I built a working product. Then realised I’d designed it for an audience I didn’t have.

SiteClash started during the .lol leaderboard rush. I saw the excitement, opened an editor, and started building.
But I wanted to try something different: free entry, with builders rallying support for their websites instead of buying their way up the leaderboard.
The idea was that everyone would bring visitors into the same arena. Someone arriving to support one project could discover the others.
On paper, it made sense. I shipped the first working version quickly.
The routes worked. Votes accumulated. Eliminations ran. A champion took the crown.
The game still felt empty.
I’d built an always-running competition without an audience to keep it moving. A few visits spread across a week didn’t feel like an event. Sites joined at different times, and builders were being asked to keep promoting a platform that hadn’t yet shown it could give anything back.
Every share costs someone a little social capital. I hadn’t given people a strong enough reason to spend theirs.
So I rebuilt the format.
Now, exactly eight websites start together in one weekend battle. Everyone gets the same reveal moment. Seven scheduled eliminations lead to one winner. Votes carry through the fight, so an early push keeps its value.
Entry is free, and paid sponsorship stays outside the competition.
The hope is that concentrating everyone’s effort into one event gives each project a better chance of being discovered. Established projects can participate too. You shouldn’t need a new launch every time you want people to notice something you’ve built.
The first roster is now full. That’s progress, but it doesn’t prove the idea works yet.
The interesting part comes next: will participants share, will visitors explore the other projects, and will anyone want to come back?
One sentence from my original write-up sums up what I learned:
“A multiplayer growth loop cannot bootstrap itself from zero just because the diagram has arrows.”
I still like the idea enough to test it. I’m just much clearer now about what needs to happen outside the code.
Have you ever shipped something that worked technically, then realised it depended on behaviour you hadn’t given people enough reason to adopt?