Financial institutions have spent years investing in data platforms, cloud infrastructure and increasingly sophisticated Enterprise Data Management (EDM) technology. Yet many firms still find themselves asking the same question: why does managing data remain so expensive, complex and operationally intensive?
The answer may have less to do with the technology itself and more to do with the operating model around it.
Across the investment management industry, the cost and complexity of market data continues to attract attention. Firms are under increasing pressure to demonstrate that they are getting value from their data suppliers, rationalizing overlapping vendors and making better use of the infrastructure they already have.
The regulatory conversation is evolving too. The FCA’s ongoing focus on wholesale data and the cost, accessibility and effectiveness of market data is another indication that the economics of data are becoming a more important industry conversation.
But there is a bigger question that firms should be asking: What is the true cost of owning and operating enterprise data?
The cost isn’t just the data license. When organizations talk about the cost of data, the conversation often starts and ends with vendor fees. But the real cost sits across the entire lifecycle.
Data needs to be sourced, ingested, normalized, validated, mastered, reconciled, enriched, remediated, stored and distributed. Exceptions need to be investigated. Corporate actions need to be interpreted. Pricing discrepancies need to be resolved. Reference data needs to be maintained. Vendor formats change. New instruments arrive. Requirements evolve.
And someone has to operate all of this, often across multiple regions and time zones. The result is an operating model that can be considerably more expensive than the headline technology or data costs suggest.
In fact, EDM is structurally expensive in-house because the workload is exception-led, time-sensitive and coverage-dependent. Firms aren’t staffing simply for average volumes; they are staffing to make sure exceptions are resolved within critical market, valuation and reporting windows.
Why hasn’t more technology solved the problem?
This is where I think the industry needs to challenge some conventional thinking.
Technology has undoubtedly transformed data management. Automation, cloud platforms, UDP’s (e.g. Databricks), APIs and increasingly AI-enabled workflows are changing what is possible. But putting better technology on top of an inefficient operating model doesn’t necessarily solve the underlying problem.
A firm can have a sophisticated EDM platform and still maintain duplicated regional teams, bespoke integrations, fragmented workflows and significant manual intervention.
In other words: You can automate the process without fundamentally changing the economics of the process. That distinction matters.
The hidden cost of the operating model
There are four structural costs that are particularly easy to underestimate.
First, global coverage. Data operations need to align with trading calendars, valuation windows and corporate-action deadlines across regions. That can create a minimum level of permanent staffing even when average workloads don’t justify it.
Second, peak capacity. Data workloads aren’t flat. Snap prices, Rebalances, corporate-action events, reporting cycles and market volatility can create significant spikes. In-house teams need enough permanent capacity to handle those peaks, even if that capacity is underutilized at other times.
Third, exception handling. The majority of data may flow through automated processes, but the exceptions are where human expertise is required — conflicting sources, missing prices, identifier issues, unusual corporate actions and other breaks.
Fourth, technology and engineering support. Firms often maintain their own ingestion pipelines, mappings, validation rules and integrations. When vendors change formats or requirements evolve, that creates an ongoing maintenance burden.
These aren’t necessarily inefficiencies caused by poor management. They are structural characteristics of running EDM internally.
So what can firms actually do?
There are several levers. Vendor consolidation and better commercial negotiation can help reduce the direct cost of market data. Standardizing data architectures can reduce duplication. Automation and AI can reduce the manual effort associated with repetitive exceptions. But increasingly, I believe another lever deserves serious consideration:
Changing the operating model itself.
Managed EDM is becoming a more credible proposition because it allows firms to move away from permanently funding regional coverage and peak capacity, while accessing specialist operational expertise and centralized engineering capabilities.
A well-designed managed-service model doesn’t mean handing over control. The client retains ownership of policy, source governance, material decisions and oversight. The service provider takes responsibility for day-to-day execution, SLA performance, workflow operations and approved operational change. That distinction is critical.
What could the economics look like?
The potential is significant, although the answer will obviously vary by institution.
An illustrative model in our recent analysis showed annual run-rate EDM costs reducing from $2.895m to $1.240m – a $1.655m, or 57%, reduction across four data domains. Importantly, the analysis argues that these savings come from structural changes – shared global coverage, better absorption of peak demand and centralized workflow and engineering support – rather than simply labor arbitrage.
That is an important distinction. The future of EDM isn’t simply about doing the same work more cheaply. It is about asking whether firms should be operating the same model at all.
The bigger opportunity
Perhaps the most interesting part is what comes next. Once data operations are standardized, centralized and measured, the path to further automation becomes much clearer. AI doesn’t need to replace the entire EDM function overnight. It can start with targeted use cases: exception triage, repetitive investigations, identifier conflicts and routine operational breaks.
So perhaps the question for CIOs, COOs and Chief Data Officers isn’t: “How can we manage our data more efficiently?” Perhaps it is: “Why are we operating all of this ourselves in the first place?”
As the cost of market data continues to rise and investment firms face increasing pressure to improve efficiency, resilience and scalability, I expect the conversation around EDM to move beyond technology.
For a deeper look at the economics of EDM, explore our white paper, The Financial Case for Outsourcing Enterprise Data Management.
