Switchfly

Switchfly is a highly customizable travel commerce platform. End users book flights, hotels, cars, and activities using loyalty points, cash, or both, but who that user is varies enormously. The platform powers travel for airlines like JetBlue, card brands like American Express, and other companies built on loyalty points.

Every client runs any number of markets, US, Canada, Mexico, Australia, Japan and more, and each market carries its own configuration: which features are on, how they behave, how they look, the language, the content, the copy, and custom overrides. The same product can feel like a completely different app depending on the client, the cobrand, and the device.

Beneath all that customization sat a harder problem. Every product is sourced from external connectors like Expedia, and each connector structures and names its data differently. Presenting an aggregated, trustworthy set of results meant first understanding those variations cold, then designing so they read as consistently as possible no matter where they came from.

So the real design challenge was rarely “make this screen prettier.” It was: how do you create one coherent, scannable experience across inconsistent source data, dozens of market configurations, and clients who can override almost anything?

Real constraints shaped every decision, and naming them early kept the work honest about trade-offs.

I supported six engineering teams as an individual contributor while taking on manager and director-level duties, including managing another designer. Work was planned in SAFe Product Increments, five PIs a year, five sprints each, ten days a sprint, so I partnered with product owners one to two PIs ahead, writing stories and epics in JIRA, with a major release every PI plus minor releases throughout.

To keep design coherent across that surface area, I built process, not just screens. I started twice-weekly “design forum” hours, 60-90 minutes open to Engineering, Product, QA, and Sales, to collaborate and surface insight across the company. And I introduced a new intake process for design requests, with a design story template product owners could actually use.

Platform work is relentless coordination. Beyond the six engineering teams, I met weekly with internal stakeholders and clients, primarily American Express, to walk through upcoming work, work in progress, and customer feedback on what we’d just shipped.

Managing another designer and running the design forum meant I was as much a connector between disciplines as a maker. The request process and story template existed precisely so collaboration could scale without everything routing through me.

One representative slice: flight search results asked users to interpret fare flexibility, baggage inclusion, and change policies with almost no visual hierarchy. I introduced a clearer indicator framework, concise labels, visual tags, and contextual tooltips that surface the critical details right at the decision point, inside the search result cards.

Because each client can edit the language, I wrote and designed the defaults to reduce ambiguity around fare rules and fees, provide immediate clarification without forcing deeper navigation, and stay visually consistent across the card. Better hierarchy and contextual disclosure lowered the cognitive load of comparing flights and helped users move to checkout with more confidence.

The work ranged widely in size and scope, the goal, where possible, was always to make it available to every client. Some were brand-new features; some were redesigns of existing pages; some were small, surgical fixes for a single client.

At platform scale, the highest-leverage design work is often invisible: the request process, the story template, the forum, the defaults robust enough to survive a hundred client overrides. Build those and quality compounds; skip them and you’re firefighting screens forever.

I’d also protect the weekly client time even harder. Standing in front of Amex every week is what kept the abstraction, all those configs and connectors, anchored to a real person trying to book a real trip.