Example Projects

The most popular tagline was the one we had to kill

A company launching a new product feature had three tagline options and needed to pick one fast. The early frontrunner had strong appeal, but when I ran a monadic survey with 1000+ participants testing emotional response, clarity, and brand fit, a problem surfaced.

Many people misread what that tagline was promising. They liked it because they assumed the product could do something it couldn't.

Once participants saw what the feature actually did, preference flipped. The original winner became the least preferred option.

Shipping that tagline would have created false expectations from day one, disappointed users at best, legal exposure at worst. The team moved forward with a message that was less flashy but actually true.

How a travel tech company stopped arguing about the roadmap

Feature prioritisation at this company came down to two things: build effort and business value. Customer opinion didn't have a seat at the table, not because no one cared, but because there was no structured way to bring it in.

I built one. Working with PMs and designers across teams, I gathered their top feature candidates, turned them into plain-language statements, and ran targeted surveys with active users to measure what they actually valued. The output was a simple three-axis framework: feasibility, business impact, and customer pull. All could be plotted visually so tradeoffs were obvious at a glance.

Teams could now point to evidence when killing a low-value feature, not just instinct. Roadmap discussions got faster and less political. And the process was light enough to repeat without adding overhead.

How much personalisation is too much? Users told us exactly.

A mental health app wanted to offer tailored recommendations — exercises and prompts matched to each user. The tension: more personalisation requires more data, and the team worried that would erode trust.

I designed a concept testing study around different personalisation scenarios, varying how much data each required and how much control users had over it. Participants reacted to prototypes and we measured emotional response, perceived helpfulness, and intent to use.

The finding was unambiguous. Users were open to personalisation — but only when they controlled it. Opt-in features with clear data choices felt helpful. Features that collected data quietly and served automatic suggestions felt intrusive, even when they promised better results.

The team launched with an opt-in model. Ethical design and good personalisation turned out not to be in conflict.

This work even got published in the most prestigeous conference in Human-Computer Intercation: CHI

Bundling more services made them feel worth less

A global online travel agency assumed that bundling more ancillary services (luggage, seat selection, insurance) would feel like better value and drive more sales. It wasn't working, and they didn't know why.

I ran a conjoint analysis with 800+ long-haul and family travellers, simulating real booking decisions across different bundle configurations and price points.

The problem was a mismatch between desire and willingness to pay. Travellers wanted the high-value extras like checked bags, cancellation protection but as soon as those went into a bundle, the total price crossed a threshold and the whole thing felt expensive. Meanwhile, bundles of smaller, cheaper services (seat selection, basic support) felt like a genuine deal, even though individually those items weren't priorities.

The company moved away from large all-in packages toward lighter, flexible bundles built around perceived value. Sometimes less really does sell more.