Market Data Has Become Too Important to Manage as an Afterthought

For an asset manager, market and reference data are not simply another operating expense. They are part of the raw material of the investment process.

Prices are needed to value portfolios. Reference data describes the instruments being held and traded. Credit ratings feed investment and risk processes. Indices define benchmarks and, increasingly, investment universes. Classification data supports portfolio construction, reporting and regulatory requirements. ESG data has added another substantial category. The list continues to grow.

Without these inputs, much of the machinery of modern investment management simply does not work.

Yet the way financial institutions manage the cost and consumption of this data has often not kept pace with its importance.

What was once a comparatively modest expense has developed into a material cost category. At the same time, it has become deeply embedded in investment processes, technology and client products. This combination makes market data unusually difficult to manage.

The question is therefore no longer simply whether an institution can negotiate a better price with its vendors.

The more important question is whether it really understands what data it needs, what it consumes, what it pays for and what it is contractually entitled to do with it.

The raw material of the investment factory

Think of an asset manager as a factory.

Investment strategies and client portfolios are the products. Portfolio managers, analysts, risk managers and operations teams operate the machinery. Technology provides much of the production infrastructure.

Market and reference data are among the raw materials entering the factory.

This distinction matters because market data is sometimes treated as a technical subject. In reality, its consumption is largely determined by the business.

An equity manager has different requirements from a fixed income manager. An active manager has different requirements from an index manager. A bank-owned asset manager may require additional data for accounting, regulatory or group reporting purposes. An institution operating across multiple asset classes and jurisdictions can have substantially more complex requirements again.

The growth of passive investing illustrates the point. Managing an index-related portfolio can require data not only for the securities currently held, but also for thousands of constituents within the relevant investment universe.

Regulation has added further requirements. So have demands for more sophisticated risk management, ESG integration and greater transparency towards clients.

The natural consequence is more data consumption.

At the same time, economics are attractive for data suppliers: Once a dataset has become embedded in investment processes, replacing it can be difficult. Models may depend on it. Investment guidelines may reference it. Historical data may have been built around it. Client agreements may explicitly name an index. Systems and interfaces may have been designed around a particular provider.

This creates considerable stickiness.

Annual price increases of 5, 6 or 7 percent may not appear dramatic in isolation. Applied repeatedly to a growing base of products and services, however, the effect compounds.

The result is a cost category that has gradually moved up the agenda of financial institutions.

The key word is gradually.

Market data costs rarely arrive as one large strategic decision. They accumulate.

A new strategy needs another dataset.
A client requests a particular benchmark.
Regulation creates a new reporting requirement.
A portfolio manager finds a useful data product.
Another team adds a credit rating source.
A new system requires an additional feed.

Each decision may be perfectly reasonable on its own.

Ten years later, the institution can be left with a market data landscape that nobody would have designed deliberately.

Death by a thousand cuts

This is why the central market data problem is not necessarily cost.

It is transparency.

Ask a financial institution how much it spends on market data and there will usually be an answer.

Ask precisely which products it receives from each provider, where those products are consumed, which investment or operational processes depend on them, which contractual rights apply and whether every product remains necessary, and the answer becomes considerably harder.

Ownership is often fragmented. Procurement may own the contract. IT may manage the feed. A portfolio manager may have requested the data. Operations may consume it. Risk may use it for another purpose. Finance may pay the invoice.

Each participant sees part of the picture.

Few see the whole thing.

This fragmentation creates a peculiar problem: an institution can simultaneously buy too much data and not license enough.

Paying for what you no longer need

The first problem is familiar.

Data tends to accumulate more easily than it disappears.

A strategy changes, but its old data package remains. An index is no longer used, but the wider index family continues to be licensed. A new provider duplicates data available elsewhere. Requests continue to run for securities no longer held. A dataset purchased for a project quietly becomes part of the permanent cost base.

None of this necessarily reflects poor management.

It reflects the reality that market data contracts and consumption evolve incrementally, while few institutions regularly return to first principles and ask:

If we were designing our market data landscape today, based on the products and services we currently offer, what would we really need?

Without that discipline, cost creep is almost inevitable.

And because contracts renew at different times, sit in different budgets and are sometimes allocated poorly across the organisation, the increase can remain relatively invisible.

There is rarely one invoice dramatic enough to trigger action.

It is, as we described during our recent Good Circle discussion, death by a thousand cuts.

Using what you have not licensed

The opposite problem can be more uncomfortable.

Seeing data does not necessarily mean you are entitled to use it.

A user may see an index, rating or other dataset through a terminal and reasonably assume that the organisation has paid for access. The institution may indeed already be paying the provider a substantial annual amount.

But market data licensing is highly specific. A contract may cover certain index families but not others. It may permit display use but not systematic consumption. It may cover an individual user but not enterprise distribution. It may allow data to be viewed in one application without permitting it to be downloaded and used elsewhere.

The distinction becomes particularly relevant when data moves from individual consumption into an enterprise process.

A portfolio manager may use a particular index because a client has requested it. The index becomes part of a mandate or reporting process. Years later, someone discovers that the relevant licence was never included in the existing agreement.

At that point, the institution does not have a cost optimisation question. It has a licensing question.

Market data providers generally retain contractual rights to audit usage. If data is required for a critical investment, regulatory or client process, there may be little practical choice other than to obtain the appropriate licence.

This is why a serious market data assessment can produce an initially counterintuitive result: it may identify substantial savings and additional licensing costs at the same time.

The objective is not to minimise every invoice. It is to arrive at the right cost base for the business.

A procurement exercise is not enough

The obvious response to rising market data costs is to negotiate harder.

Negotiation matters. But starting there is often a mistake.

You cannot negotiate your way out of data you genuinely need. Nor should you negotiate a contract before understanding whether the underlying consumption makes sense.

The first task is therefore to establish transparency. That means building an accurate view of the providers, contracts, products and spend across the organisation. More importantly, that commercial inventory needs to be connected to actual consumption:

Which processes use the data?
Which products depend on it?
Which systems receive it?
Who consumes it?
Which contractual rights apply?
Is the data still required?
Is equivalent information already available elsewhere?

This is where the subject becomes an operating model question rather than simply a procurement question.

A procurement specialist can understand the commercial terms of a contract extremely well without knowing whether a particular rating is essential to an investment process. A portfolio manager can know exactly which data is useful without understanding the contractual consequences of redistributing it through an internal system. IT can understand the data architecture without deciding whether a particular index family still has business value. Legal can interpret the licence without deciding whether the underlying product is necessary.

Effective market data management requires these perspectives to meet.

Start with what the business actually needs

A useful way to approach the problem is to reverse the traditional logic.

Do not start with the existing contracts. Start with the business.

What products and services does the institution offer today? Which asset classes does it manage? Which regulatory and client obligations must it fulfil? Which investment, risk, valuation, accounting and reporting processes are necessary to support them?

From there, determine the minimum data requirements needed to operate that business effectively and compliantly.

Only then compare that requirement with what is currently being purchased and consumed.

The differences become interesting.

Some services will prove essential.
Some will be convenient rather than necessary.
Some will be redundant.
Some data may be available through an existing provider that the organisation is already paying.
And some critical consumption may not be adequately licensed at all.

This creates the basis for an informed discussion with the business. Instead of asking a portfolio manager, “Can we cancel this?”, the organisation can ask, “What process requires this, what would happen without it, and what alternatives do we have?”

That is a much better conversation.

Technology helps, but it does not create governance

Market data inventory and vendor management platforms can support this process. They are useful for maintaining contracts, products, renewal dates, entitlements, spend and other information in a structured way.

But buying a tool does not solve the underlying problem.

A perfectly maintained inventory of poorly governed market data is still poorly governed market data.

Someone needs to own vendor relationships. Someone needs to understand usage rights. New requests need to be assessed against what the organisation already owns. Business users need to understand that data has a cost. Material new commitments should not be negotiated independently by individual users without considering the wider vendor relationship.

Cost allocation can help here. If the entire market data bill disappears into a central operations or IT budget, consumers have little economic incentive to question their requirements. Greater transparency does not mean every portfolio manager needs to receive an invoice, but users should understand the cost consequences of their choices.

There is also an architectural dimension. Where practical, institutions should avoid making their core data infrastructure unnecessarily dependent on a single data provider. A more vendor-neutral architecture can preserve flexibility and make future provider changes easier.

Consolidating more business with one provider may generate attractive commercial terms. It can also increase dependency.

The right answer therefore cannot be determined by price alone.

The economics extend beyond the first saving

A well-executed market data review can create immediate savings. Redundant services can be cancelled. Unnecessary consumption can be reduced. Contracts can be renegotiated. Providers can sometimes be consolidated or replaced.

But focusing exclusively on the first-year saving understates the value.

Consider a cost base that increases by 6 percent each year. Reducing that base today does not only save money today. It prevents future increases from compounding on expenditure that should never have existed in the first place.

There is a second structural benefit.

Many institutions effectively carry market data debt. Like technical debt, it is created when individually sensible short-term decisions accumulate into a landscape that becomes expensive and difficult to change.

A proper clean-up creates the opportunity to reset that landscape around the actual business. Instead of adding another provider every time a new requirement appears, the organisation can first ask whether the need can be met within its existing strategic data relationships.

That makes future growth easier to manage.

It can also improve time to market. If an institution understands its licences, available datasets and consumption rights, answering the question “Can we use this data for this new product?” becomes considerably easier.

Compliance improves for the same reason.

And management gains something that is surprisingly rare: the ability to forecast how the market data cost base is likely to evolve as the business changes.

Market data needs active management

Market and reference data will not become less important. Investment products are becoming more data-intensive. Regulation continues to demand information. Clients expect greater transparency. New technologies, including AI, will create further opportunities to consume and combine data.

Trying to solve the resulting economics simply by pushing vendors for lower prices misses the larger issue.

Financial institutions need to know what they need, what they own, what they use and what they are allowed to do with it.

That requires an inventory, but also an understanding of the investment and operational processes behind it. It requires commercial discipline, but also business judgement. It requires technology, but also governance.

And it requires regular attention.

A market data landscape that was appropriate five years ago is not necessarily appropriate today. The products may have changed. The clients may have changed. Regulation certainly has. The available providers and solutions may have changed as well.

The most effective institutions will therefore treat market data in the same way they treat other strategically important parts of their operating model: understand it, assign ownership, measure it, challenge it and redesign it when the business changes.

Because market data is not simply something financial institutions buy. It is part of what allows them to operate.

The objective should not be to spend as little as possible on it.