Product Designer

Friend groups still wanted to stay close, but coordinating became harder once everyone had different schedules, schools, jobs, locations and routines.

Most of the existing tools only solved one piece of it. Group chats were good for conversation, but plans got buried. Calendars were useful, but separate from the actual friend group. Camera rolls saved memories, but not in a shared group space.

Ying Zhen, my business partner, and I ran a validation survey while the app was still an idea, and it reached 40 people. When we asked which features they would find most useful in an app for connecting with the people they love, a group chat with planning tools inside it beat every feature that stood on its own.

• A group chat where you can make polls and share video: 17.
• A photo-sharing widget: 13. A shared calendar: 13.
• Customizable profiles, widget and group: 12.
• An events list: 10. Shared to-do lists: 9.
• None of the above: 0.

Asked whether they would download an app with all of it, 23% said very likely and 60% said somewhat likely.

Ying also interviewed families and young adults in Chile and contacts in China. The finding from Chile changed my build order: tight-knit families who already live near each other need planning tools more than they need another place to post photos. That is why the shared calendar moved to the center instead of staying a nice-to-have.

I wrote this line in my notes and then built against it.

In January 2025 the whole product was one diagram: an app with five modules hanging off it. Almost everything that followed is a child of this drawing.

The first proof of concept was not mine. Ying is not technical, so she generated a placeholder interface and labelled what it should carry: a poll tab, an upcoming events tab, a shared calendar, and a palette of eight tools. Generic chat, real argument about what belongs in it, and something to react to before I built anything.

I learned Figma at the start of this project so I could test the idea visually before committing it to code. These are the first screens people ever reacted to.

Then it had to run on a phone. This is the first build in the spring of 2025, on mine. The modules exist and almost nothing is finished: home is a placeholder telling me where memories will go, the group chats have no names, and the settings drawer is a list of words.

I sat with 15 people one at a time, across five rounds spread over several months, and watched them use whatever version existed that week. Every note went into one document. When several people hit the same wall, it became work. When opinions split, my business partner made the call.

The notes are blunt, which is what makes them useful. Four of them changed the app the most.

People got three screens into sign up and then hit a button that did nothing. The handwritten username field looked charming to me and read as decoration to them. I rebuilt the step so the field explains itself, Skip goes somewhere, and every screen can be left.

The frames were images laid over a photo, so the photo sat in the wrong place and the polaroid ran off the edge. I rebuilt framing as one component that every screen shares, so a memory looks the same on the board, in the chat and on the screen where you make it. The caption now sits on the polaroid strip, where you would write it.

The survey had already said a chat with tools inside it was the most wanted thing in the app, and the sessions said the tools were hiding. The palette became a horizontal row of things that work: a poll, a shared to-do list, a bill split, the camera, a nudge, and the day's question pinned at the top.

Home was a list of chats and nothing else, so there was no reason to open the app unless someone had messaged you. It now opens on what changed since you last looked: the newest memory, the next thing on the calendar, and the group step streak.

This is what shipped. Every screen below came out of the current build on the same phone, on the same afternoon.

Onboarding: welcome, log in or sign up, pick a username, a welcome by name, and the introduction questions.

Then the circle itself: home, the settings drawer, making a group, the chat room, and the calendar with its month, week and day views.

And the memory side: the board of everything a group has saved, one memory with its reactions and replies, and the screen where a photo gets its frame and stickers. The stickers are my own drawings, cut from a page of hand studies.

I was the only developer, so every design decision came back to me as an implementation problem. Three of them changed the design.

• The sidebars are overlays, not routes. Expo Router turns the file tree into navigation, and making settings a route meant losing the screen underneath it. Sliding it over the top kept the app feeling like one place.
• Everything saves on the phone first. A demo cannot depend on a network in a room full of judges, so the app keeps its own state and the cloud version sits behind the same screens.
• Stickers ship as local assets. Pulling them from a service meant a network call in the middle of making a memory, so the set is small, drawn by me, and always there.

We pitched Clique at the University of Portland Pilot Venture Challenge and placed second. The pitch carried a business model, freemium tiers, a five-year projection and a $100K ask, and the app in the demo was the real one, running on a phone, not a slide.

Two questions from the judges did more for the product than the placing.

• How would privacy work if the app stores personal friend-group memories, messages and calendars?
• How does Clique avoid becoming another app that keeps people scrolling?

Both pushed in the same direction: private circles, clear permissions, calendar control, and memory-based interaction instead of a feed. The answer to the first one is in the app now. Your own calendar is read into anonymous busy blocks that never leave the phone, so a group can find a free hour without anyone seeing anyone else's day.

I cut a feature, and the next round of testing brought it back. The step tracker was dropped early because it read as a fitness feature in an app about staying close. Later sessions said the opposite: step counts are already something friend groups compare. It came back as a streak that unlocks sticker sets, which is the version people described rather than the one I first built.

My prettiest idea was my least usable one. The handwritten username field is still the thing I liked most in the first build, and it is the thing testers stumbled on hardest. Watching someone hesitate over it taught me more than any amount of staring at my own screens.

I trusted the survey too late. The number that mattered, a chat with tools inside it beating every standalone feature, was in front of me before I wrote any code, and I still spent the first build making screens look right instead of making the chat do things.

What I would test next. Everything multi-person is still unproven: whether reminder schedules feel like care or nagging, whether a group actually settles a bill split, and whether find a time survives contact with five real calendars. That needs a second phone and a server, which is the next build, not this one.