 3 ops team members on usertesting.com.
Key findings:
✅ 100% of users found the interface more intuitive and aligned with their workflow.
✅ Search confirmed as a needed feature — prioritized and added to the design.
“The preview shelf is definitely an upgrade from what we were previously using.”
Chris, Catalog Analyst
Caption
What's the Impact?
Accelerating workflows while meeting strategic goals
🛠️
Less engineering involvement
This frees up dev resources and allows them to only support when needed.
“Preview 2.0 helped us launch faster, reduced engineering dependency, and empowered more Ops team members to improve product data quality.”
Walt, Product Manager
Reflection
Collaboration and iteration
This multi-year project taught me how to design within a large, evolving scope and collaborate across teams to modernize a legacy tool. I supported developers post-handoff, adapting designs as new use cases surfaced. It was rewarding to see the tool evolve through multiple iterations and contribute to a high-impact system.