Restructuring research and design at ADSS

Context

ADSS is an Abu Dhabi-based financial services firm offering trading platforms across regulated MENA markets. ADQ-owned, regulated by the UAE Capital Market Authority since 2010, with subsidiary operations in London, Hong Kong, and Singapore. Four products serving institutional, private, and retail clients across forex, equities, indices, commodities, and bonds.

When I joined as Head of User Experience in October 2022, reporting to the Chief Product Officer, the UX function operated as a service layer rather than a strategic capability. Each of the four trading products ran isolated user research, with no shared methodology, no shared insight repository, and significant duplication across studies asking overlapping questions. Time-to-insight on a typical study sat at three to four weeks. Findings landed in slide decks and rarely informed roadmap decisions after the launch they were originally commissioned for.

The CPO’s mandate was clear: turn UX from a delivery function into a strategic capability that could measurably influence product decisions across all four platforms. The work that followed restructured how research and design operated together, what they shipped, and how the rest of the organisation consumed their output.

The problem

Research costs were rising faster than the value the function returned. A representative quarter looked like this: 12 to 16 separate studies commissioned across the four products, of which roughly half were asking variations of the same question: onboarding friction, KYC drop-off, pricing comprehension, platform switching behaviour. Each was budgeted, recruited, and synthesised independently. The same insight surfaced four times, three weeks apart, paid for four times.

Worse, very little of it converted into product decisions. Findings sat in 40-slide decks distributed by email. Three months after a study landed, neither the PM nor the designer who commissioned it could reliably summarise what it had concluded. New PMs joining a product would commission research on questions that had already been answered. Vendor invoices kept coming.

The structural diagnosis was straightforward: the function was set up to produce research, not to retain and apply insight. Until that flipped, no amount of additional headcount or budget would compound. The CPO needed evidence that the function could operate as a strategic capability before the next budget cycle.

My role

I joined ADSS reporting directly to the Chief Product Officer, with end-to-end ownership of the UX function across four products: the proprietary ADSS trading platform, MetaTrader 4 / 5 customisation, the institutional desk experience, and a retail-facing onboarding flow under development at the time.

The mandate covered both research and product design, unusual at this seniority level, where the two usually split into separate director-level functions. Owning both is what made the restructure possible: most of the duplication had been hiding in the gap between “what we ought to study” and “what we ought to design,” and only one function owning both could close it. That let me redesign the operating model end-to-end.

I worked alongside KPMG (operating model and digital transformation), Adaptive Financial Consulting (cloud-native trading platform technology), and Turing (engineering integration support) on platform partnerships during this period. On UX strategy it was just the CPO and me; everything else was delegated.

Approach

I structured the restructure across three phases over six months. The sequence mattered more than any individual move.

Phase one, repository before method. Before optimising how studies got run, I built a centralised insight repository, indexed by user intent, product surface, and decision type, and tagged to the specific roadmap decision each insight had informed or could inform. The point was to file insight by the decision it served, not the study that produced it. This took six weeks and produced no new research output. I pushed back against pressure to start commissioning studies immediately. The reasoning: optimising research throughput against a function that didn’t retain its own findings would compound the original problem, not solve it.

Phase two, shared usability framework across all four products. With the repository in place, I built a single usability testing framework that all four platforms used. Same recruitment criteria, same task taxonomy, same scoring rubric. The trade-off was real and worth naming: less platform-specific tailoring on individual studies. The gain was cross-platform comparability, shared tooling and vendor pricing, and the ability to spot insights that applied across the portfolio rather than to one product. Within three months we had benchmark data for every major platform flow that we could reuse against any new feature decision.

Phase three, embedded research in product strategy decisions. With the framework operating and the repository compounding, I changed how research participated in roadmap conversations. We started co-moderating executive strategy workshops, and we did it as a research function: we came into those rooms with a defensible cross-product evidence base rather than study summaries. The CPO and I worked the framing of the function’s value upward, into the C-suite TAM expansion conversations that were happening in parallel.

The pattern across all three phases: optimise the operating model first, then improve specific outputs. Most UX restructures invert this and end up shipping faster studies that nobody acts on.

Key decisions

01 — Repository before method

Centralising insight came before optimising how research got run. The pushback was internal: PMs wanted faster turnaround on the studies they had already scoped. I held the line for six weeks because optimising velocity against a function that lost its own findings would have compounded the underlying problem, not solved it. The cost was visible, slower study delivery for a quarter, and the gain only showed up in months four through six, when the same insight stopped being commissioned twice.

02 — One framework across all four products

Imposing a shared usability framework across four products with different user bases (institutional, retail, MetaTrader power users, new-to-market retail) was a real trade-off. Each product team had reasonable arguments for bespoke methodology. I chose comparability over tailoring because the strategic question, which platform actually serves which user best, and where do they overlap?, couldn’t be answered without a common measurement frame. We accepted some loss of fidelity on individual studies in exchange for cross-portfolio insight that the function had never produced before.

03 — Co-moderating executive strategy workshops

Most UX functions earn their seat at the strategy table by building credibility study-by-study, year over year. With a CPO-level mandate already in place, I made the function visible at the strategy layer immediately rather than working up. We co-moderated regulated TAM expansion workshops within the first two quarters. The risk was real: research speaking into strategy conversations before it had built a deep evidence base could read as overreach. The mitigation was the repository, by the time we were in those rooms, we had a defensible cross-product baseline behind every claim.

Outcome

The quantitative outcomes:

$5M annual savings, through eliminated duplication, vendor consolidation under the shared framework, and a 20% reduction in average per-study cost.

~35% faster time-to-market for new features, driven by integrated customer journey mapping and JTBD frameworks that let product teams reuse existing insight rather than commissioning fresh research for every roadmap question.

30+ research studies and design sprints delivered across the four products during the 14 months I led the function, but the more meaningful number is the ratio: the same headcount that had been delivering 12-16 studies per quarter pre-restructure was producing roughly the same volume post-restructure with cross-product impact rather than per-product impact.

The bigger change was qualitative. By the second half of my tenure, the function was in TAM-expansion strategy conversations with the CPO, regulators, and partner technology providers (Adaptive, Turing), consulted before a decision was scoped rather than after it was made.

Worth being straight about the timescale: at 14 months, some downstream effects of the restructure were still in early flight when I moved on. The repository compounded for as long as it was maintained; the framework only worked while a single function owned both research and design. Both are choices my successors made about whether to preserve.

What I’d do differently

Two product bets I’d make harder next time.

I’d have pushed harder for social trading. ADSS’s user research consistently showed that new-to-market retail clients trusted other traders’ decisions more than their own, particularly during the early months of trading. We treated this as a behavioural insight to design around (better defaults, better education, conservative recommended positions). It was actually a product opportunity. A social-trading layer, copy-trading from verified ADSS clients, social proof on position sizing, community-vetted strategy templates, would have addressed the core trust gap directly rather than mitigating it. eToro built a $10B+ company on this exact insight. We had the data to see it; we didn’t push hard enough on the product implication.

I’d have designed an explicit beginner-to-expert experience curve. The trading platforms operated on a single UI tier, same density, same options, same complexity for a 20-year forex veteran and a first-time retail client. We discussed dual-mode interfaces (Starter vs Expert) and validated the demand in research. We didn’t ship it because the engineering cost of forking the UI was real and the regulatory implications of a “simplified” trading interface needed careful navigation. Both reasons were valid. Both reasons would have been worth solving. Forcing every retail user into the same UI as the institutional desk meant we measured retail attrition as a UX problem when it was a product-architecture problem.

The most valuable insight from regulated trading research is about who would trade at all if the product met them where they are, not how to serve existing traders better. That market-expansion thesis is where the function’s second year should have gone, and both of the regrets above point straight at it. I led the function for fourteen months; with another twelve, that’s the bet I’d have placed.

Artifacts