AI feed boosted retention 3× in a 1M-user app
Me (design + medical insight) + 1 developer
Roles I played:
UX Designer – Led the product from research through delivery
Doctor – Ensured medical accuracy in emotional forecasts, cycle education, and tips
Visual Designer – Created UI, data visualizations, and motion effects (including haptics) to make the app feel warm and human
Strategist – Turned thousands of feedback points into habit loops, product decisions, and monetization ideas
Feature 1
Women struggled to understand how hormones influence emotions in repeating patterns — like “seasons”
Solution:
Designed a visual mood forecast, like a weather app for feelings
Paired it with gentle illustrations and personal tips (“Low energy today? Try a slower start”)
Impact
Became one of the most shared features
Validated how users felt and gave them clear language to describe it
Feature 2
HRV (heart rate variability) gave a stress score, but raw numbers felt boring and meaningless
Solution:
Designed a visual: stress shown as liquid rising or falling in a soft-colored container
Used different colors to signal intensity
Paired with a daily tip matched to their energy level
Impact
Made stress data tangible and easy to interpret
Turned an abstract metric into something users could act on immediately
Feature 3
We saw users struggling to scroll through a long symptom list every day. So the feed showed the most likely symptom for their current phase: “Cramps”, “Backache”, “Mood swings”. No typing, no extra taps → log in seconds
Impact
Higher completion rates
Users felt remembered — like the app “knew” them
Feature 4
We wanted users to recognize patterns in their own body:
So we showed the top symptom of the month
Clicking it showed how often they logged it, and when
We added “You’re not alone” context. e.g., “3,201 women logged this today”
That one stat made people stop, read, and reflect.
Feature 5
Users didn’t have a reason to log in daily but retention is important for both helping the user and increasing the revenue.
Solution:
Added daily streaks with small, joyful celebrations
Sent personal notes after logging, tailored to symptoms:
“Add omega-3 rich food for PMS”Gave weekly streak discounts that felt earned, not like generic offers
Impact
Turned logging into a dopamine loop users wanted to maintain
Discounts performed better and reduced “offer blindness”
Feature 6+7+8
We pulled sleep data directly from Apple Health and translated it into a friendly summary:
“You’ve been sleeping on time 4 days this week.”
This wasn’t just data for the sake of data, it became a signal of recovery, and helped users make sense of their mood or energy dips.
It reinforced a subtle behavior loop: better sleep = better day.
Period analysis gets boring and We designed period patterns to feel visual, not clinical.
Instead of a dry calendar, we used color gradients and cycles to show rhythm.
We added insights like, “Your average cycle length is 28 days,” paired with visuals that updated as they tracked.
BBT is a guide for fertility tracking. For users trying to conceive, we used BBT (Basal Body Temperature) to offer timely, clear nudges:
“Your temperature is up — this might be a good time to try.”
We made this opt-in and privacy-first, and it became one of the most trusted features.
Simple cues, no jargon.
Feature 9
We introduced an AI-powered chat, a space where users could ask anything, and the system would respond based on their individual data.
But we knew people don’t always know what to ask.
So we made it easier to start.
We added a daily card in the feed with a hormonal health tip, changing every day.
When clicked, it opened up one actionable tip each across four areas: Food, Mental health, Relationships, Exercise
Why these?
They were the most asked topics in our feedback and community forums—especially around PMS, emotional health, and lifestyle habits.
We even went further: offering follow-up questions like:
“Can I get a simple recipe for this phase?”
“How can I manage PMS in my relationship?”
“What's a gentle 5-min workout for today?”
Users didn’t just engage, they came back more often.
AI chat usage jumped 200% in the first few weeks.
Paywall A/B test
When we gave away the Daily Feed on day one, and then hid it behind a paywall, conversions jumped 25%.
Impact
Once the user experienced the value of daily feed and then see it locked, they wanted the value back and were more likely to convert.
In App Feedback
After users had seen the Daily Feed 7 times, we gently asked:
“Was Today’s feed useful?”
70% said yes.
We followed up with:
“Which part helped most?”
The top picks were:
Period history graph
Emotional forecast
Stress, energy & health tips
And then we asked, “What could be better?”
Post-feedback
Looking at the feedback we realised, users were interested in exploring each feature in-depth, which means they are curious and our feed is helpful.
We delivered it in next update.
Infact we went one step ahead and added Health quiz to make things more interactive.
