Working with AI leaves gaps of time. Those short stretches while you wait for a request to finish. I used to watch YouTube during those gaps, but at some point I ran out of things to watch. So I switched to Chzzk. It's live streaming, so the content never dries up. I started watching more and more.
Watching streams, I noticed how streamers interact with viewers. They run polls and raffles a lot, but the tools supporting those events felt lacking. Since I'd been enjoying the content, I thought I could build something. And honestly, there was another calculation. If streamers used my service during broadcasts, viewers would see the service name, and that might bring traffic to Q-Fit.
Starting with Polls: Chat Integration Was Quick, the Real Challenge Came After
I started with polls because they seemed simpler than raffles — just receive chat messages and tally them. Gemini handled the Chzzk chat integration quickly. It was done in no time, and getting the chat data itself turned out to be less difficult than I expected.
So I built it to show the voter list and real-time chat history on screen. Both streamers and viewers can see who voted for what, and you can immediately tell whether your vote was counted. The functionality landed this way, but honestly I'm not sure how well the UI turned out. I didn't study design, and I can't claim I have a good eye for making screens look nice.
The Raffle Feature: Building on Top of the Poll
The raffle was built as an extension of the poll. Since the poll already tracked which users voted for which option, using that data as the pool of raffle candidates was a natural next step. Having that foundation made it feel less like building something new and more like adding onto what was already there.
I also tried to address the pain points of existing raffle tools. By watching streams, I paid attention to how streamers ran raffles and noted the parts that looked inconvenient.
Things I Thought About from the Streamer's Perspective
- Automatic prize display: When there are multiple prizes or multiple quantities, old tools required streamers to manually check off winners. I made it so you register prizes and quantities in advance, and winners are automatically shown which prize they get, with records saved.
- Subscriber multiplier: For streamers who want to reward their subscribers more, I added a multiplier setting so subscribers have a higher chance of winning than regular viewers.
- 1-on-1 chat with winners: I added the ability to connect directly with winners via chat. The goal was to eliminate the hassle of having to track down a DM after the winner is drawn.
- Absent winner skip: When a winner isn't responding, you can skip them and continue drawing the next person. This keeps the stream from stalling because of one unresponsive winner.
Adding all these features made the raffle screen pretty complex. With more information to show on one screen, I had to think about where to put everything, but honestly I can't tell if it turned out well — I didn't study design. If the screen feels cluttered even with all the features, people might not use it.
Promotion: The Communities Gemini Recommended
After finishing the features, I figured I needed to promote it. But I had no idea where to post. I didn't frequent Chzzk-related communities much. I asked Gemini for promotion channels and it suggested three: Chzzk Lounge, FMKorea, and DCinside.
I posted promotional messages on each community directly. The content was an introduction to the service, an explanation of the features, and a prompt to try it out. It wasn't particularly difficult, but there was one unexpected challenge. I wasn't familiar with community culture, so it took a while to understand what tone to write in, what expressions to use, and the terms that circulated on each board. Every time I came across an unfamiliar term, I asked Gemini what it meant one by one.
Results: Buried, and Community Slang Was Harder
The results were minimal. The boards move fast, so posts got buried under new ones not long after I uploaded them. Views were low, and I have no idea if any streamers actually tried the service. I never saw anyone using it, and I never heard from anyone who did.
Still Glad I Built It
The promotion didn't work, but I'm satisfied with having built the features themselves. There was a certain satisfaction in actually improving things I had personally found inconvenient while watching streams. Without user feedback I can't know whether the features really hit the mark, but I did what I could at the time.
These were features built with an expectation of driving traffic, and that didn't pan out. But I don't regret it. Q-Fit gained a new feature, and I came away with the experience of building Chzzk chat integration. If I build something similar again, I honestly still don't know how I'd handle promotion. I can tell that posting on communities the way I did this time didn't work — but I have no real sense of what would. Maybe I need to actually study marketing at some point.