Skip to content
Jaka FeferMessage me
All work

Crypto derivatives trading

EasyX

Beginner traders were quitting after costly early mistakes, so I designed education-first onboarding and a practice sandbox, and onboarding completion rose about 40%.

Role
Founding Product Designer
Dates
Jan 2022 - May 2024
Team
Small team, product and engineering
EasyX trade screen on a laptop and a phone, with the collateral and position size sliders next to the price chart.

Outcome

  • Onboarding completion rose about 40%
  • About 60% fewer errors on irreversible actions than the pre-design baseline
  • Users who practised in the sandbox first were three times as likely to complete their first real trade without errors

The product

EasyX staking page. A large APR figure, stake and claim actions, and a list of staking positions with deposit and unlock dates.
Three onboarding steps on phones. Choose an asset, choose real or dummy money, and the explainer for long versus short.

The problem

EasyX was built for people who want to start trading but don't yet understand the risks, the terminology or the mechanics. Most platforms assume prior knowledge and push beginners into irreversible actions too early. In interviews and tests, beginners hit three failure modes. They didn't know what a screen was asking, so they hesitated or clicked through. They couldn't tell what would happen next, especially when an action could not be undone. And an early mistake cut their confidence, so they dropped off. Users who made an early mistake rarely came back, however fixable it was. The business goal was to win the beginner segment by cutting churn from those mistakes.

What discovery showed

  • Beginners didn't lack information, they lacked the confidence to act on it
  • Predictability and transparency built more trust than safety features did

Key decisions

Education-first onboarding

Beginners were dropped straight into trading screens and froze. In 6 of 8 interviews people described hesitating on screens where they couldn't predict the outcome, and fear of mistakes was the top reason for not finishing onboarding in the survey.

The CEO wanted the usual connect-and-trade flow. He saw the product as an advanced trader would, and this was a product for beginners. I took him through the interviews and tests with 500+ people and showed where the friction was, then designed onboarding that teaches the core concepts one step at a time and brings up responsibility and safety early.

Made against the founders' preference at the time

Sixteen onboarding steps from choosing an asset to setting position size, each shown plain and with its tutorial version that explains the screen inline.

Result. The completion gain under Impact came from the prototype test of this flow.

Sandbox before real money

Every first trade risked real money, so beginners avoided acting at all and learned nothing.

I designed a sandbox where users practise the full flow with dummy money before they touch real funds. The choice sits inside onboarding as its own step.

Onboarding step four, real money or fake money, shown plain and with tutorial callouts that explain each option.

Result. In usability testing, users who practised first completed their first real trade without errors far more often than those who did not.

Make state and consequences explicit

Users couldn't tell what an action would change or whether it could be undone, and confirmations built for speed let accidental trades through.

I designed flows that state the current state, the consequence and the next step, with confirmations on irreversible actions, without hiding how the system works.

Staking deposit screen with callouts that explain the lock period, the amount being committed and what the confirm button does.

Result. Post-launch analytics showed the drop in errors on irreversible actions against the pre-design baseline.

Reflection

  • I would run structured usability testing earlier, aimed at onboarding comprehension and the moment before an irreversible action.
  • I would define success metrics on day one, so learning outcomes and confidence could be measured over time rather than read off completion rates.