Cut the path from opening the app to an order from seven steps to four, without losing a single feature
- Margin Trade is an exchange where people trade the price of Bitcoin, gold or Nvidia going up or down, often with borrowed money. Its desktop terminal is built for professionals and is dense on purpose. Simple mode is the same product on a phone, for everyone else: four tabs, one Buy / Sell bar, every professional feature one layer down, and the full terminal one switch away.
Seven steps between opening the app and placing an order.
The first mobile app was the desktop terminal squeezed onto a phone. It opened on a chart with nothing about your money. Buy / Long opened eleven settings at once, then margin mode, leverage and time in force; seven steps in, Place Order. Pros could use it. Everyone else left. Three words you need for the rest of this story:
Trading with borrowed money. At 10×, a $500 deposit controls a $5,000 position, so gains and losses are ten times bigger.
The price at which the exchange closes your position because the losses have used up your deposit. The one number a trader never wants to be surprised by.
The desktop screen professionals trade on: a chart, an order book of live bids and asks, and an order form with every setting visible at once.
Chart, Buy / Long, margin mode, leverage, price and amount, time in force, Place Order.
All visible at once, none explained, and Place Order greyed out until every one is right.
It opens on a candlestick chart. No balance, no positions, no list of markets.
The launch app, step by step.
Simple mode, step by step.
There were no analytics on the launch app yet, so the problem was established three ways: I audited the launch app in Figma and counted every control and step; the founder set four success metrics (trades completed, fewer steps and errors, activation and retention, cost and risk understood); and our own traders kept using the same two words when asked what they wanted, flexibility and control. Published studies on retail traders and crypto onboarding backed each call, and they are cited next to the screen they changed.
Four steps from open to order, down from seven. Risk shown before you confirm, on every order. Nothing taken from professionals: every terminal feature stays, one layer down, and the full terminal stays one switch away.
Hide the terminal. Don’t remove it.
Follow one trader through a day. Each moment: what broke, what we built, why, and what it cost.
Opening the app, placing a first trade, protecting it, and coming back tomorrow. For each, the old flow and the new one, the evidence that decided it, and the trade-off we accepted.
A candlestick chart with fourteen drawing tools, made for someone who already knows what they are looking for. No balance, no positions, no list of markets.
One number you understand, a chart under it, two rows (positions, USDC), then markets with sparklines and a floating Trade / Deposit bar.
Nobody opens a trading app to fill in a form. Our traders opened it to answer two questions, how am I doing and what is moving, and the launch app answered neither. A 2026 study of retail traders found the same: people choose by simplicity (WSU). So the first screen became one number, with the markets under it.
A professional who opens the app to read the chart is now one tap further from it. We accepted it because that trader has Advanced mode, which still opens on the chart; a first-timer has nowhere else to go. Result: the first screen now shows your balance, today’s change, your positions and the markets, where the launch app showed none of them.
Every setting visible whether you need it or not, and Place Order disabled without saying why.
Type what you want to risk, see the size in BTC, tap Preview. Leverage, margin mode, time-in-force and TP/SL are chips a pro can open and a first-timer can ignore.
The audit said a first order asked eleven questions and then greyed out the button. Industry data says what happens next: 60–75% of people quit a crypto app before their first transaction (Swapper). Every question a form asks costs a trader, so we ask one: how much, in dollars. The rest became chips.
Defaults now carry weight: a first-timer trades at 10×, cross margin, good-till-cancelled unless they open a chip. We accepted it because the terminal already used those defaults, and we put the leverage chip and the liquidation price on the sheet itself, under the amount. A professional who changes settings on every trade pays two extra taps. Result: 4 steps from opening the app to an order, down from 7, and no button is ever greyed out.
Tick a box and the form grew: TP price, SL price, gain, loss. Fees, deposit and the liquidation price were not on the screen at all.
One step before the fill: position size, entry, fees, the liquidation price in the terminal’s mark yellow, and a TP/SL row. Then Buy now.
The words our traders used most were flexibility and control, and control starts with knowing what a trade costs and where it dies. The launch app told you neither before the fill. The founder’s fourth metric said the same thing in other words: cost and risk understood. So one preview step sits between the amount and the order, and the app sets an alert at 5% from liquidation without being asked.
Every trade now takes one more tap, which a scalper placing twenty orders an hour will feel. We accepted it for Simple mode only; Advanced mode keeps the one-tap terminal form. Result: the liquidation price is on 100% of previews, where the launch app showed it on none.
The launch app opened on the same chart every time. Unless you had an order to place, there was no change in your own numbers to check and nothing social to see.
The leaderboard opens on your rank and the distance to the next one, then a plain list of top traders. With today’s change on Home, it is a reason to open the app without a trade.
A terminal only rewards you when you trade. The founder’s brief named activation and retention as two of four metrics, and the market explained why: perp DEXs now take 26% of crypto futures volume (BEX), the professionals are already taken, and growth means people who open the app on a day with no trade in mind.
A leaderboard can push people to trade more than they should. So it ranks volume and equity, never profit, and the share sheet shows a rank, not a return. We accepted a less exciting board for one that does not reward recklessness. Result: this is the moment we can measure least directly; the +15% volume lift below includes it, and the next step is to instrument retention on its own.
Nine screens, and where every terminal feature went.
It starts inside the terminal: Simple mode is a toggle at the top of the More drawer, the same drawer that already held Testnet, API and Language. Flip it and the same account opens on four tabs, one trade flow, and the terminal one switch away.
One design system for both modes, so switching never feels like leaving.
Advanced mode (the full terminal) and Simple mode are built on the same design system: the same tokens, type, components and colors. Simple mode changes what is on the screen, never what it looks like. Glass cards, a floating blurred tab bar, pill buttons, tabular digits. Lavender Mist (#A89DE8) is Margin Trade’s brand color, and it is the only color an action can be. Green and red mean long, short, gains and losses; mark yellow means the liquidation price. Nothing else gets a color.
In a trading app, green and red are already spoken for: they mean money made and money lost, on every price, chart and position. A brand color that is green or red would make every button look like a gain or a loss. Lavender belongs to neither side, so a tap target never reads as a market signal. It also holds strong contrast on the near-black surfaces traders stare at for hours, and it is calm, which matters in a product built around risk. That is why it is the brand, and why it is reserved for actions: if it is lavender, you can tap it.
Traders switch modes, and the founder asked for one product, not two. With one design system behind Advanced and Simple mode, a trader recognizes the app on either side of the toggle: the same Buy / Long green, the same lavender button, the same digits in the same places. Engineers built one component set instead of two, so a fix or a new token lands in both modes at once.
Blur costs GPU on older phones, and a dark surface loses contrast in sunlight. We accepted both for the launch and kept two fallbacks in the file: flat surfaces when blur is unavailable, and a light variant for a later release.
Three months in: more trading, and traders who feel in control.
Three months after launch: +15% trading volume, 7 → 4 steps to an order, 11 → 1 questions.
More money traded on the exchange than before Simple mode shipped. The team’s own measurement.
Counted on the screens, launch app against Simple mode.
Eleven settings became one: how much. Nothing removed from the terminal.
Three months after launch, the team’s own measurement.
Flexibility and control, the same two words they asked for: a fast path for the everyday trade, the full terminal one switch away.
The published numbers shaped the brief and are cited where they changed a screen. The +15% is the team’s measurement and includes everything that shipped in the quarter. Margin Trade overall clears $1M+ daily across 14+ countries.
The challenges, and what I’ll do next.
Every terminal feature had an owner who would notice if it went missing. The fear of ‘dumbing down’ only went away once each one had a home one layer down.
A professional and a first-timer open the same app. Serving both on one screen is how the launch app reached eleven controls; the answer was two modes on one token set.
The launch app had no funnel data, so the brief had to stand on counted controls, traders’ words and published studies. A number you can count aligned the room faster than any adjective.
Time to first order, order completion and error rate per cohort, so the next lift is attributable to a screen, not a quarter.
Hand them a phone with Simple mode and a $20 position to open. Count the hesitations.
Learn which mode a trader uses per market and open there, so the switch becomes invisible.
