Lead designer for the entire LogixTrader product surface at ACY — the web trading platform plus intellogix, the admin portal that governs every symbol, every account, and every IB white-label tier behind it. Four audiences served by one platform: retail traders, introducing brokers, small hedge funds, and institutional integrators embedding our quote/chart widgets. The thesis: density done right is faster than minimalism done wrong — and the back-office is where senior product design actually lives.
My role: Lead Product Designer · end-to-end UX/UI · interaction spec · design-system governance.
Team: 4 Frontend + 1 Backend (real-time data feeds) · Product Manager (competitive analysis, feature roadmap) · Legal & Compliance Officer (order validation rules, risk disclosures) · 45 beta traders (usability testing).
Context: Cross-functional team inside the ACY ecosystem. Worked within established brand guidelines while building platform-specific interaction patterns for execution and risk surfaces.
The Problem
MT4/5 isn't just a platform — it's a network effect. Brokers offer it because traders demand it; traders demand it because every broker offers it. Breaking that loop required understanding three forces working against any challenger.
Constraint 01
Active forex traders operate in 200–800ms windows. A cluttered interface doesn't just frustrate — it directly costs them money. Every design decision had to pass: does this help or hinder a split-second call?
Constraint 02
Professional traders have spent years building unconscious motor patterns around MetaTrader. Any alternative that disrupts those patterns reads as "slower" regardless of actual performance. We mapped MT4's most-used shortcuts and intentionally preserved them.
Constraint 03
A trading interface must simultaneously surface price feeds, open positions, account metrics, order forms, charts. Too little = workflow gaps. Too much = decision paralysis. The challenge isn't reduction — it's hierarchy.
Architecture Decision
Three paths forward. Each had a believable case for it. The chosen one — hybrid — got us live in under 8 months without compromising UX or shouldering the operational risk of replacing battle-tested execution infrastructure.
Build proprietary matching engine, data feeds, charting library, UI from zero. Complete stack ownership.
Why not: 2–3 year timeline, unproven execution infrastructure. MetaQuotes has spent 20+ years on matching-engine reliability. Replicating that from zero is unacceptable operational risk for a regulated broker handling real client funds.
Apply custom CSS/UI layer on MetaTrader's existing web terminal. Fastest path to market.
Why not: Locked to MetaQuotes' component architecture and update cycle. No room for contextual data surfacing, custom shortcuts, or density-aware layering. No competitive moat — perpetually constrained.
Fully custom proprietary UX, connected to MT's proven execution backend via API bridge. Design freedom + battle-tested reliability.
Why this: Full design control over the experience — progressive disclosure, custom layouts, keyboard-first workflows, contextual data surfacing — with MetaTrader handling execution. Live in under 8 months. demo.acylogixtrader.com runs on this architecture today.
What Doesn't Work
I designed for the portfolio I wanted to show, not the users I was serving. Removed keyboard shortcuts. Hid quick-access panels. Collapsed data density for whitespace. The brutal feedback below is what made V2 possible — and what made me a better designer.
| Metric | MetaTrader 4 (Baseline) |
LogixTrader V1 (Too minimal) |
LogixTrader V2 (Hierarchical density) |
|---|---|---|---|
| Task failure (rapid execution) | 12% | 82% | 4% |
| Avg. time-to-trade | 2.1s | +3.5s | 1.8s |
| Projected abandonment | Baseline | 42% | <5% |
| User satisfaction (1–5) | 3.8 | 1.4 | 4.7 |
Directional signals on n=45 cohort. Order-placement isolated test on n=15 paired within-subjects (8.2s → 2.9s, Cohen's d = 2.47).
"Looks nice. Can't trade on it." — direct quote from 8 separate testers on V1. The hardest feedback I've received in my career, and the one that taught me the most.
Beta tester · n=45 cohort · ACY usability lab
The lesson that carried into every project since: Minimalism ≠ clarity in data-dense professional environments. Professional users don't need less information — they need better hierarchy. The design challenge is organisation, not reduction. Validated downstream on Finlogix, ACY Connect, TradingCup.
Outcomes
Numbers below split design-attributable from team-attributable. Live demo above is the strongest evidence — the platform you can click on right now is the artefact this case study is about.
Business
Reduces dependency on third-party licensing. Custom feature development on ACY's own timeline. Differentiator in client acquisition.
UX
Power traders get density when needed; newer traders get clarity by default. Both audiences served without mode switching.
Engineering
Rapid UI iteration without touching execution infrastructure. Design-system components shared with the broader ACY ecosystem accelerate cross-product consistency.
Trader
One-click execution + persistent position management + keyboard-first design. Faster than MT4's baseline on the same hardware, validated head-to-head.
Validated Principles
Each one was tested under fire on this project. Each one carries a counterargument I had to talk past in a real meeting.
Principle 01
The instinct to "simplify" trading interfaces by removing data is wrong. Professionals need dense information; the design challenge is making 100 data points scannable in under a second through size, weight, colour, grouping.
Principle 02
Designing against an incumbent with 20+ years of habit formation, "objectively better" isn't enough. The platform had to feel familiar enough to lower switching cost while different enough to justify the switch — that meant preserving the most-used MT4 shortcuts deliberately.
Principle 03
A custom frontend on proven execution infrastructure was the single most important enabler of good UX. Without it, we'd have been constrained to the incumbent's layout paradigm — no progressive disclosure, no density-aware layering, no shipped product in 8 months.
Design Artifacts · Figma
In a trading platform, a validation rule is a UX decision; an interaction state is a visual system. Each artefact below is one place those three layers collapse into one.
Artefact 01
Highest-frequency surface — traders interact 30–80× per day. Every state (default, hover, press, search, result) mapped and rebuilt for friction reduction without losing density. Old → new comparison shows typography scale, colour tokens, and spatial hierarchy refined deliberately.
Artefact 02
A Take Profit inside the spread zone = rejected order. An invalid Stop Loss = unprotected exposure. Prevent errors at input, not at submission. Real-time visual feedback shows traders exactly where constraints exist before they commit.
Artefact 03
Symbol Manager, Watch List, Trading Terminal — three data-dense tables, one interaction grammar. Hover, press, select, checkbox: same behaviour, same affordances, same feedback. A trader's muscle memory works identically across every table on the platform.
Artefact 04
Date/Time pickers required exhaustive state mapping — typing vs. calendar input, every validation error, year/month rollover. Three responsive layout strategies were evaluated against real trader workflows at different viewport widths. New users get progressive complexity; power users skip onboarding.
Why these four artefacts matter: data constraints (spread zones, validation rules), user experience (interaction states, onboarding flows), and visual design (typography scales, colour tokens, spatial hierarchy) are one unified system. In trading-platform design you cannot separate these three — change a validation rule and the UX changes; change the UX and the visual feedback changes. This is what "design" means in this context.
Beyond the Trading Surface
The trader-facing platform you've seen so far is one of three surfaces I designed. Behind it sits intellogix — the admin portal where every symbol's tradability, every account's risk posture, and every IB's white-label customisation is configured. Case studies that show only the chart miss what senior PD work actually is.
Admin Portal · 01
~1,000+ tradable symbols across stocks, forex, CFDs, indices, crypto. Real-time bid/ask quote feed status dots, mass-edit/mass-delete operations, CSV import/export for bulk onboarding, sortable by any column. Every row here is a per-symbol governance config — change the spread, change the trade-access posture, change the precision and the trading frontend reflects it within one tick cycle.
Admin Portal · 02
Click any symbol from the list and you land in a 7-tab detail surface — General, Quote, Trade, Calculation, Swap, Margin, Session. Each tab is a different governance lens on the same instrument. Three of the most consequential tabs shown below.
Admin Portal · 03
15,026 trader accounts · 35,663 open positions · live margin-call queue with equity tracking and margin-level alerting. The left panel surfaces accounts trending toward stop-out so risk officers can intervene before auto-liquidation; the right is a paginated, filter-by-account-or-position-ID stream of the entire open book.
Why the admin portal matters for senior-PD evaluation: the trading frontend is where the muscle memory lives, but every per-symbol decision a trader sees was made in this table by an ops admin or a compliance officer. Designing both surfaces consistently — same interaction grammar, same component library, same escalation paths — is the work that separates a chart designer from a platform designer.
Multi-Audience Design
Same backend, same symbol catalogue, same execution infrastructure. But what each audience needs to see first is fundamentally different. The platform serves all five without mode-switching, because each surface was designed knowing the others existed.
Audience 01
Chart + order panel + watchlist + position bar. Reads price first, decides next, executes within 200–800ms. Doesn't need to know the symbol's currency triplet — just that the order will route. Surface: LogixTrader web platform.
Audience 02
White-labels the platform under their own brand. Configures a curated symbol subset (forex-only, or indices-only, etc.). Tracks commission accrual across their referred trader base. Surface: branded LogixTrader + IB admin slice of intellogix.
Audience 03
Manages 5–20 sub-accounts under one master entity. Needs aggregated P&L, position roll-up, risk overlays per strategy. Routes orders programmatically through FIX 4.4 — UI is for monitoring, not entry. Surface: LogixTrader portfolio view + FIX API endpoint.
Audience 04
Multi-jurisdiction routing (ASIC + ESMA + FCA + ASA) · custom session windows per market · MiFID II Art 27 best-execution reporting (RTS 27 quarterly venue quality, RTS 28 top-5 venues). Surface: full intellogix admin + reporting layer + custom audit trail.
Audience 05
Embeds chart, quote, or order widgets into their own affiliate site / news portal / strategy blog. Requires sandboxed auth (OAuth + partner key + scope), branding controls, and graceful degradation when API rate-limits. Surface: standalone embed.acylogixtrader.com widget endpoints.
What designing for five audiences taught me — and made demo-trading-terminal possible: every audience reads the same data in a different order. The retail trader's screen has the same symbol catalogue as the institutional platform's reporting layer, but the IB sees only their curated subset, the hedge fund sees aggregated positions across sub-accounts, and the widget integrator sees a single instrument in a sandboxed iframe. Designing the platform meant designing all five reading orders simultaneously — symbol logic, trade logic, back-office governance, multi-tenant isolation, and embeddable surface. The discipline carried directly into demo-trading-terminal — that demo is the consumer-facing distillation of every domain insight LogixTrader produced.
Live Demo · Execution Quality Analytics
MiFID II requires systematic best execution reporting. LogixTrader's analytics layer surfaces fill quality across venues, slippage against VWAP benchmark, and latency distribution — so a trader can see not just where their order was routed, but whether that routing decision was optimal given market conditions at submission time.
Refreshes every 3s · Simulates real-time execution feed · All data synthetic
Portfolio thread
LogixTrader sits inside two threads that run across the portfolio.
Thread
Retail → Institutional Translation
What breaks when consumer UX patterns meet regulated institutional contexts — and what replaces them
Thread
Evidence & Verification Discipline
How quantitative claims are sourced, validated, and turned into design decisions