In the high-stakes world of modern finance, firms have long been juggling the competing imperatives of efficiency, compliance, and cost. Nowhere is this more evident than in the architecture of their trade surveillance programmes. For years, a split system model where off-the-shelf solutions monitor listed markets while in-house builds handle the idiosyncrasies of Over-the-Counter (OTC) products was seen as a pragmatic compromise. But as markets evolve and regulators sharpen their focus, the cracks in this fragmented approach are beginning to widen. The result is a mounting consensus: split surveillance is no longer suitable.
This bifurcated architecture has its roots in the age-old build-versus-buy debate. When it came to surveilling vanilla products, cash equities, futures, and other liquid listed instruments, buying vendor tools was an easy win. These systems, often calibrated for high-volume exchange trading, promised quick implementation and cost containment. But they were poorly suited to more complex or bespoke asset classes. OTC derivatives, structured credit, and fixed income instruments priced off illiquid curves all demanded more tailored, often internally developed controls.
Thus, a hybrid model emerged: buy where simplicity allows, build where complexity demands. What followed was a gradual descent into a labyrinth of siloed platforms, disconnected datasets, and inconsistent controls. Each system did its job in isolation. Together, however, they created a fragmented framework prone to duplication, inconsistency, and blind spots.
The hidden cost of pragmatism
At first glance, this model offered short-term gains. Internal builds afforded bespoke flexibility, while off-the-shelf tools ticked regulatory boxes without straining the budget. But what began as a tactical compromise has ossified into a strategic liability. Firms now find themselves shackled to systems that neither communicate nor scale. Surveillance teams must reconcile divergent methodologies, maintain overlapping infrastructure, and retrain analysts for each system's idiosyncrasies.
More troubling still is the threat to data governance. In an era of increasing regulatory scrutiny, the capacity to explain, audit, and defend one’s surveillance framework is no longer optional, it is existential. When JP Morgan was fined $450 million in 2021 for surveillance and data governance failings, it sent a clear message: having controls is insufficient; they must be coherent, integrated, and demonstrably effective.
Regulators today demand more than coverage, they demand cohesion. The old "control-by-control" mentality, in which each risk was addressed independently, is displaced by a new standard: holistic oversight. Surveillance data must be traceable, reconcilable, and complete. In firms with fragmented systems, these qualities are often aspirational rather than actual. Data lineage becomes murky; gaps in feed coverage persist unnoticed; and inconsistencies arise between systems meant to monitor the same behaviour. The risk is not just operational failure, it is regulatory non-compliance.
A blunt instrument in a sophisticated market
The greatest irony of the split model is that, in trying to optimise for complexity, it fails to capture it. Market abuse, after all, does not respect product silos. It operates across asset classes, jurisdictions, and trading venues. Traders hedge OTC exposures with listed futures. They construct multi-leg strategies that are benign in parts but abusive in sum.
A surveillance architecture that analyses trades in isolation: one order at a time, one trade at a time, one venue at a time, misses these subtleties. False positives proliferate: a legitimate hedge looks suspicious when stripped of context. Conversely, actual misconduct goes undetected, its intent hidden in the seams between systems.
The operational consequences are equally stark. Each platform requires its own support, tuning, and subject-matter experts. Escalations become slower, development cycles longer, and institutional memory narrower. When regulatory demands shift, as they invariably do, firms must retrofit new controls into multiple systems simultaneously. This duplication slows the firm’s ability to respond and inflates costs without improving outcomes.
The challenge of coordination
Beyond inefficiency lies an even more pressing concern: the absence of a unified view. In a typical split environment, rolling out a new control requires coordination across disparate technology stacks, each with its own data model and release cycle. A surveillance rule introduced on one platform may take weeks or months to replicate elsewhere. Worse, inconsistencies in how data is ingested or interpreted can undermine the integrity of the rule itself. A trader’s activity may appear anomalous in one system and unremarkable in another, not because of what was done, but because of how it was seen.
This problem is more than technical; it is conceptual. Surveillance should mirror how risk manifests: interconnected, dynamic, and contextual. Instead, firms are forced to act as translators, reconciling fragmented outputs into something resembling a coherent picture. That translation process is slow, manual, and fraught with error.
Towards an integrated future
What is needed is a step-change in design philosophy. Rather than retrofit legacy systems to talk to each other, firms should aim to consolidate surveillance onto a unified platform that interprets trading behaviour holistically and adapts in real time.
A modern surveillance system must connect dots across desks, products, and venues. The aim is not simply to monitor activity, but to understand it, to see risk in context, not in isolation.
Key to this vision is adaptability. Static rules, hard-coded thresholds, and rigid workflows are increasingly out of step with the fluid nature of today’s markets. Instead, best-in-class platforms leverage real-time data and dynamic analytics to recalibrate risk indicators as conditions change. The result is fewer false positives, more precise alerts, and a better allocation of investigative resources.
The operational benefits are equally compelling. Training requirements fall as staff learn a single system. Development accelerates, since new models need only be built once. Escalation workflows are streamlined, and controls can be benchmarked against centralised risk measures such as P&L or Value-at-Risk. Data lineage improves, allowing firms to trace anomalies back to source and demonstrate end-to-end governance.
From compliance to competitive advantage
Seen in this light, integrated surveillance is not merely a compliance obligation, it is a source of strategic advantage. Firms that embrace a unified architecture gain the agility to respond quickly to emerging risks, the insight to detect sophisticated manipulation, and the resilience to withstand regulatory scrutiny. They move from defensive compliance to proactive control.
Indeed, in a world where surveillance failures carry reputational and financial costs, robust oversight becomes a differentiator. Clients, counterparties, and regulators are interested in dealing with firms with demonstrably effective controls. Surveillance, once a cost centre, becomes a pillar of trust.
The shift will not be easy. Legacy systems, entrenched interests, and short-term budget pressures will all conspire to preserve the status quo. But the direction of travel is clear. Fragmented surveillance is no longer defensible. The firms that recognise this and act will be better positioned to survive and thrive in tomorrow's markets.
