Case study · The network's primary B2B platform

Four years inside the network's front door

Mastercard Connect is the primary touchpoint between the network and every customer it has — issuers, acquirers, and processors. I spent roughly four years embedded in it, running the research behind a platform-wide modernization.

Role
Staff UX Researcher (platform)
Context
Mastercard Connect · B2B portal
Tenure
~4 years embedded
Focus
Legacy → next-gen migration
MASTERCARD CONNECT STORE · MY ITEMS Application Report Product Service My Company Manager APPLICATION Manage My Company APPLICATION ? Operational Reports REPORT Pricing & Billing PRODUCT Issuer Loss File REPORT
The Connect "Store" — applications, reports, and products, color-coded by type (illustrative reconstruction). The two near-identical names are real.
The platform

One portal, dozens of business-critical apps

Connect is where the network's customers actually do business with Mastercard — a self-service portal hosting a sprawling catalog of applications across issuing, acquiring, disputes, tokenization, data, and billing. A representative slice of what lives there:

My Company Manager Mastercom (disputes) ISSM (enrollment) MDES Manager Account Level Management Data Integrity Online Operational Reports Pricing & Billing Resource Center Mastercard Market Trends Location Administration Tool

Application names shown are publicly documented in Mastercard's own Connect materials. Specific client data and internal detail are kept confidential.

The taxonomy

Even the categories were a usability problem

Connect doesn't just host apps — it sorts everything into item types: applications, reports, and products, each behaving differently. The platform signaled those types with a color-coded card system — three or four colors, each carrying a distinct meaning that very few users actually understood. I worked on that taxonomy directly: how items get categorized, labeled, and visually coded so people can tell what they're looking at before they click.

Taxonomy is invisible until it's wrong. When a system's own categories don't match how customers think, every search and every menu quietly costs time — multiplied across an entire network's customer base.
The mandate

Move a generation of apps from legacy to next — without breaking the customers who depend on them

The platform was migrating its applications from a legacy generation to a new one — major design overhauls, and sometimes functional ones too. The catch: these are tools real businesses run their operations on every day, so nothing could regress. The work was sequenced from the most critical applications down, and it ran for years.

My job was the research and design guidance underneath that migration: understanding how each app was actually used, where it failed people, and what "better" had to mean before a single screen was rebuilt.

The clearest example

Two apps. Nearly the same name. Years of quiet friction.

For a long time, customers complained about Connect and no one could quite say why. The pattern only showed up once you watched people use it: the platform had two separate applications with names so close that nobody could remember which did what.

"My Company Manager"
one application
vs.
"Manage My Company"
a different application

Everyone felt the friction; nobody had named the actual problem. It wasn't a visual-design issue or a missing feature — it was an information-architecture and naming problem hiding in plain sight, the kind that only surfaces when you sit with real users instead of reading the org chart.

This is the whole job in one anecdote: finding the load-bearing confusion no one had put into words — then giving the team the evidence to fix it for good.
The app I designed

Customer Parameter Management — from a spreadsheet no one could touch to an app anyone could use

One application in this migration I didn't just research — I designed it. Customer Parameter Management began as a legacy two-axis spreadsheet, dense with the logic that codes how card programs behave. I rebuilt it as a single-axis web application that lost none of that functionality — working from Connect's design system and style guides, and from my own usability research, with the few true domain experts almost entirely unavailable to help.

Two months on paper. It took me a year — and it's the deepest financial-services knowledge I've ever built.

The internals stay confidential — this is core network configuration, and some of it still should. It's here because it's the clearest proof of the range: research and design, on infrastructure the business couldn't afford to get wrong.
What the work looked like

Full-stack research, at platform scale

Understand
  • End-user persona development across issuers, acquirers, and processors
  • Ethnographic and contextual research into how each app was actually used
  • Customer journey mapping across the apps people moved between
Shape
  • Information-architecture work to untangle overlapping, confusingly-named tools — and the platform's own taxonomy of item types
  • Proof-of-concept and usability testing on redesigns before build
  • Design guidance to the teams rebuilding each application — and a rework of the platform's Help pages so customers could self-serve
Land
  • Synthesis and executive readouts to align stakeholders on what to change and why
  • Iteration across years and releases as the migration moved app by app
Why this one matters

Spend four years inside one platform and you stop seeing screens and start seeing the system — the place where a small naming decision quietly taxes thousands of customers every day. Finding those is the work I'm best at.

Methods on this engagement
Ethnographic & contextual inquiry Persona development Journey mapping Information architecture Card sorting & taxonomy design Nomenclature / naming research Proof-of-concept & usability testing Interaction & UI design Design-system–driven design Executive synthesis & readouts

A slice of a deeper toolkit — 70+ named research, product, and facilitation methods, drawn from a working library of 175+ structured activities. The right ones get pulled for the problem in the room.