Rebuilt Supplier Performance
When a platform used by 100,000 suppliers needs to move homes, the easy answer is to move it as-is. We didn't take the easy answer.
Enterprise
Data Visualization
Role
Sr. UX Designer
Timeline
Feb 2025 – Jan 2026
team
Merchants, Suppliers, PM, Engineers, Stakeholders, Me
platform
Web, Mobile
The Real Problem
SQEP had been running for eight years. It was Walmart's Supplier Quality and Excellence Program — the platform that 100,000+ suppliers and merchants used to track performance, compliance, and program health. It worked, more or less. But it had been built incrementally, screen by screen, over nearly a decade, and it lived completely outside the new Supplier One platform that Walmart was moving everything onto.
The brief was straightforward: migrate it over. Match the new design system. Ship it.
What nobody had done yet was actually sit with the people using it.
When I started talking to merchants, it became clear quickly that "migrate it as-is" would mean migrating eight years of usability debt into a brand new platform. Suppliers couldn't tell what was driving their scores up or down. Merchants were clicking through multiple places just to get a picture of one supplier's situation. Data that should have told a story was sitting in tables that told nothing.
The feedback from support tickets and analytics made it concrete:
'I can see my score but I have no idea what's causing it.'
'I have to visit five different screens to understand one supplier's situation.'
'By the time I find what I need, I've already lost the thread.'
The platform wasn't broken. It was just never designed — it had grown. And moving it into a new home without fixing that would have been a missed opportunity that wouldn't come around again for another eight years.
Finding the Fix
I started where the frustration was loudest — support tickets, usage analytics, and existing data about how people actually moved through the platform. The patterns weren't subtle.
Users weren't failing to find features. They were failing to find meaning. A score without context is just a number. A trend without a visual shape is just a table. The platform had years of rich performance data and no way to make any of it readable at a glance.
Three things stood out:
Suppliers needed causation, not just scores. A SQEP score of 72 tells you nothing actionable. A score of 72 because on-time delivery dropped in the Southwest region tells you exactly what to fix. The old platform gave you the number. Never the reason.
Merchants needed to triage, not read. They weren't reviewing one supplier at a time. They were scanning portfolios — looking for who needed intervention before problems grew. The existing layout forced them to go looking. They needed something that surfaced the answer before they had to ask.
Eight years of trend data was being shown like a spreadsheet. The history was there. Nobody could see it. Multi-year performance patterns presented in rows and columns don't reveal anything — they just confirm that the data exists.
I took these findings back to stakeholders alongside the designs. What I had in mind wasn't just a migration anymore — it was a migration with a usability overhaul woven in. The timeline was tight. The platform was huge. And the work would have to happen in three phases while SQEP stayed live in production the entire time.
Stakeholders looked at the designs and backed it. The work was turning out better than anyone had expected for users — and that made the ambition worth the risk.
What Actually Happened
Because SQEP was live in production throughout, every phase had to ship cleanly before the next one could begin. There was no staging environment where we could quietly get things wrong. The platform had real users, real scores, and real business consequences every day it was running.
That shaped everything about how the work was sequenced.
Phase one covered the foundation — core navigation, information architecture, and getting the platform living inside Supplier One with the new design system applied. New components were built from scratch where the design system didn't yet have what SQEP needed. Discovery happened alongside delivery — I was learning what the next phase needed while shipping the current one.
Phase two was where the visualization work took shape. The treemap went through the most iteration — my first version grouped suppliers by score range, which was logical but wrong. Merchants kept trying to read it geographically. I restructured it around region first, with score as a secondary layer. Immediately more useful. The trend graphs had a similar story: early versions showed everything, which meant users saw nothing. Stripped back to three default views with the option to drill deeper. The geo view came last and landed best — supplier performance mapped across the US, color showing health, density showing volume. Merchants could spot regional patterns in under ten seconds.
Phase three brought it together — the full experience, both supplier and merchant views, the new data visualizations, and the refined flows that came out of ongoing feedback as each previous phase shipped.
Throughout all of it, the design didn't get compromised to meet the timeline. It got tighter.
What Changed
The SQEP redesign moved the platform's ASUS score from 65 to 80+. In a post-launch survey, 90% of suppliers said the redesign was genuinely helpful — a number that means something when the platform has 100,000 active users.
Stakeholders who had approved a migration ended up with a platform that performed significantly better than the one they'd been running for eight years. The timeline held. Production stayed live throughout. Nothing broke.
The feedback that stayed with me was simpler than any metric:
'I finally understand why my score changed.'
'I used to spend twenty minutes figuring out where to start. Now I can see it immediately.'
'This actually looks like someone thought about how we work.'
The next step for SQEP is to take this further — connecting the performance data to direct, actionable guidance. The vision is an Action Center that surfaces specific recommendations for suppliers on how to improve their scores, and gives merchants a structured way to review supplier performance for assortment planning. The foundation is there. The data is readable now. What comes next is making it actionable.
What I Had to Work With
A live platform that couldn't go down. Every phase had to ship correctly because real suppliers and merchants were using SQEP every day. There was no room for "we'll fix it in the next release" — each release had to stand on its own.
A platform too large to redesign all at once. SQEP had accumulated eight years of features, flows, and edge cases. Trying to overhaul everything in one go would have broken the timeline and probably the platform. Phasing the work meant making constant decisions about what to fix now, what to fix next, and what to carry forward temporarily — and owning those calls.
Two user groups with genuinely different needs. Suppliers wanted to understand their own performance deeply. Merchants needed to triage across large portfolios fast. One platform, two mental models, zero room to design for only one of them. Running workshops with each group separately was the only way to keep both needs clear without letting one compromise the other.
A tight timeline with on-the-go discovery. There was no phase of "research, then design, then build." Discovery happened while delivery was already moving. That meant staying close to feedback at every stage and being willing to adjust the next phase based on what the current one taught.
What I'd Do Differently
I'd push for a short discovery sprint before phase one began. The on-the-go research worked, but some of the visualization decisions in phase two would have come faster if I'd done more structured merchant interviews before any design started. The treemap restructure, in particular, could have been avoided with one more conversation upfront.
I'd also create a more visible record of decisions made under constraint. When you're shipping live in phases with a tight timeline, decisions happen fast — and some of them are "good enough for now" rather than "right." Documenting those explicitly would have made the handoff between phases cleaner and given the team a clearer picture of what was intentional versus what was deferred.
What I Learned
Migration is a design opportunity, not just a technical task. The brief was to move SQEP into Supplier One. The opportunity was to fix eight years of accumulated usability debt at the same time. Recognizing that difference — and making the case for it with designs rather than arguments — is what turned a routine project into something stakeholders were proud of.
Data without shape is just noise. The information was always in SQEP. What was missing was a way for a human eye to read it quickly and trust what it was seeing. The visualization work wasn't decoration — it was the difference between data that existed and data that got used.
Tight timelines and good design aren't opposites. They require different decisions. When the timeline is fixed, the question isn't "how do we do less?" — it's "how do we sequence this so nothing important gets cut?" Phasing the work correctly meant the experience didn't get compromised. It just got shipped in the right order.