I find out what people actually do, and help teams build for all of them.

Context

Verizon was consolidating personalized offers into a single My Offers experience across the app and website. Engagement was falling, and new multi-offer tiles were about to roll out.

Approach

I paired behavioral telemetry with in-depth interviews, then synthesized findings across several offer studies, including a color study and tile usability testing run by my team. The data showed where people dropped off. The interviews explained why: competing visual hierarchies, crowded tiles and brand color combinations that added friction.

What changed

The synthesis became a set of experience principles for offer strategy. Design used it to rework tile layout, color and placement, and I made the case to executives for end-to-end usability over isolated feature scores.

Context

Teams were building line and device pickers into the Verizon Assistant chat interface with no shared standard. Carousels, radio lists and cards all coexisted.

Approach

I ran unmoderated interaction tests with 193 participants, comparing layout orientation, selection controls and four levels of information density, from a bare phone number up to name, number, device image, plan and price.

What changed

Vertical layouts beat carousels on speed, and richer detail beat minimal cards on confidence. The findings became two standards: vertical orientation on mobile across plan management flows, and a device image on every troubleshooting line selection.

Context

Verizon's Next Gen Digital mobile sales flow was being reskinned under the new Monarch brand. Visual changes like this often break things that only assistive technology users notice.

Approach

I ran a study with 27 participants: 16 screen reader users, 5 switch and voice control users, and 6 low-vision magnification users. They worked through real purchase tasks while I logged where and why they got stuck.

Findings

• Sign-in loops that sent authenticated customers back to the login screen
• A focus trap in the Estimated Next Bill modal
• Unlabeled trade-in and porting options that a screen reader announced as "Object Object"
• Layout density that made magnified screens hard to follow

What changed

I turned the findings into a prioritized remediation list for engineering and design, aimed at the checkout steps where high-intent customers were dropping out.

Context

Product teams were detaching from central VDS components and building their own. Separately, the research team's archive of more than a thousand studies was hard to search, so findings got repeated instead of reused.

Approach

I configured an AI design audit tool to find the 10 components with the highest detachment rates, then interviewed designers and engineers about why. In parallel I designed a 7-facet taxonomy and content model for the research archive, consolidated three glossaries into one acronym list, and pressure-tested it against real datasets like MaxDiff surveys and monadic tests.

What changed

The adoption study named the gaps (a web-first bias, missing complex components, poor discoverability, fragile version upgrades) and led to recommendations including leadership acceptance criteria, sprint time for visual quality defects, and a contribution path for teams. The taxonomy shipped as production files for semantic search in the team's Stravito repository.

Context

Creatability is an open-source set of experiments that use machine learning to make creative tools more accessible, built by Google Creative Lab with NYU's Ability Project.

Approach

I ran user research with people with diverse abilities and tested new interaction modes as they were built, working across research, design and engineering.

What changed

The research shaped how the tools handled input and feedback. The project went on to win the 2019 Cannes Lions Design Grand Prix.