Case studies

Examples and case studies pulled from past work

Client names are withheld where required. Each write-up states the decision on the table, how the research was built into the workflow around it, and what changed afterwards. Happy to walk through any of them on a call.

Segmentation and analytics · A U.S. shopper-research agency · 2022-2023

Rebuilding a shopper segmentation for repeated use

The director running a 33,000-respondent shopper study left the agency mid-project with no documentation, and no one internally could pick it up. Working as an outside analyst with a four-week scope, I reverse-engineered the design and built a two-layer segmentation, six shopper types by four trip types, with short typing tools developed alongside the model so any future survey could reproduce it. Weeks after delivery the framework sized a seven-million-shopper growth gap for a packaged-food client. It then grew into a 20-chain retailer extension, custom studies across eight-plus categories, and a syndicated offering the agency was still selling a year later.

Read the write-up →

33,000respondents in the base study

7Mshopper growth opportunity sized

20retail chains sized with the same model

Concept and product testing · An Ontario beer brand · 2025

Choosing between two beer concepts when the scores tie

Two lager concepts, one built on strength and one on dependable quality at a value price, came out statistically tied on appeal and purchase intent among 600 Ontario beer buyers. A virtual shelf test showed the quality concept lifted total shelf revenue by 11 percent against 5 percent for the strength concept, with nearly all of its volume new to the retailer and minimal cannibalization of the brand’s own products. Claims analysis showed the strength concept’s main weakness was a formulation ceiling, while the quality concept’s gaps were a communication job. The recommendation came from understanding the evidence around the headline score and the practical levers the marketing team had to pull.

Read the write-up →

600beer buyers, three-cell shelf experiment

+11%total shelf revenue per visitor, quality concept

3lines of evidence behind one recommendation

Product and design research · A major Greater Toronto condominium developer · 2023

Which condo layouts residents actually want, and why

A developer and its interior-design partner needed evidence before committing new suite layouts and upgrade options to a design brief. My practice designed, fielded and analyzed a survey of 575 prospective residents in the Greater Toronto and Hamilton Area, combining MaxDiff preference share, latent class segmentation, TURF analysis and click-on-plan diagnostics that let people say exactly where a floor plan worked or failed. Residents put storage first, then natural light and separate spaces for separate activities. Attractive upgrades that cost counter or closet space lost support once the trade-off was visible. The wide one-plus-den plan won, seven preference segments showed why upgrade packages should be chosen project by project, and the study surfaced a net-new furnished-package revenue opportunity.

Read the write-up →

575prospective residents, investors screened out

7preference segments guiding upgrade packages

97%of residents reached by three appliances

Program evaluation · Common Compass, an Ontario registered charity · 2026

Connecting program objectives to evidence people can use

Common Compass runs mental health and belonging programs for students, parents and educators across Ontario schools. Its surveys had grown complex across 16 program modules with no formal data model behind them, so insight was hard to reach, facilitators sometimes saw results at the same moment as the schools they were debriefing, and analysis depended on R skills the organization did not have. As a volunteer research advisor through Data For Good, I audited program objectives and learning outcomes, mapped them to constructs and survey items, built a common core with tailored modules, designed the analytical data model, and automated results emails and debrief slides in the Google Workspace tools staff already used. Every response now carries its context, and routine reporting runs without an analyst.

Read the write-up →

16program modules on one measurement framework

3audiences with shared and tailored measures

0lines of code staff write to run reporting