Case study
UpNext: start with an action
Independent product · Product design, interaction design and AI-assisted development · Web & iOS
UpNext offers everyday activities as cards. Choose one, try it, and see what you want to keep doing. I designed the product and its interactions, then built and shipped it with AI assistance. This case follows the first choice, from the web app to the details I had to work through in Swift.
Start with an action
Try a small action. Find what fits.
UpNext offers everyday activities as cards. Choose one, try it, and see what you want to keep doing. I designed the product and its interactions, then built and shipped it with AI assistance. This case follows the first choice, from the web app to the details I had to work through in Swift.
Before the first step
Before I could start, I had to make a plan.
Habit apps asked me to work out what I wanted to do and make a plan for it. I found that preparation tiring. I wanted to try an activity first, discover whether I enjoyed it, and build from there. That became the question behind UpNext: how could the app give me something manageable to try?
A hand of cards gave me a starting point.
I liked how roguelikes gave me a hand of possibilities to consider. Nudge research also influenced me: help someone make a choice while leaving it theirs. I brought those ideas together in UpNext. The app would offer activities; the person would read the cards and decide what fit their day.
Making the choice yours
Give the first choice something concrete.
The app draws six different cards from the available pool. Each names an activity, such as walking 1,000 steps or stretching for five minutes. Depending on the mode, you choose one, two or three. The cards give the first choice a manageable boundary. They are drawn from a pool, without assessing someone’s individual needs.
Give people time to read and reconsider.
A card opens into a preview before it joins today’s selection. You can remove an ordinary card, and the board starts only after confirmation. An early tester report also led me to add space below the confirmation control so the navigation would not cover it. Reading, changing a choice and confirming needed to feel like separate actions.
Keeping the feel in Swift
Swift needed the same room to explore.
The first Swift version put the cards in two columns and used a tap to draw. I could choose a card, but the web’s hold, reveal and inspection sequence was missing. I rebuilt that sequence in the native app so there was still time to explore the hand before confirming. The three recordings compare the restored early Swift commit, local iOS build 35 and the current web component.
Let a tap stay a tap.
On iOS, tapping a card could compete with its drag gesture. I separated the tap action and required a clear vertical movement before treating a drag as selection or dismissal. I also shortened the card’s entrance and aligned the selected-card slots. I made these adjustments to keep an exploratory touch separate from a decision.
The response should end with the gesture.
Both versions use an 800 ms hold to open the deck. Progress, sound and haptics must start together and stop if the person releases early or leaves the screen. I tied their start and cleanup to that gesture. Testing the iOS version in TestFlight helped me catch differences that were hard to judge from code.
After the first action
Show what happened after the action.
Confirmed cards become today’s board. Completing an activity changes its state and awards the applicable XP once. A level-up opens a card pack, adding activities to future draws. I wanted a finished action to leave something visible and give the next visit a starting point. Completion is still reported by the person using the app.
Give the next visit a way back in.
Finishing the first activity is also a moment to set up the next visit. The app offers a reminder after the first completion and widget guidance after the second. These recordings show those actual transitions. I placed setup after an action so it has a purpose the person has already experienced. I have not measured its effect on return visits.
Where I would take it next
Does a card look read, selected, or confirmed?
The preview, selected slot and confirmed board are different states. The next usability session would ask someone to inspect one card, compare another, then change the selection. I would watch whether the slot and exit motion are enough to explain what changed, without asking them to narrate every gesture.
Can they resume without starting over?
Choosing a card can be interrupted by a notification or another task. I would test returning while reading a card and returning after selection: is the current state clear, does the next action remain visible, and does the person repeat a step unnecessarily? This is a focused follow-up test, not a result already measured.
Try it for yourself.
UpNext is available on iPhone and Android. Draw a hand, explore the activities, and choose something to try today.
About
About the designer
I'm Jongmin, a product designer exploring design engineering. I care about how things work, and how they feel.
I think through the whole experience: what someone needs, how they get there, and the details that make it feel right.
Product design is my starting point. I'm exploring design engineering to bring more of those ideas into working interactions.
Contact
Something on your mind?
A role, a project, or just a hello. I'd like to hear it.