With an industry-wide cost of more than $28 billion per year, market and reference data is often cited as financial firms’ third-largest expense, behind staff costs and office space. And since the financial crisis, data professionals have faced ever-tightening purse-strings and demands from management to reduce data costs year after year, while also coming up against end-users hungry for increased access to new data types.

“The volume of data being created worldwide is tripling every year, and the expected growth in what we spend on data is less than 20 percent. So the gap between the growth in data and our spend on it is growing much faster. And to close that gap, we have to be more efficient with data,” says Caroline Sherman, managing director and head of the Enterprise platform at crowd-sourced hedge fund Quantopian.

So how are data professionals tackling these demands? Often, in many of the same ways that they have for decades. But with significant savings proving harder to come by, firms must become more creative in their approaches, and be willing to embrace—or initiate—more innovative solutions.

“We would expect all firms to have found and managed the low-hanging fruit. But that doesn’t mean there aren’t more opportunities to reduce costs out there,” says a market data manager at a large US investment bank.

But these are becoming increasingly difficult to find and execute on, with traditional cost rationalization exercises yielding “very little,” according to one former buy-side data manager. 

“I spent six months on one index provider, and we found some locations that could be cut off … but their standard price increases will eat up those savings—so you’re just treading water,” the buy-side data manager says.

Broadly speaking, methods for managing data costs—and potential areas for innovation—fall into two areas: technical and commercial. Technical opportunities refer to the ability to reduce data spend by analyzing usage and data delivery, while commercial factors include new approaches to negotiating terms, or new business models around data fees and licensing.

On the technical side, the investment bank data manager highlights the potential to reengineer delivery of market data, removing inefficiencies, and replacing or outsourcing incumbent systems with lower-cost alternatives. “One would imagine there are many legacy systems on Wall Street that could benefit from new technologies or from outsourcing processes to reduce the technical cost and footprint of market data,” he says.

Some methods of identifying costs and savings are antiquated, such as manually listing individuals and the services they use—which can be costly and time-consuming, often requiring consultants to be on-site for several months—then recommending eliminating services altogether or replacing them with cheaper alternatives. Some more “modern” services—such as basic inventory management platforms—merely automate this process without leading to a more strategic conclusion. And while some of these processes are run in-house on an ongoing basis, others are performed ad-hoc by external consultants, which can allow bad practices and costs to inch upward again between evaluations.

“There are natural inflators of market data costs, so costs will always creep back up,” especially when many data administration departments are still under-staffed, post-financial crisis, says Stephen Veasey, CEO of data consultancy 3D Innovations (3Di).

The Times They Are A-Changing

“The days of looking at a static report to see if you could save costs are over. Now, you need to see intraday where data is going and to who, and why it’s going there, and you need to assign metadata to data—for example, is it for generating reports, or for front-office trading decisions, because the cost of using something for trading in the front office can be more expensive,” says Amjad Zoghbi, director at Xpansion Financial Technology Services, whose Xmon platform can assign and track costs by individual consumer, desk, office, region, application, job function or business process.

“So you can see how much data is being used for testing or in reports, and how much each of those uses is costing. It elevates the discussion and shifts it from market data administration teams to business decision makers. They think data is cheap. So when we put this in front of them and show them how much it costs … in some cases, it opens their eyes to the cost of data,” Zoghbi says.

Part of Xmon is a caching layer that prevents duplicative re-requests of the same data. West Highland Support Services’ Reference Data Framework (RDF), launched earlier this year, provides a similar approach, and—by employing an abstraction layer, allows users to change the data source for a user or service, says Jeff Hays, RDF product manager at West Highland.

Distinguishing usage types is also a key feature of the Market Data Grip tool offered by Dutch vendor Rivium Business Solutions, which can segment data consumers with lower demand and determine suitable data solutions, rather than automatically giving them a full-featured premium terminal.

Hans van Sligtenhorst, business unit manager for market data at Rivium, says the tool helps counteract bad habits at end-user firms—such as copying the profile for new hires from previous user profiles, which can mean that from day one, new joiners are automatically given access to a more expensive package of data than they may need.

“We have about 80 different business profiles for content and capabilities within terminals and how they contribute to the business … so you can combine a theoretical profile of what we think someone should need to do their job with actual usage data,” says Kees Brooimans, CEO of the Screen group of businesses at TRG Screen. But simply developing more granular profiles isn’t innovative in itself, he says. Rather, leading-edge firms are handing off inventory management and usage tracking to vendors like TRG Screen so they can focus on strategy, business analysis and procurement. “That’s more innovative—spending time on areas that add more value, rather than on processing an invoice,” he adds.

A new twist on an older approach is offered by EasyFeed.io, a Paris-based startup that prices every request for data in real time and suggests alternative sources that could allow banks, asset managers, pension funds, asset servicing companies and insurers to save money.

“We work at the request level … and we calculate costs based on requests, so we can price every request, and in the context of other requests coming in during the month,” says Vincent Goubert, managing partner at EasyFeed. “We capture and process every request in 300 milliseconds. So, for example, if you are a securities servicing company and you are about to onboard a new asset manager client, you can assess the cost for their funds and vendors in advance.”

However, old-school displacement strategies are still commonplace.

“The idea at a lot of larger firms is to explore low-cost alternatives, but usually as a stalking horse to gain leverage over incumbent vendors,” says Barry Raskin, managing director at market data and management consulting firm Jordan & Jordan, noting that the scale of a project to replace services—especially, for example, direct feeds—can be prohibitively complicated.

Instead, firms need to be smarter about what they buy and how they buy it, he says. “Firms should take a deep dive into their infrastructure and workflow to make sure they are buying smart and not paying for data more than once,” such as when one firm acquires another with similar subscriptions, or buys the same data direct from an exchange, and again via real-time vendor services, as well as end-of-day files, Raskin says. “There are massive overlaps at large institutions that are costing a lot of money. But that kind of cost-saving analysis is complicated because you have to look at the plumbing, and many people own different parts of the plumbing.”

The way firms organize their data access and processing can also have a big impact on cost management, says Dessa Glasser, principal at the Financial Risk Group, and former chief data officer (CDO) at JPMorgan Asset Management.

“People can now leverage the cloud and data services: You build a logical model with a data services layer, so you can re-use your data. … Once data becomes a shared asset, it’s in one logical place rather than having every unit bring data in themselves, so you don’t have everybody redundantly cleaning it and moving it around,” Glasser says. “It’s about being smarter about how you use and purpose the data.”

Instead of a business unit acquiring, onboarding, formatting and scrubbing data separate from other areas of the firm, which could mean it takes a month before the data is actually usable, firms should federate some of the work and apply “appropriate governance,” rather than leaving a CDO office to implement a heavier-handed approach.

“Federated ownership is actually cheaper to maintain, and gives access to data closer to the source,” Glasser says. “It makes it much easier to allocate costs, because you can see who is creating and using data, and who isn’t … and you can put restrictions on who can access it. In addition, it causes vendors to become more streamlined and more competitive on costs. And by having a data layer, you can outsource, or move to a utility, or replace vendors more easily—so it forces vendors to focus on competing and on value-add.”

Negotiation Needs Knowledge

Since data organizations are being held to new standards when it comes to saving money, they also need to hold their providers to new standards when negotiating deals. 

“People are looking at enterprise deals more than in the past, including all-you-can-eat deals. Or maybe a firm already has an all-you-can-eat deal and wants more, or only uses a fraction of the data included in the deal and wants to only pay for what it uses,” Raskin says. 

“Other new areas for cost savings include, for example, what Symphony Communication Services is trying to do around open APIs. Firms are seeing that they don’t have to be married to any single data platform, but can use APIs to bring in their preferred data sources, whether internal or third-party data,” he adds, but warns that this approach will not yield immediate short-term savings. “It’s a longer-term play, and requires a big investment of capital.”

Web services data provider Xignite has been focusing on emulation application programming interfaces (APIs)—which create an abstraction layer between data source and consuming application, allowing firms to switch data sources and phase out legacy infrastructure—for several years, says CEO Stephane Dubois. And while this model and data delivery via the cloud can save millions of dollars, he warns that “if you really want to drive cost control, control is knowledge of what costs how much. … If you don’t know how much something is costing you, and you can’t associate it with a business area, then you can’t fix it.”

Harald Bina, COO of MarketMap at financial technology provider FIS, says that while initiatives like virtualizing datacenters and automating processes can deliver lower costs and greater controls, knowledge is key. “A market data manager must have a pretty deep knowledge of what’s available out there and understand what vendors provide, to really drive efficiencies within a firm. It’s getting more sophisticated for a market data manager to stay on top of things,” he adds. 

3Di’s Veasey says that over 16 years the consultancy has delivered an average 11 percent reduction in clients’ market data budgets, but adds that the “big wins” come from “technical rebalancing and rationalization of supporting infrastructures.” Other areas of potential savings—such as license arbitrage between ratings and index providers—can also yield results, but require detailed work. 

“There’s a myth that you need all the top three ratings agencies. Often, you find that some functions just need one, some just need two, and some can use an internal rate … and there might be roles where you need more,” Veasey says. “You have to work very closely with the business side—and often the business doesn’t allow you to get that close because it’s busy doing its job, and doesn’t have time for a debate about data costs.”

Once a firm has decided its position, if negotiations fail to deliver the desired result, “nothing succeeds like demand management,” the investment bank data manager says. “Present it to the users and ask if they need those costs. The users will make intelligent choices. In our firm, every user has to re-certify their usage every year—and every year, we take between $4 million and $5 million out of our costs.”

While re-certification programs can prove useful for controlling costs of individual subscriber-based services, some of the more challenging sources of data costs—such as index data and ratings—are typically licensed on a firm-wide basis.

“Really, the only way to save a significant amount is by disengaging from a vendor or dataset. … You could take out $500,000 to $1 million of costs annually in one fell swoop by eliminating suppliers entirely,” the buy-side data manager says. “To get to the next level of savings requires a dialogue with the business and asset owners to review the mandates required,” such as the indexes used to benchmark performance, or requirements to use more than one source of ratings data.

“There seems to be a disconnect between knowing what you need to do and actually doing it—for example, being outraged over a ratings agency raising fees versus going to clients and saying you’re going to remove that agency,” he adds. “Have you engaged the asset owners? They could achieve better cost savings on investment products if they would be more flexible about what benchmarks you can use.”

Tipping the Playing Field

There is a third way of managing data costs: acting preventatively or disruptively to challenge or offset expense. “When a provider or industry group is recalcitrant, we look at what we can do to introduce competition, such as initiatives to support new entrants into the index space to make the industry more competitive,” the investment bank data manager says. “If we can’t improve a deal with a vendor through negotiation, can we improve the landscape through competition? The more competition there is, the less you see behavior such as raising prices without adding value.”

MarketMap’s Bina concurs: “One thing we’ve observed among our major accounts is that where index providers become very expensive—if not cost-prohibitive—and inflexible, this has triggered a number of firms to create their own indexes to track the market, or to base investment products on alternative indexes,” he says.

Of course, they say offense is the best defense, and some firms are turning their data assets from an expense into a revenue source. Goldman Sachs and JPMorgan have both begun investigating the potential of selling proprietary data to quantitative firms as “alternative” datasets. Could firms with high data consumption bills explore ways to monetize data assets and offset demands to reduce data spend?

“If you want to pay for new data types by cutting other costs, then you have a limited ceiling,” says Quantopian’s Sherman. “When you’re looking at growing the top line, budget is far less constrained—so I think these are treated differently.”

The individual with oversight of these separate areas would usually be the chief data officer, though the traditional demands of this role may be at odds with the desire to commercialize data assets.

“In my previous roles, I’ve been not just the guardian of data, but also the person responsible for commercializing it—or, more accurately, for leveraging the asset value of data. That’s a big distinction, because it’s not always just about making money from data. Sometimes maximizing value is about reducing risk and streamlining operations as well,” says a data consultant and former CDO, adding that the CDO role is typically a shared service supporting the needs of the entire organization, while a commercial function would generally be associated with specific P&L verticals. 

Glasser says this shift in the structure of CDOs is already underway. 

“I do think we will see chief data officers becoming more aligned with the business, and not just holding a purely governance-related role. Unfortunately, most CDOs have traditionally been aligned mainly to risk and regulatory functions, and anything revenue-generating is usually handled separately… so as the CDO moves away from a purely regulatory role and more towards finding insights in the data, this could be a natural extension of that,” Glasser says.

John Bottega, executive director of industry association the EDM Council, who has served as CDO at Bank of America, the Federal Reserve Bank of New York, and Citigroup, acknowledges that the role of the CDO is changing to encompass data acquisition, and to work with other key roles within an organization—such as the other CDO, the chief digital officer, to create “digital storefronts” for data—but perhaps not all the commercial responsibilities that others envisage.

“Governance and commercial functions are different skillsets. I’m not a marketing expert, so to sell a store of data, I’d need a marketing expert,” Bottega says. “If a bank has stumbled into a good source of data and wants to commercialize it, then it probably needs experts to do that. … It takes a lot of infrastructure and people to sell data—it’s a full-time job. But as the industry has changed, we’ve seen more cases where a firm may well have data that could be used by other organizations.”

However, he says he expects cost controls to be ever-present, and that firms are unlikely to accept the promise of revenues to offset higher costs. With no obvious way to counter the rising cost of data, the pressure on firms to adopt more innovative approaches to cost management will only increase.