Heuristic Evaluation & Usability Study of the Flo App

The study combined complementary methods — planning, a quantitative survey, qualitative interviews, and a full heuristic evaluation — to triangulate self-reported satisfaction against actual usability.

• ✦Planning — CSD Matrix
• ✦Quantitative Survey (42 respondents)
• ✦Qualitative Interviews (4 participants)
• ✦Heuristic Evaluation — Nielsen's 10 Heuristics (38 usability problems identified, 4 catastrophic)

Before defining the research plan, I mapped what was already known, assumed, or unclear using a CSD Matrix (Certainties, Suppositions, Doubts) to separate fact from assumption before choosing a method. The doubts that surfaced directly shaped the four research questions guiding the study, and the assumptions became hypotheses to validate against real user behavior in the survey and interviews that followed.

To reach both current and non-users of the app, I designed a branching questionnaire — respondents were routed to different question sets depending on their app usage history, ensuring every question was relevant to who was answering it.

The survey reached 42 respondents, of which 18 were active Flo users — the group whose feedback most directly shaped this study.

• ✦71.4% had used a women's health app before
• ✦Flo was the most recognized app among them, used by 70%
• ✦Average satisfaction among active Flo users: 4.4/5
• ✦Almost all recommended Flo and reported it improved their routine — most often citing accurate cycle predictions and reduced reliance on hormonal contraception

• "It's very accurate — the more data I give it, the more it gets my cycle right."
• "I stopped taking birth control, and now I use it to understand my body instead."

Four interviews explored how participants actually related to the app day-to-day — how (or whether) they used specific features, and where friction showed up in practice.

• ✦Usage is more reactive than routine. Most participants didn't open the app daily — they checked in around expected symptoms or relied on push notifications to prompt logging, rather than proactively tracking every day.
• ✦The app's value went beyond tracking. One participant pointed out she has no on-call doctor to ask quick questions to, and wished Flo offered a lightweight chat or AI assistant for that — "even a paid one" — showing the emotional support role these apps can play, not just data logging.
• ✦Reminders were a recurring pain point. One participant said pill reminders were "too easy to miss" inside the app and ended up setting a native phone alarm instead — a workaround that directly echoes the notification visibility problem flagged in the heuristic evaluation.
• ✦One participant experienced a calendar desync after switching phones, though she wasn't sure whether it was a sync issue or user error — flagged as a minor reliability concern, not a dominant theme.

With the heuristics and severity scale defined, I audited the app screen by screen, annotating each violation directly on the interface — which heuristic it broke, its severity, and where it occurred (a single screen, multiple screens, or the app's overall structure). Organizing the audit by feature area (Onboarding, Home, Content, Private Chat, Messages, Flo a Dois, Calendar, Settings) made it possible to spot patterns.

Every annotated problem was consolidated into a spreadsheet logging the heuristic violated, its location, severity, and a suggested fix — turning scattered annotations across 8 feature areas into a single, prioritized dataset. This spreadsheet is the direct source for the severity and heuristic-frequency breakdowns shown in Key Findings below.

Self-reported satisfaction closely tracked the app's public store rating — the polished surface experience holds up. But it masked a deeper pattern: underneath a well-liked interface, 38 usability problems were still present, 4 of them catastrophic.

• Survey4.4/5
• Interviews4.1/5
• App Stores4.7–4.8/5

Strengths

• ✦Intuitive design, useful reminders, and health insights that empower users.

Weaknesses

• ✦Free version lacks necessary features (AI Chat, advanced reports).
• ✦38 usability problems identified. The most frequently violated heuristics were:
  • Consistency and Standards (11 problems — 28.9%)
  • User Control and Freedom (5 problems — 13.2%)

• ✦Most usability problems were minor: out of the 38 identified problems, 16 were classified as minor problems, suggesting that the majority of adjustments needed would have only a small impact on overall usability.

• ✦The 4 catastrophic problems
  • No way to undo a submitted chat response
  • No back navigation in several flows
  • No way to preserve entered data when linking an account
  • Premium settings toggles disabled with no explanation of why

One interview participant independently reported missing pill reminders and resorted to a native phone alarm instead — echoing a notification visibility problem also caught in the heuristic evaluation. A finding surfacing through two independent methods is a stronger signal than either alone.

• ✦Add the ability to undo or edit a submitted chat response, and restore consistent back navigation across all flows — directly addressing the two most severe control-and-freedom violations found.
• ✦Preserve user-entered data by default when linking an account, removing an unnecessary point of data loss.
• ✦Replace the disabled Premium toggles with a clear explanation of why those settings are locked.
• ✦Consolidate inconsistent icons, labels, and component patterns across Settings and Content — the two areas most affected by Consistency and Standards violations.
• ✦Reassess reminder visibility: even satisfied users reported missing notifications and relying on external alarms instead.

Beyond usability, this study reinforced how much Flo already does to empower women with accessible health information — which is exactly why closing these gaps matters. This study was submitted, approved, and graded as part of my UX Design thesis at PUC-RS. It's shared here as a portfolio reference — please don't copy or redistribute the content without permission.