Clue · Case study · Health app · Sep – Nov 2023
A useful app that felt distant.
Self-initiated case study on the Spanish version of Clue. I contrasted what the app prioritizes with how its users actually use it, and reviewed the conclusions with a lead designer from Clue’s first design team.
Each stone is a metric from this case.
Users in unmoderated tests (Maze)
Usability score of the prototype
Moderated tests
Companies in the benchmark
Context
Clue is one of the best-known period-tracking apps. We chose its Spanish version for a self-initiated case study: with no client or brief, we defined the scope ourselves across three areas — sign-up and onboarding, home screen and tracking.
With no product team, the hard part was deciding which problem was worth solving. So we crossed four sources before drawing anything: 9 moderated tests, 14 forms, a 5-company benchmark and tests with 29 users. At the end we checked our conclusions with someone who had worked inside Clue.
My roleresearch, synthesis and scope
We were two designers working as peers for four sprints. With no product team, we also did that work: defining scope, deciding what to prioritize and what to leave out.
Research
- 9 moderated tests and 14 forms
- Benchmark of 5 companies
- Unmoderated tests with 29 users on Maze
Synthesis and prioritization
- Findings by area: onboarding, home and tracking
- One main insight that organized the work
- Prioritization with MoSCoW
Prototype and cross-check
- Sketches, wireframes and prototype, with KPIs defined up front
- Iteration based on the Maze results
- Review with a Clue lead designer
The bottleneckthe app was talking about something else
The app worked: sign-up was simple and its inclusive language built trust. The problem wasn’t one screen; it was what the product talked about.
Storytelling and tone centered on the reproductive side of the cycle.
That’s the side of the cycle the app’s tone pushes toward.
As a self-knowledge tool and a source of information for their medical appointments.
For that they need to remember, tell things apart and log clearly.
Onboarding
Simple sign-up, but a design with “almost no identity”.
Home screen
Trust thanks to the inclusive language; confusion between calendar and analytics.
Tracking
Icons and colors that look too alike, and buttons that get in the way when saving.
“Me parece un diseño muy sobrio casi sin identidad”
Claudia, 36
“Siento que tengo la información pero no la forma de recordarla”
Candy, 24
“Me es confuso el uso variado de colores y cómo puedo agregar mis síntomas”
Sara, 28
What I did
- I organized the findings by area and reduced them to one main insight: the app is useful and intuitive, but feels distant.
- With my partner we defined the KPIs before prototyping: time per task, success rate and error rate.
Decisionswhat goes in and what stays out
We prioritized with MoSCoW. The most important decision wasn’t what we added, but what we left out.
Consistency first, features later
The “must have” was harmonizing fonts, sizes, numbers and icons and strengthening UX writing, before adding anything new.
Bring analytics to the home screen
Analytics on the home screen and reminders in the bottom bar, so users can remember and understand their data.
Not building the forum
A forum between users went into “Won’t have”: it was a different product bet, and the research pointed to clarity and self-knowledge.
Must have
- Harmonize fonts, sizes, numbers and icons
- Strengthen UX writing
- Integrate analytics on the home screen
- Reminders in the bottom bar
- Key icons to improve the information
- More value in the Plus invitation
Should have
- Logo and slogan on certain screens
- Cycle chart colors
- Calendar palette
Could have
- Icon on the “What do you want to do?” onboarding question
Won’t have
- Discussion forum between users
Testingwhat worked and what didn’t
We tested the prototype with 29 users on Maze: three tasks and three KPIs — time, success and error. The usability score was 80/100.
Task 1
Task 2
Task 3
Source: unmoderated Maze tests with 29 users, run on the prototype.
What the data said
Task 3 was the weak spot.
- 70% didn’t complete it, 60% dropped off and the click error rate reached 57%.
- Tasks 1 and 2 did better on success, although task 2 showed a high click error rate.
Cross-checkvalidating with someone inside
During my job search I had mentorships on ADPList. One was with a lead designer from the first design team at Clue, the German company, who reviewed the case.
What I heard
Several of the insights matched topics their team had already worked on.
- The proposals pointed to features and improvements they had already identified.
- This isn’t validation of implementation: it’s a signal that the research, without access to their internal data, reached similar questions.
Evidencefrom sketch to prototype
Process
Final prototype — onboarding
- Home cards that are easier to read with a simplified design
- Onboarding copy reviewed for clarity
- Smaller buttons so they don’t compete with the information
Final prototype — home screen
- Plus invitation with relevant information in context
- A single screen with the key period information
- Redesigned cycle icon and alerts in the main bar
- Option to switch the app mode from the home screen
Final prototype — tracking
- Symptom-customization button separated to stand out
- Different type weights to tell important dates apart
- More legible icons with color categorization
UI Kit
#225270 · #3B8EA5 · #B8D4DD · #D4848A
Hierarchy of headings, body, call-outs and subheads.
What I learned
An insight is a difference, not an opinion
What mattered wasn’t saying the app “felt distant”, but showing the gap between what it communicates and how it’s used. That gap organized the prioritization, and it’s what a Clue designer recognized as close to their own work.
“In this project, UX isn’t only about simplifying; it’s about inspiring confidence and control over one’s own body.”
Melany Olivares