From Manual Workarounds to Data-Driven Care
My first product research project — leading a multi-cycle program from a broad, ambiguous problem space to a focused, validated product direction.
· Accurx · Healthtech · 2023–2024
“ We’ve been recording this data using clunky Excel sheets since March of 2020. So immediately I can tell you — this is the data I’ve been asking for.”
the problem.
Primary care practices lacked critical operational visibility into their own patient demand data. Reporting mechanisms were manual, fragmented, and fundamentally reactive—occurring at the exact moment the UK government initiated intense public scrutiny on practice performance metrics. The lack of standardized data capabilities left healthcare providers defenseless against regulatory pressure.
my role.
As the sole User Research Lead across three six-week product cycles, I maintained end-to-end ownership of the space: policy analysis, survey design, moderated qualitative interviews, and usability testing. Recognizing a fundamental data gap, I independently engineered the product team's first quantitative research infrastructure from scratch to enable data-backed decision-making.
the outcome.
Successfully validated demand management as the core product use case, driving the launch and rapid iteration of a new Reporting MVP. Additionally, I architected a user-persona framework adopted across the organization and fundamentally reframed the product's value proposition from a basic "reporting tool" to a high-impact decision support engine.
my approach.
I initiated the program by analyzing the broader policy landscape rather than internal product features. While the government had begun publishing national appointment data, the definition of an "appointment" varied wildly across practices—a structural fragmentation that led the NHS to classify its own dataset as “experimental." This analysis reframed the entire initiative: practices were not underperforming; they simply lacked the data architecture to validate their actual labor. This was a measurement problem before it was a product problem.
To validate this friction at scale, I executed a mixed-methods study comprising a survey of 350 respondents across five care settings and subsequent moderated interviews. Rather than relying on opt-in volunteers, I targeted high-volume power users by leveraging MixPanel behavioral analytics. The research uncovered a fundamental mental-model gap: hindered by a lack of reference points, users struggled to articulate advanced reporting solutions. Shifting strategy, I introduced live, data-driven interactive prototypes to anchor their feedback.
The pivotal insight was that users did not actually want static reports—they wanted to forecast demand patterns, optimize practice staffing, and determine triage closure thresholds. Reporting was merely the mechanism; high-stakes operational decision-making was the ultimate goal.
“We calculate averages ourselves every month to work out when to close triage — it takes hours we don’t have.”
the impact.
The research successfully defined the strategic roadmap for a new enterprise reporting capability, now deployed across practices managing tens of thousands of monthly patient requests. The resulting user-types framework was unified across all cross-functional product teams.
Furthermore, a self-initiated behavioral study I conducted revealed that 73% of high-volume users were completely unaware the existing dashboards even existed. This critical discovery successfully redirected engineering capital away from an expensive, unnecessary redesign that would have solved the wrong problem. When the optimized dashboards reached production, user adoption was immediate—delivering the automated infrastructure clinicians had spent years rebuilding by hand.
The comprehensive write-up — full methodology, artifacts, and all three research cycles — lives in my password-protected portfolio. Reach out for access! 📧hello@smresearchdesign.com
To access the full project, reach out for access to my comprehensive Notion portfolio.
📧hello@smresearchdesign.com