Demand exists. The participation model doesn’t.
Polymarket, Augur, and a wave of DeFi protocols already proved that people want to predict outcomes on-chain. What never scaled was who could participate.
Every existing platform routes users through orderbooks, liquidity, and trading mechanics before they place a single prediction. The interest is real. The technical barrier is what kills it before it becomes action.
Two audiences. Opposite needs. One product.
This wasn’t a UI simplification problem. It was a participation model problem, and solving for only one audience meant losing the other.
The curious newcomer has heard of prediction markets but never joined. Drop-off happens if orientation takes more than 30 seconds. The experienced DeFi user wants fast engagement cycles, not long-term position management, and bounces when the product feels too shallow.
Designing for traders alone keeps the category small. Designing for newcomers alone alienates the power users who drive volume. The product had to hold both.
Light
Dark
Benchmark the category before opening Figma
Competitive benchmarking across six platforms, Polymarket, Augur, Drift Protocol, Binance Predictions, Azuro, and Limitless, cross-referencing interface patterns with perceived friction.
Five insights guided direction. The decision to avoid orderbooks wasn’t a late pivot, it was clear before any prototype existed.
- Users stall on orderbooks not from lack of intelligence, but because that system was built for financial professionals.
- The more data visible at the moment of decision, the higher the hesitation and drop-off.
- Weak feedback loops mean users don’t understand what happened after their own prediction.
- Participation data is a decision signal. Hiding what others are doing ignores real behavior.
- Short cycles build habit. Long cycles tied to external events don’t.
Orderbooks were the single biggest barrier to entry across the entire category. We chose a different path from the start, not after testing and discarding the obvious one.
Five decisions that changed the participation model
The platform was designed for desktop and mobile from the outset, with every state built and tested across both breakpoints.
01
Pooled positions, not an orderbook
Users deposit into a shared pool and pick a side, UP or DOWN. The decision becomes binary, not a trading operation. No counterparties, no liquidity literacy required.
02
Time-based rounds, not a continuous market
3-minute preparation: choose amount and direction. Room locks, no new entries. 2-minute resolution counts down to outcome. Total cycle ~5 minutes: short enough to build habit, contained enough to reduce anxiety.
03
Participation data as interface, not backend
Pool distribution is public from day one, a visible table of what each participant bet. Social proof for newcomers. Contrarian signal for experienced users. Either way, the decision stays inside the product.
04
One dominant action
UP or DOWN is the strongest visual element on screen. Amount and confirmation are deliberately secondary, the interface answers one question at the moment that matters.
05
Resolution as a loop, not a dead end
Win/loss appears as a modal at round end. Acknowledge it and you stay on the same screen, ready for the next round without navigating away. The emotional peak channels directly into the next action.
Iteration after beta: Live chat wasn’t in the original launch. During the closed test, some users still hesitated when choosing a side, even after understanding mechanics. Chat opened to everyone with an active prediction in the round, so decisions stopped happening in a vacuum.
Production beta, not A/B theater
No version comparison in a lab. Validation happened in production with people who already wanted to use the product.
A closed beta ran 5 days: 5 groups of 20 users (100 total), all from the product’s own waitlist, an audience that arrived with real intent, not a random sample.
More than half already understood the product’s purpose before entering, thanks to education on the landing page. Combined with a self-explanatory flow, 100% of participants understood mechanics by the second round. Individual win/loss outcomes varied normally, that was never the success criteria.
Founders and stakeholders who followed the beta shared the same read as users: intuitive, clear, delivers on what it promises.
Measured outcomes and delivery impact
Time to first action
Understood mechanics by round two
Waitlist conversion · zero paid
100% of beta users understood mechanics by round two. Time to first action dropped from ~10 seconds to under 2–4 seconds across desktop and mobile.
- Design system: 120+ components cut handoff time by ~35%
- Waitlist funnel: the same audience that validated interest on the landing page validated the product in beta
- Qualitative alignment: users and internal team agreed, intuitive, clear, purpose delivered
Where users still hesitate
The most critical moment today is the 3-minute preparation window, when a user decides alone on amount and direction, with no real-time guidance, before the room locks.
That’s where newcomers are most likely to get lost. The top priority is bringing guided support at room creation, before lock, not after, building on what live chat taught us in beta.
When users are confused, fix the mental model before the UI. We never polished our way out of an orderbook. We replaced the model, then designed the loop around it.
