A prediction market experience designed for people who follow the news but have never traded before.
Product Designer · 10 weeks · 2024 · 2 PMs, 3 Engineers, 1 Designer · Responsive Web
Tools·Figma · FigJam · Amplitude · Framer
Prediction markets let people forecast real-world events, but most products in the space look and feel like trading terminals. I redesigned the experience for people who follow the news and have opinions, without requiring them to learn order books in order to participate.
First-time users didn’t understand what “shares” or “odds” meant in this context.
Order books assumed financial literacy that most participants didn’t have.
There was no clear first step between browsing markets and placing a forecast.
Help newcomers place a first forecast with language and visuals they can follow, while preserving depth for more advanced use.
Most participants had never used a prediction market, so research started with how they already reason about likelihood and risk — before introducing interface patterns that only made sense to traders.
Tested language and visual metaphors for 'probability' with 15 non-expert participants.
Analyzed 6 existing markets to map the exact moments where newcomers dropped off.
Ran open card sorts to understand how people naturally categorize events and topics.
newcomers understood the default order book
higher comprehension with % framing
preferred topic-based discovery
31 · Product Manager
“I follow the news obsessively and have opinions — I just don't want to feel like I'm day-trading to share them.”
I restructured discovery around topics and questions, matching how people already discuss the news.
Low-fidelity work focused on the highest-risk moment: placing a first forecast. I explored sliders, cards, and plain-language toggles until one framing felt clearer in walkthroughs.
The brand direction aimed for an editorial feel closer to reading the news than opening a brokerage product. Color and type needed to stay clear and restrained.
Warm paper backgrounds soften the finance association. Lavender marks primary actions. High-contrast text keeps probabilities readable; graphite supports secondary metadata.
Ink
#121212
Primary text
Paper
#F7F5F2
Page background
Lavender
#C4B5FD
Primary actions & focus
Graphite
#5C5C5C
Secondary text
Line
#E6E2DC
Dividers & borders
Fraunces gives market questions a more human, editorial tone. Instrument Sans handles the interface, probabilities, and supporting copy. Large percentage numerals carry most of the visual hierarchy on each card.
Instrument Sans
UI and body copy
Fraunces
Question headlines & market titles
High-fidelity design took a more editorial direction. Each card leads with the probability, and a lavender accent marks primary actions — moving the visual language away from a trading terminal.
Every market opens with a plain-language question and a clear probability, making forecasting feel closer to sharing an informed view on current events.
A plain-language question and a large probability on each card.
A slider shows potential outcomes in dollars before confirmation.
Browse by politics, sports, tech, and other familiar topics — aligned with how participants sorted events in research.
newcomer activation rate
drop-off on first forecast
7-day return rate
The most effective change wasn’t a single interface detail — it was rebuilding the product around questions. Once markets read like something people would say out loud, much of the intimidation fell away.
Next