Selected work
Evidence that moves organizations forward
Onboarding was losing 87% of users before the product had a chance
The team had written off an 87% exit rate as unavoidable compliance friction. It wasn't — and the real reasons had nothing to do with compliance.
The stylist tool was built for stylists. It was missing the two things they actually needed.
17% adoption on a tool built specifically for the people using it every day. Leadership was already weighing a full rebuild. Before they did, someone finally asked the stylists what was actually missing.
Two weeks post-launch. Users were finding it. Zero bookings.
Behavioral data found what no one was looking for — a missing return path. One nav change. Bookings within a week.
Users were falling into a fashion rabbit hole. The app was making it worse.
Users came looking for style inspiration and left overwhelmed. A diary study revealed the tension between browsing and deciding — and what it actually takes to bridge the two.
A brand new loyalty discount operation at checkout. Staff had never seen it before — and it showed.
Loyalty discounts were hitting checkout for the first time. Testing with staff on live transactions — before a single customer touched it — found exactly what would have broken it on day one.
Investors were logging into their dashboard and leaving more confused than before.
Users were logging in, checking their numbers, and leaving more uncertain than before. The same frustration surfaced in annual surveys year after year — until someone asked why.
The listing said "breakfast included." It was a 30-course seasonal Kaiseki ceremony.
A 30-course seasonal Kaiseki ceremony listed as "breakfast included." When travelers from three countries explained what they saw, the gap between experience and expectation turned out to be far bigger than a translation problem.
The 14-day early access perk was real. Nobody at the store knew how to activate it.
Members had 14-day exclusive access to their favourite brands. Most never used it. Diary studies and in-store tag-alongs revealed why — and it had nothing to do with the perk itself.
The benchmarking tool was solving half the problem. Research surfaced the other half.
A global real estate org was solving the right problem — but only half of it. Discovery interviews across four business functions surfaced a foundational need nobody had put on the roadmap.
The resort team was shipping features for "families" and "couples" — without ever testing with either.
Fast-moving teams don't always have time for full research cycles. A two-hour workshop was enough to show how far the team's assumptions were from what families and couples actually wanted.
100 research questions. One researcher. No system to decide what mattered most.
Nearly 100 open research questions and no shared way to prioritize them. What started as a personal problem turned out to be a structural one — and the fix was simpler than anyone expected.
The platform had a roadmap. A JTBD study across five roles revealed a much bigger picture.
17 users. Five roles. 200+ individual jobs mapped. What the team thought was a validated roadmap turned out to be missing something none of the five roles had ever said out loud.
