Designing a bedtime app that wants you to put your phone down
Mude — a behavioural design exploration into bedtime procrastination
- Project
- Self-initiated product concept
- Role
- Product Designer / UX Researcher
- Timeline
- 1 month
- Methods
- Secondary research · Survey · Journey mapping · Prototyping · 5 guerrilla usability tests · 3 moderated usability tests · Iteration
We've probably all done it.
You're tired. You know you should sleep. Tomorrow-you would very much appreciate it.
And yet somehow you're still scrolling at 12:47am.
Mude is a self-initiated product concept exploring bedtime procrastination — the gap between intending to go to bed and actually doing it.
The challenge wasn't simply:
How do I help people sleep?
It was:
How might a digital product help people emotionally wind down and leave their phone — without becoming another reason to stay on it?
The 30-second version
Problem
People sometimes delay bedtime even when they intend to sleep, often filling that time with digital activities.
Insight
Going to bed isn't only a scheduling problem. For some people, staying awake provides a final period of personal time or emotional decompression after the day.
Design response
I explored a short ritual built around capturing meaningful moments, reflecting on the day and then deliberately ending the digital interaction.
What testing changed
My original assumption that journaling was inherently calming was wrong.
Users pointed out that revisiting a frustrating or upsetting journal entry before bed could do exactly the opposite.
That changed the product.
I started with the behaviour, not an app idea
I wanted to understand why someone who wanted to sleep would repeatedly choose not to.
My exploratory research combined secondary research with a survey examining bedtime habits, stress, digital behaviour and activities people already found helpful before bed.
The research pointed towards a more interesting problem than poor time management.
Bedtime procrastination could represent an intention–behaviour gap.
Someone might understand perfectly well that sleep is important while still choosing the immediate reward of another episode, another video or a little more personal time.
So instead of designing another app that said:
"It's bedtime. Go to sleep."
I became interested in the transition between being active and being ready to stop.
That became the product opportunity.
The core tension
How do you use a phone to help someone stop using their phone?
This became the most important constraint in the project.
Most digital products are designed around engagement:
Open → interact → discover more → come back
That felt wrong here.
If Mude succeeded by increasing minutes spent in the app, it could actually be making the original problem worse.
So I defined a different design principle:
Mude should earn a few intentional minutes of attention, then give the user's attention back.
That meant avoiding infinite content, unnecessary browsing and interactions designed purely to prolong engagement.
The experience needed an ending.
Research helped me narrow the behaviour
My initial exploration generated plenty of possible features:
journaling, mood tracking, sleep tracking, galleries, personalised prompts, achievements, environmental recommendations, reminders, integrations and more.
Early concept sketch — explored, not part of the final scope
This week
But a one-month concept did not need to become an entire wellness ecosystem.
I narrowed the core experience to three behaviours:
Notice
Capture small moments during the day before they're forgotten.
Reflect
Return to selected moments as part of a short bedtime ritual.
Switch off
Create a clear ending that encourages the user to leave the digital environment.
This gave me a much simpler product loop:
Notice during the day → Reflect at night → Say goodbye to today → Switch off
Decision 01
Don't make users become poets before bed
Journaling sounded useful in theory.
In practice, it creates a very obvious problem:
"What am I supposed to write?"
My early journey mapping identified the blank page as a likely drop-off point. If reflection started to feel like homework, it wouldn't make a convincing bedtime ritual.
So I explored multiple low-effort ways to capture something:
- A short note.
- A photo.
- A voice entry.
- A lightweight prompt.
The goal wasn't to produce a perfect journal entry.
It was to help someone notice:
"That was a nice part of my day."
Prompts could help when people were stuck, but users always retained control over what they recorded.
Decision 02
Capture moments when they happen — not only when you're already tired
Originally, I thought about Mude mainly as something people would use at bedtime.
Mapping the journey changed that.
Meaningful moments happen throughout the day.
If users were expected to reconstruct everything at 11pm, they would have to remember what happened, decide what mattered and then write about it when they were already tired.
So I introduced multiple entry points.
Someone could quickly capture a moment during the day and return to it later during their wind-down.
That divided the cognitive work:
Daytime: notice.
Bedtime: reflect.
The bedtime experience could therefore remain lighter and calmer.
Decision 03
Make wind-down a ritual, not another content feed
I wanted the evening interaction to feel deliberately different from the products users might have been scrolling moments earlier.
Instead of an endless feed, Mude had a finite sequence.
The user could prepare their environment, revisit selected moments and move towards a deliberate ending.
One interaction I explored was a "goodbye to today" ritual inspired by tearing a page from a calendar.
It gave an abstract action — finishing the day — a visible moment of closure.
The interaction wasn't there just to create delight.
It communicated:
Today is finished. You don't need to keep doing things.
Then users challenged my biggest assumption
After creating the first high-fidelity prototype, I ran three moderated usability-testing sessions with follow-up interviews.
And one participant asked a question that changed the direction of the project:
"What if I write about something that annoyed me?"
I had assumed that journaling followed by reflection would create a reassuring wind-down.
But journaling isn't automatically positive.
Someone might use it to process:
an argument, a stressful meeting, anxiety about tomorrow, or something they were angry about.
Automatically resurfacing that entry just before sleep could be the opposite of calming.
It exposed an important flaw in my original concept:
Capturing an emotion and reliving an emotion are not the same user need.
That was more useful than a participant simply telling me they liked the prototype.
Decision 04
Separate "things I want to remember" from "things I need to let out"
Instead of abandoning journaling, I reconsidered the role different entries should play.
Some moments are worth revisiting:
- A good meal.
- Something funny a friend said.
- Finishing something difficult.
- A small thing you're grateful for.
Other thoughts may simply need somewhere to go.
So I explored separating positive reflection from emotional release.
Content intended as an outlet didn't need to automatically return during the bedtime recap.
This gave the user more control over their emotional experience and challenged my earlier assumption that more personalisation or more resurfaced content was necessarily better.
The lesson was simple:
Don't optimise an experience around the happy path of someone's emotions.
Let it out
Testing also made the experience quieter
The first round of guerrilla testing with five participants exposed several more practical problems.
My opening screen felt redundant.
Some button choices needed better accessibility.
The wind-down flow contained too much text.
And parts of the sequence felt more like using an app than winding down from one.
The second testing round challenged the visual experience further.
Participants questioned whether the interface felt sufficiently calming, and some found the slideshow-style playback potentially more stimulating than relaxing.
So I reduced unnecessary content and reconsidered motion, visual density and the sequence of the experience.
The objective wasn't:
"How can I make this more impressive?"
It became:
"What can I remove?"
Decision 05
Design a finish line
A bedtime product needs something most apps actively avoid:
a reason to leave.
So the final part of Mude deliberately closes the interaction.
After reflection and the goodbye ritual, the interface transitions towards Switch Off.
Rather than suggesting another activity or presenting more content, the experience encourages the user to put the device down and can connect with the phone's Sleep / Do Not Disturb mode.
There is no infinite feed waiting underneath.
No "you might also like…"
No streak demanding one more action.
Because, in this context:
Closing the app can be a successful conversion.
Good night
The final concept
Mude became a three-part experience.
01 — Notice
Capture meaningful moments throughout the day with minimal effort.
02 — Wind down
Revisit selected positive moments through a short, finite bedtime ritual.
03 — Switch off
Mark the end of the day and deliberately leave the digital environment.
The product isn't intended to diagnose or treat sleep disorders.
It's a behavioural design exploration around a narrower question:
Can we make the transition from "one more thing" to "I'm done for today" a little easier?
What did I actually validate?
Because this was a self-initiated prototype, I did not have a live product, production analytics or a large enough sample to claim improvements in retention or sleep.
What I could evaluate was the design itself.
Across two rounds of testing, I identified issues with:
navigation, accessibility, information density, interaction sequencing, emotional assumptions and the overall wind-down concept.
Most importantly, testing invalidated part of my original hypothesis about journaling and led me to change the product model.
For this project, that is the outcome I would stand behind:
I finished with a better problem definition than the one I started with.
If Mude were shipped, what would success mean?
Traditional engagement metrics could be misleading here.
More sessions and more screen time might actually indicate a worse outcome.
Instead, I would evaluate whether Mude helps users follow through on their own bedtime intention.
A primary behavioural signal could be:
Intended-bedtime adherence
How often does a user finish their wind-down close to the bedtime they set for themselves?
I would support that with:
Wind-down completion — Do users reach the intentional Switch Off moment?
Time to switch off — Does the ritual remain short rather than becoming another prolonged activity?
Repeat voluntary use — Do people find the routine useful enough to return to over time?
Self-reported calmness — Do users feel more ready to disengage after the ritual?
Post-routine screen behaviour — With appropriate permission, do users actually leave their device after finishing?
Sleep quality could be explored as a longer-term self-reported outcome, but I would not claim that Mude caused improved sleep without substantially stronger research.
What I wouldn't build yet
This project originally generated ideas around AI-generated content, sleep tracking, wearables, achievements, social features and deeper personalisation.
I wouldn't prioritise those next.
Not until the core behaviour works.
Before increasing scope, I would want to answer:
Do people actually find the Notice → Reflect → Switch Off ritual useful enough to repeat?
Only then would I decide which additional capabilities genuinely strengthen that behaviour.
What I learned
The best usability test doesn't always validate your design.
The most valuable feedback I received wasn't:
"I couldn't find this button."
It was a question that challenged the premise of the experience.
That reminded me that testing isn't just about checking whether users can operate a solution.
It's also an opportunity to ask whether the solution should work that way at all.
User engagement isn't universally good.
Mude also changed how I think about product metrics.
For many digital products, longer sessions can look positive.
For a bedtime product, they could represent failure.
Good product design means defining success around the user's goal, not automatically around more product usage.
Sometimes the best thing your product can help someone do is: