Usability Testing Affinity Mapping Insight Synthesis Health Tech Enterprise Wellness

Enhancing Wellbeing
for Essential Workers

ESNTL Wellness is an enterprise wellness platform developed by Concentrix in collaboration with Apple. As Lead Researcher, I conducted a comprehensive usability study with my team to evaluate how well users can understand, find value in, and take action based on the health insights provided by the app.

Company

Concentrix × Apple Inc.

My Role

Lead UX Researcher

Duration

3 months

Team

1 PM · 2 UX Designers · me

67%

of essential workers couldn't locate their own health data during testing

Overview

What is ESNTL Wellness?

An enterprise wellness platform that uses passively and actively collected data to provide personalized insights and actionable recommendations for workplace wellbeing. Designed for essential workers who carry high stress loads and have the least time to manage their own health.

$ESNTL Wellness across Apple Watch, iPhone and admin dashboard$

💻

App Screenshot 2

Replace with your image URL

The Brief

Concentrix engaged our team to conduct a formative usability evaluation of the ESNTL Wellness app, identifying friction points in the insights experience and delivering evidence-based recommendations to inform the product roadmap.

Platform Goals

Improve user awareness of how daily decisions affect their wellness and guide them to the incremental changes they can make to support their health.

The Problem

The app was built to empower. It was doing the opposite.

In our initial cognitive walkthrough, we assessed the ESNTL app across four key flows: onboarding, daily check-in, reviewing insights, and seeking support. One pattern emerged immediately. The insights feature, which was the core value of the app, was buried and disconnected from any actionable guidance.

"Users could see their health data. But they couldn't make sense of it, couldn't navigate to it, and in many cases felt no motivation to act on it."

Key Numbers from Testing

33%

could find their mood quality data

50%

could navigate to insights page

Task completion across 6 essential workers: 3 nurses, 2 firefighters, 1 law enforcement officer.

Research Timeline

3 months from brief to delivery.

Weeks 1–2

Problem Definition

Cognitive walkthrough · FigJam mapping · Stakeholder alignment

Weeks 3–4

Research Design

Screener · Task design · Recruitment · Pilot test

Week 5

Data Collection

6 sessions · Think-aloud · Zoom · Transcription

Week 6

Synthesis

Affinity mapping · Pattern ID · Prioritization

Weeks 7–12

Delivery

Report · 6 recommendations · Stakeholder presentation

My Role

Sole researcher on a cross-functional team of four.

I owned the research strategy end-to-end, defining the research questions, designing the study, recruiting participants, and leading the synthesis while moderating 5 of 6 usability sessions, running think-aloud protocols, probing for behavioral insight, and adapting in real time when participants went off-script.

Participants

3 nurses, 2 firefighters, 1 law enforcement officer.

Methods

Recruitment Survey

Screener questionnaire · profession, familiarity, stress levels

Snowball Sampling

Network-based referrals via Concentrix team

Cognitive Walkthrough

4 app flows assessed independently before testing

Moderated Usability Testing

6 sessions · remote via Zoom · think-aloud · I led 5

In-depth Interviews

Discovery questions + in-session probing

Affinity Mapping

FigJam · participant × task · theme clusters

Post-Test Survey

8 Likert questions · n=6 · satisfaction + intent

Observation

Dedicated notetaker · verbal + non-verbal cues

Key Insights

Four things we learned that Concentrix needed to hear.

Each finding is backed by behavioral observation, completion data, and participant quotes, weighted by frequency and severity.

Users couldn't navigate to their own health data.

The mood quality section was nested below the check-in module and the insights summary, meaning users had to scroll past two other sections to reach the feature the app was built around. If essential workers under high stress can't find the data in three taps, they won't open the app a second time.

"I'll click Summary maybe? No, but that's where I was before."

, P05, during Task 2

Users saw their data. They had no idea what it meant.

Every participant encountered the "baseline" metric. Not one correctly understood how it was calculated. Beyond baseline, users were confused by the 1–7 scale used for most metrics. "Why is it out of 7? What does 4.9/7 mean? Is that good?" The scales were internally consistent but externally meaningless without context.

"Whose baseline is this? Health Institute or all users? I don't know who you're comparing me to."

, P02, Task 2

The heart rate notification created anxiety, not empowerment.

The app surfaced an elevated heart rate notification during the check-in, a feature designed to prompt reflection and proactive health management. Across 6 participants, reactions ranged from confusion to indifference to genuine alarm. A nurse, someone trained to interpret medical data, still found the notification ambiguous without plain-English context.

"I see that my heart rate is elevated, but what does that mean? Is it something I should be concerned about?"

, P01, Task 1 (after HR notification)

A one-size-fits-all experience didn't fit anyone.

Every participant mentioned personalization without being prompted. A nurse's wellness goals are different from a firefighter's, which are different from a law enforcement officer's. But the app asked everyone the same questions, on the same scale, with the same generic tips.

"In this day and age… it's all about personalization."

, P01, Task 1

The Process

How we got from an open brief to six prioritized recommendations.

Step 01

Current state assessment, before a single user

Each team member explored the app independently and mapped interaction flows in FigJam. One pattern appeared in everyone's notes: the insights flow had the most friction and the widest gap between intent and experience.

Step 02

Stakeholder alignment, defining the brief we weren't given

Concentrix gave us an open direction. Rather than running a broad evaluation, we aligned on a specific hypothesis: that users couldn't understand or act on their health insights. Concentrix agreed. This became our north star for every subsequent decision.

Step 03

Study design and pilot, building in rigor

Three scenario-based tasks, a facilitation script for consistency, and a pilot test before any real participant saw the study. The pilot caught two protocol gaps we fixed before going live.

Step 04

Six sessions with essential workers

We tested with 3 nurses, 2 firefighters, and 1 law enforcement officer, the exact user group ESNTL was built for. I moderated 5 of the 6 sessions, using the think-aloud protocol to capture both behavioral patterns and the reasoning behind them. Sessions were recorded via Zoom and transcribed using TurboScribe.

Step 05

Synthesis, from observations to patterns

After all six sessions, we used FigJam to build an affinity map, organizing observations by participant and task, then clustering by theme. We prioritized findings using two criteria: how often it appeared across participants, and how severely it blocked the core use case. The result was a clear hierarchy of 6 issues, each with a recommended intervention.

Affinity map — HR Notification findings clustered by theme

Interaction flow map — Check-In flow cognitive walkthrough

Research Artifact

Interaction Flow Map, FigJam

Cognitive walkthrough outputs: full interaction architecture mapped across Onboarding, Check-In, Insights Review, and Support-Seeking flows.

[Upload interaction flow export here, see Asset 03 in Step 3C guide]

A Research Decision

"An open brief is a gift, if you treat it as a design challenge."

Concentrix didn't hand us a problem to solve. They handed us permission to find one. The cognitive walkthrough before any formal testing was our most important research act, it let us define the question that would make the study matter. Rather than running a general evaluation, we shaped a focused brief that produced findings Concentrix could actually act on.

Design Recommendations

Six recommendations. Prioritized by impact and implementability.

Each recommendation maps directly to a finding. We used a Now / Next / Later framework to give Concentrix a clear implementation roadmap, not just a list of problems to fix.

🗂️

Restructure information hierarchy

Surface health data in primary navigation. Mood Quality and Work-Life Balance belong in the top level, not buried below check-in.

💬

Explain metrics at point of use

Every metric needs a plain-English tooltip. "Your baseline is calculated from your last 30 check-ins" is clearer than a bare number.

⚡

Add actionable guidance to every data point

Health data without context causes anxiety. Every metric needs a "what this means" and a "what you can do."

🎯

Introduce personalization

Let users prioritize the metrics they care about and set goals by profession. A nurse's stress data is not a firefighter's stress data.

📊

Redesign graphs for comprehension

Interactive overlays, tap-for-context, zoom by time period. Clarity over accuracy. A graph no one can read in 5 seconds adds cognitive load, not insight.

👀

Improve infographic readability

Review contrast ratios, color choices, and annotation density. Design for first-time readers, not power users.

Impact & Outcomes

The first formal usability data this feature had ever seen.

Before our study, Concentrix had no structured user data on the ESNTL insights experience. The feature had been built and shipped without formal usability testing. Our research generated the first evidence-based baseline for this critical part of the product.

Our research also produced a strategic framework, the Now / Next / Later prioritization, that gave the product team a clear sequence for improvement rather than a flat list of problems.

Post-Test Survey Results (n=6)

Overall satisfaction with the app

3.58 / 5

Easy to use

3.79 / 5

Can effectively track wellness

3.50 / 5

Would use 3+ days per week ⚠

2.75 / 5, lowest score

Resources and tips useful

3.33 / 5

Insights data useful ⚠

3.00 / 5

Data feels private and secure

3.17 / 5

The low repeat-use score (2.75/5) is the most significant signal: comprehension failures don't just frustrate users in the moment, they kill the habit.

What I Learned

Three things I'd do differently, and one thing I'd always do again.

🎯

Define the question before you answer it

The cognitive walkthrough before any formal testing was our highest-leverage act. An open brief is permission to define a better question. Better questions produce more useful findings.

⚡

Health data without context can cause harm

Good intent doesn't override cognitive response. Alarming numbers without guidance trigger anxiety before reflection. Emotional response to health data is a first-class design variable, not an afterthought.

📋

What I'd change about the recruitment

A nurse and a firefighter interpret "elevated heart rate" differently, and that difference was the data. Contingency recruitment planning upfront means you're never one refusal away from a pivot.

Enhancing Wellbeingfor Essential Workers