Case Study 01 · Desktop Web · Web3

Professional DEX Trading

Designing a high-control trading experience for experienced Web3 traders.

Users

Professional and active Web3 traders

Platform

Desktop web

Role

Product strategy, research, UX and UI design

Focus

Market data, order execution and trading efficiency

SECTION 01 · EXECUTIVE SUMMARY

Professional traders were leaving value on centralized exchanges because decentralized alternatives lacked control. We defined and shipped a desktop DEX experience built around advanced order execution, dense market data, and transaction confidence — validated with traders at every stage.

Professional traders were leaving value on centralized exchanges because decentralized alternatives lacked control. We defined and shipped a desktop DEX experience built around advanced order execution, dense market data, and transaction confidence — validated with traders at every stage.

Professional traders were leaving value on centralized exchanges because decentralized alternatives lacked control. We defined and shipped a desktop DEX experience built around advanced order execution, dense market data, and transaction confidence — validated with traders at every stage.

Business and market context

Business and market context

Competitive analysis across leading CEX and DEX platforms showed a consistent pattern: decentralized venues had caught up on liquidity but not on trading tools. Professional traders kept a CEX open for execution and a DEX open for assets — a split workflow with real costs. The business opportunity was to close that gap and capture high-volume traders directly.

Product opportunity

Product opportunity

Product data and trader interviews pointed to a specific wedge: advanced order execution. Traders would tolerate a learning curve for self-custody, but not the loss of limit orders, position context, and execution feedback. We framed the product around one promise — CEX-grade control with DEX-grade ownership.

Target users and segmentation

Target users and segmentation

Segmentation separated three groups by trading frequency and tooling expectations. We deliberately prioritized the professional segment — the smallest group, but the one defining credibility and volume.

Professional traders — daily, multi-position, tooling-critical

Primary

Active traders — weekly, momentum-driven, chart-first

Secondary

Occasional swappers — served by default flows, not the MVP focus

Out of scope

SECTION 02 · RESEARCH

How research changed the product.

Twelve interviews with professional traders, a competitive teardown of eight venues, and behavioral data from the existing swap product. Each key insight below traces from observation to validated product change.

Insight 01 — “I won’t trade blind. Market orders alone are a dealbreaker.”

Insight 01 — “I won’t trade blind. Market orders alone are a dealbreaker.”

Observation

10 of 12 traders named limit orders as the single feature keeping them on a CEX.

Why it mattered

It reframed order types from an advanced feature into the core switching trigger.

Product decision

Limit orders moved into the MVP; simple swap was demoted to a secondary mode.

Design response

A persistent order ticket with market/limit tabs, price presets, and slippage control.

Validation result

In prototype tests, all participants placed a limit order unaided in under 30 seconds.

Insight 02 — Simplified views read as amateur, not friendly.

Insight 02 — Simplified views read as amateur, not friendly.

Observation

Traders judged low-density screens as untrustworthy for serious execution.

Why it mattered

Our early principle of “simplify everything” was wrong for this segment.

Product decision

New principle: organize complexity, don’t remove it. Density became a feature.

Design response

A three-tier market-data hierarchy: glanceable price, scannable book, drill-down detail.

Validation result

Expert reviews rated the dense layout more “professional” and faster to scan than the simplified variant.

Insight 03 — On-chain uncertainty breaks trader confidence.

Insight 03 — On-chain uncertainty breaks trader confidence.

Observation

Traders refreshed block explorers mid-trade because pending states were opaque.

Why it mattered

Execution anxiety — not fees — was the biggest emotional barrier to DEX adoption.

Product decision

Transaction lifecycle became a first-class product surface, co-designed with engineering around confirmation constraints.

Design response

Inline status stages — submitted, confirmed, settled — with explicit failure and retry states.

Validation result

Explorer checks during test trades dropped to near zero; confidence ratings rose in follow-up sessions.

SECTION 03 · RESEARCH PLAN

Four research workstreams feeding one decision model

Rather than a generic process diagram, research was structured as four parallel workstreams. Each carried its own question, method, and output — and every output was tied directly to a product decision.

01

Market & competitor analysis

CEX spot · DEX swaps · Aggregators · Meme terminals · Perps · Wallet trading

Research question

Where do existing products sit between simplicity and professional control?

Method

Feature teardown across 6 product categories

Output

Positioning map + control/context gaps

Product decision

Target high-context, controlled-complexity space

02

Trader interviews

Trading routines · Position sizing · Execution preferences · Risk · Failed-tx recovery

Research question

How do active and professional traders actually decide and execute?

Method

Semi-structured interviews with active traders

Output

Behavioral patterns + recurring pain points

Product decision

Prioritize repeatable sizing and direction clarity

03

Contextual observation

Chart ↔ panel movement · Preset usage · Decision sequence · Ignored vs re-checked data

Research question

What does the real order-entry sequence look like in context?

Method

Observed live trading sessions

Output

Attention flow + interaction friction points

Product decision

Keep the order panel persistent beside the chart

04

Prototype validation

Direction · Market vs Limit · Strategy discoverability · Cost transparency · Error recovery

Research question

Do the proposed patterns hold up under realistic tasks?

Method

Task-based prototype testing

Output

Comprehension + recovery evidence

Product decision

Surface execution conditions before the CTA

SECTION 04 · USER SEGMENTATION

Designing for depth, not the lowest common denominator

Three trader levels emerged. The interface deliberately targets the two with the highest execution demands — expressed here through scale and density rather than identical persona cards.

SECONDARY

Occasional Trader

Trades infrequently

Primarily uses simple swaps

Focuses on price and final received amount

Limited familiarity with execution settings

PRIMARY AUDIENCE

Active Trader

Trades frequently

Uses watchlists, alerts, presets, and market data

Values speed and transaction reliability

Sometimes adjusts slippage or execution settings

PRIMARY AUDIENCE

Professional Trader

Monitors multiple information sources

Uses repeatable position sizes

Requires granular execution control

Actively manages risk, fees, and market timing

PRODUCT DECISION

The interface should not be reduced to a basic wallet swap. It needed to preserve advanced controls while making repeated actions faster.

Product scope and feature prioritization

Product scope and feature prioritization

I facilitated scope workshops with product and engineering, scoring features against the switching-trigger evidence. The MVP shipped market and limit orders, token discovery, positions, and transaction status. Stop orders, portfolio analytics, and API access were sequenced behind validated demand.

MVP · Market + limit orders

MVP · Token discovery

MVP · Positions + status

v2 · Stop orders

v2 · Portfolio analytics

Information architecture

Information architecture

One trading workspace instead of paged flows: market data on the left, chart at center, order ticket pinned right, positions and open orders docked below. Every element answers one of three trader questions — what is the market doing, what do I want to do, what is happening with my orders.

Primary workflow

Discover

Analyze

Configure

Review

Execute

Monitor

Design exploration, validation and iteration

Design exploration, validation and iteration

Three order-entry structures were prototyped and tested against real trading tasks. The pinned-ticket model won on speed and error rate. Two usability rounds reshaped slippage controls and confirmation review; engineering constraints around quote refresh timing drove a co-designed price-staleness indicator. Edge cases — failed transactions, partial fills, stale quotes, disconnected wallets — were specified as system states in the design system, not left to implementation.

Product outcome and reflection

Product outcome and reflection

The workspace shipped with limit orders adopted as the dominant order type among the target segment, and the design system scaled to two further market types without redesign. Next: deeper position analytics and the v2 advanced order suite. The lasting lesson — for professional tools, trust is built through visible control, not visual simplicity.

SECTION 05 · INTERACTION DESIGN

Market and Limit are different structures, not one extra field

Buy and Sell set the first-level context. Market and Limit act as a lighter second-level control. Rather than bolting a price field onto the Market form, Limit is designed as its own layout — comparing them side by side keeps each model coherent.

MARKET ORDER

SOLA / USDC

BUY

SELL

Market

Limit

AMOUNT

1.00

BNB ▾

0.5

1

2

5

Custom

Estimated receive

≈ 248.6 SOLA

Network fee

0.0004 BNB

Price impact

0.12%

BUY SOLA

LIMIT ORDER

SOLA / USDC

BUY

SELL

Market

Limit

LIMIT PRICE

0.0421

USDC

AMOUNT

250

SOLA ▾

Total payment

10.53 USDC

Expiration

7 days

Status

Open until filled

PLACE LIMIT BUY

SECTION 06 · VALIDATION

Testing specific hypotheses, not general impressions

Usability sessions were framed around comprehension and recovery questions the design had to answer — each mapped to a concrete task and a design change.

RESEARCH QUESTIONS

Can users immediately identify Buy or Sell?

Do users understand the unit of quick presets?

Can users tell whether Advanced Strategy is enabled?

Can users complete an order without losing chart context?

Can users find fees and slippage before submission?

Do users understand how the form changes between Market and Limit?

Can users respond correctly when the price changes?

EXAMPLE TASKS

TASK 1

“Use 1 BNB to buy the token with a Market order.”

WE MEASURED

Direction recognition

Preset usage

Strategy awareness

Order summary comprehension

TASK 2

“Create a Limit Buy at a specified target price.”

WE MEASURED

Limit-price comprehension

Total-payment understanding

Expectation of delayed execution

Ability to locate open orders

TASK 3

“Adjust slippage and enable transaction protection.”

WE MEASURED

Advanced-setting discoverability

Understanding of active preset

Ability to return to defaults

Awareness of execution impact

HYPOTHESIS → EVIDENCE → DESIGN CHANGE

HYPOTHESIS

A large first-level Buy/Sell selector prevents direction errors.

EVIDENCE

Participants selected direction without hesitation and self-corrected before submission.

DESIGN CHANGE

Kept direction as the dominant, redundant-cue control above order type.

HYPOTHESIS

Preset numbers read as amounts, not percentages, when the unit stays visible.

EVIDENCE

Participants read preset values as amounts on the first attempt once the unit chip stayed pinned beside the field.

DESIGN CHANGE

Locked the asset unit beside the input and presets.

HYPOTHESIS

A separate enable vs. edit model clarifies Advanced Strategy state.

EVIDENCE

Users could state whether a strategy was active and how to change it.

DESIGN CHANGE

Split the switch (on/off) from the strategy summary and edit entry.

SECTION 07 · OUTCOME

Measurable clarity, without stripping away control

The redesign held its ground with experienced traders while removing the friction that slowed first orders — validated in testing and reflected in the shipped experience.

IMPACT SNAPSHOT

-38%

Time to first completed order in moderated tasks

+29%

Correct direction (Buy/Sell) identification on first attempt

0

Critical mis-submissions across final validation round

9/10

Traders rated the order panel as “professional-grade”

WHAT SHIPPED

Redundant direction cues

Buy/Sell stays the dominant, always-visible control above order type.

Locked asset unit

The unit chip sits beside inputs and presets so amounts never read ambiguously.

Persistent chart context

Order entry no longer covers the chart — traders confirm without losing the market view.

Separated strategy state

Advanced Strategy exposes on/off independently from edit, clarifying what is active.

Next case study

Cross-platform Prediction Market ↗

Cross-platform Prediction Market ↗

Cross-platform Prediction Market ↗