When it comes to market data, equation is simple: sources + content + users + means = Bn 30 USD .
And if at first sight, we easily sense that it comes to selling shovels in a gold rush, some comments on each element of the equation may be useful to grasp what this is all about.
SOURCING: GIANTS AND DWARVES
a. The market data “Ivy league” of intermediaries.
While Bloomberg and Thomson-Reuters are first names to come to mind when thinking market data and account for almost 60% of the market , landscape is wider. The fact is these two still stand above all by benefitting from the stickiness of the information they provide (e.g. tickers and data are often widespread through company systems), relying upon their coverage, the terminals and messaging system they offer, the latter two being key differentiators.
Along with these two super aggregators, other traditional players such as Markit, S&P, FIS (ex Sungard), MSCI…still compete mainly by, to put it in a nutshell, providing more specialized information whether focusing on specific asset class whether on indices creation.
There from, current options boil down to choosing one of the two available “one-stop shops” or cherry-picking a range of specialized providers , the trigger being the ratio depth and breadth of information received over total cost.
However, traditional vendors also have to face primary sources and to cope with the rise of Direct Market Access as well as the threat of tech unicorns . Indeed, market data is both (hello world) market and data, and competition may come from both sides.
b. Market: trading places already stand as competitors.
1. Exchanges: the primary source
♦ It comes out as a roughly Bn 2 USD business for exchanges , market data revenues counting for a 10/15% range of total revenues for the 9 main exchanges.
♦ LSE is the main provider with a quarter of all market data revenues generated by main exchanges, market data counting for almost a third of LSE revenues.
♦ However, only the largest ones have the footprint to compete as data providers  while others still miss largest institutional global investors by lacking of sufficient liquidity due to their limited reach.
2. Inter-Dealer Brokers: the OTC skill
♦ Broker data revenues including post trade, risk, and analytics contributed for M 550 USD, i.e. 7%  of IDB revenues.
♦ ICAP and Tullett Prebon showed the highest revenues among all.
♦ Specialized IDB, e.g. in commodities and energy, remain out of the business for now.
c. Data: tech titans to create shade on current players sooner or later?
♦ Amazon, Google, IBM, Microsoft, on an already US-centric market, but also TenCent  appear much stronger “in terms of […] technology, adaptability, market reach, and financial firepower”.
♦ Google already partnered with some exchanges to provide live prices for free .
But, whether market data came out as a by-product or a core offer, service provided is manifold.
CONTENT: CARROTS AND PEAS
a. Evolution: consumption drives content
Information evolved with “a subtle but important switch from ‘How to Source’ requirements, to a more interventionist ‘What to Source’ requirements” .
And three steps can be identified so far:
1. Vendor broadcasting prices directly to users for any need they may have (with thus terminal business booming in the 80’s).
2. The “enterprise-wide usage” (i.e. from trading to risk monitoring systems to reporting tools) along with the sort of blossom of analytics.
3. And a more regulatory (LEI, AML related data) and technologic oriented one.
b. The range: it is all about added value
As mentioned previously, to compete with Bloomberg or Thomson Reuters, other providers had to take the path of specialization and/or proprietary data. As showed below, data value thus depends on whether it is “original work”, “original sources” or data publicly available.
USE: “CLIENTS BUY DATA TO MAKE MONEY FROM IT, OR BECAUSE THEY HAVE TO” .
If we stereotype, information is possibly provided for:
1. Building a strategy and taking investment decisions.
2. Processing of trades.
3. Market reporting.
However, very final users tend to change and reasons why data is needed also evolved as the need for data from traders decreased with the rise of algotrading when in parallel:
1. Retail market increases its usage when more High Net Worth Individual and New Mass Affluent markets are trading on their own (only needing information on bid/ask and not the full depth of the order book) while getting their data from brokers before that.
2. Compliance and regulations are calling for more market data use to comply with new obligations (e.g. MiFID2, the Fundamental Review of the Trading Book, the IOSCO Principles on Financial Markets).
Market data thus spreads all over companies from pricing engines, through internal analytics and models, to trading platforms, through all calculations performed (e.g. collateral, margin, capital adequacy) as well as portfolio and valuation management or also regulatory and client reporting.
There from, such a diverse use calls for a global picture to grasp all stakes for the company consuming data, hence in some cases the creation of Head of Market Data positions which also smooth the overall relationship for vendors.
TECHNOLOGY IN THE DRIVER SEAT
When picturing market data, anyone visualizes a rainbow-colored-so-70’s Bloomberg keyboard. And when the leading company would possibly get up to 75% of its revenues from its terminal business , other players also provide such feature. However, range in prices is wide from Money.net starting at 1.8k USD per terminal per year to Thomson Reuters, charging from 3.6k to more than 22k, and Bloomberg and its 20 or 24k fee for a 2-year minimum subscription (with in between others like FIS MarketMap, IDC eSignal, Morningstar Quote Speed or SIX ID).
Sticky and expensive, in particular with “additional exchange fees and third party services [that] can add up to 40% more” , terminals remain very profitable (325,000 terminals for Bloomberg  and 190 000 for Thomson Reuters : do the math), this business stream nevertheless tends to decrease as:
1. Large dealing desks are shrinking , with thus less people using terminals with the development of electronic trading.
2. Direct Market Access is cheap  and promotes greater usage of feeds, very sticky once shared all over company systems for different uses, hence new competition to one-stop-shops  with lowered barriers to entry (at least for exchange sourced data, not OTC nor contributed data).
Thus, even if not at the same scale (e.g. not providing the same breadth of OTC market data), selecting providers per feature  is to be considered as a serious option when eased by API and in particular when Symphony for example is spreading as an alternative to Bloomberg messaging.
And in addition to this switch from terminals to feeds, technology already started to re-shape market data landscape with the Cloud, by enabling “co-location” of infrastructures and by-passing intermediaries.
And down the road, we can also easily imagine blockchain (the magic word), as a mean to generate smart contracts, to help users in monitoring both their consumption and their contract application.
CONSIDERED AS BLOOD FOR FINANCE COMPANIES, MARKET DATA CALL FOR AN EXECUTIVE LEVEL SPONSORSHIP
Already among top operational expenses, market data is most of the time underestimated as only considered from a licensing costs perspective, when some companies may spend even more on hidden market data management costs mainly including human resources involved  e.g. in data cleansing for a proper use within the company.
Because of the level of spending and of the low readability of market data usage, due to the low correlation between AUM and market data licensing costs , the hidden costs mentioned before as well as the complexity of vendors contracts, the topic has to be addressed at executive level of companies.
That being said, market data is obviously a costs question to be tackled with, by order of potential of savings, four levers to work on:
1. Demand management.
2. Vendor management.
3. Technology management.
4. Administrative management to ease transparency and cost-awareness.
Nevertheless, if only “viewed through the limited lens of cost-savings, market data management is caught in a catch-22 [as] funding and executive support emerge when savings are found, but savings can only be found when funding and executive support for a program exists” . Hence the need to tackle the topic as a strategic one and to take action proactively ahead of shocks as for example:
1. “A damaging vendor audit (financial and reputational risk).
2. Declining capital markets (margin preservation).
3. Management change (bold vision)” .
MARKET DATA MANAGEMENT MAKES NO EXCEPTION: COMMON SENSE IS BEST
At last, market data cost is so important for users that doing nothing but paying is not an option and, whether steered by a Head of Market Data or not, sequence of actions to be taken is not rocket science:
1. Needs call for an objective inventory across departments within a company.
2. Once done, sourcing options have to be assessed to best meet the requirements.
3. A sharp and up-to-date costs monitoring framework is to be put in place to enable a) to digest all providers fee grids and b) a close follow-up of data consumption.
Bottom line: it’s all about sponsorship on the topic at top level to allow to a clear strategy definition and a smooth and efficient implementation.
 Keiren Harris lists for example CME on equities, Pricing Direct on FI, FX and Money Market via BARX, Commodities and Energy through Marex Spectron, Reference Data for Exchange Data International and the OECD for Economics
 For example LSE, Deutsche Bourse and NASDAQ OMX remarkably stand or stood as data vendors (through respectively MNI, Proquote and NASDAQ IR Insight)
 Noticeably with LSE since 2012: see Reuters “Google offers live London stock prices for free”
 Business Insider reports on Wall Street 10,000 front line jobs have disappeared since 2010 with flagship examples such as the ditching of UBS’ Connecticut world largest Dealing Room or Goldman Sachs cutting one of its three trading floors in Manhattan.
 In a way, it does not really add costs on sources as they already have to develop infrastructure to deliver prices.
 Such as Activ Financial, Quant House, and Xignite
 See supra
 Shankar Subramanian assessed that 75% of non-license market data costs can be attributed to people, 10% to market data systems (e.g. reference data management systems, ETL tools, data warehouses or hubs), 9% to other systems (such as add-ons services provided by vendors to ease acquiring, cleansing and delivery of data) and 6% to service providers (e.g. custodians) offering market data management as a by-product. “The True Cost of Market Data: Operational Impacts”, Cutter AdvantEdge, 2014