So You Want a Big Data Revolution?

In the interest of roadmapping Wall Street priorities for this year, we recently met with MDs and leading architects in various banks and financial services firms. What we got from those meetings is an inside look at what kinds of analytics projects these organizations are investing in; they include Anti-Money Laundering (AML) monitoring, trade surveillance, and Know Your Customer (KYC) analytics. The investments around those analytics in 2013 include:

  • Strengthening the golden sources – Security Master, Account Master, and Customer Master.
  • Implementing and managing various enterprise data-management initiatives – Data Quality, Data Lineage, Data Lifecycle Management, Data Maturity, and Enterprise Architecture procedures.

That’s easier said than done because the data complexity behind these analytics is staggering. For example, global powerhouse Citi has approximately 200 million accounts and business in more than 160 countries and jurisdictions. Indeed, many banks wrestle with a variety of data challenges, including:

  • Instrument Identification: All financial instruments, derivatives and loans must be precisely and uniquely identified. This is one of the basic building blocks of data management and business analysis.
  • Product Hierarchies: how to handle product dimensions /hierarchies effectively
  • Entity Identification: All business entities need to be precisely and uniquely identified so that links and relationships concerning the business structures underlying the financial industry can be evaluated. This “legal entity identifier” (LEI) standard is another one of the building blocks of data management and business analysis.
  • Business Ontology: All financial instruments and business relationships are defined by the terms of the underlying contract. The language of the contract must be both precise and comparable in order for financial institutions, investors, and regulators to fully understand rights, obligations, constraints, interconnections, and relationships.
  • Classification Schemes: These allow for the aggregation of granular data into analytical categories. Classification, according to underlying attributes, enables analysts and regulators to look at operations and investment strategies from a variety of perspectives (i.e. the flow of money, the structure of the instrument/business deal; concentration of liquidity or exposure, role performed, how one component relates to another, etc.).

This focus on data management fundamentals contradicts the hype around analytics. In the business world, the buzz is that banks are using sophisticated online and offline techniques to assess who their customers are and what they need, in order to present more successful upselling and cross-selling opportunities. They’re supposedly following the tracks of retail phenoms Target and Amazon, whose innovative strategies have ushered in a new era of marketing savvy.

At least, that’s what the media hype is telling us. But the reality is that the financial services industry, especially big banks, still has a long way to go, as analytics adoption (even customer analytics) in most financial institutions is still in its infancy. A recent survey by American Banker found that 71% of the 170 bankers polled don’t use any analytics, although that might not be true within a year. Among those non-users, the plans to buy analytics are less than ambitious. Only 2% plan to buy customer analytics in the next six months, 4% in the next six to 12 months, and 14% in more than a year from now.”

Why the hesitation in financial institutions?

According to the survey, 36% said that cost and competency were the biggest barriers; other IT issues also took precedence, with nearly 32% of those surveyed indicated that a focus on other initiatives got in the way of implementing customer analytics at their institution. And a full 23% of bankers polled also cited skepticism about the software’s ability to provide business value or a robust ROI.

But for some financial institutions the priorities are evolving. And as the operating environment improves, investments will undoubtedly increase in areas like AML and KYC analytics. The ROI may be there, but first companies have to strengthen the basics: enterprise data management (EDM) and master data management (MDM).

Achieving trust and confidence in data is a challenge in today’s business environment, thanks to independent business silos, inflexible IT environments, a lack of standards for data content and obstacles associated with gaining stakeholder alignment across the organization. EDM and MDM are both necessary for maintaining data quality, consistency, and integrity, and they have become absolutely critical in the complex regulatory regime facing financial institutions today.

Anti-Money Laundering – AML Analytics

The key to survival in today’s financial services market can be summed up as “Better know your customer.” In December 2012, U.S. authorities announced a $1.9 billion fine against British bank HSBC Holdings PLC for failed anti money-laundering controls that allegedly allowed drug money and transactions from sanctioned nations to flow through the U.S. financial system.

Led by the Justice Department, the Treasury and the Manhattan prosecutors, this sweeping investigation has ensnared six foreign banks in recent years, including Credit Suisse and Barclays. In June 2012, ING Bank reached a $619 million settlement to resolve claims that it had transferred billions of dollars into the U.S. for countries like Cuba and Iran that are currently under U.S. sanctions.

Also in December last year, U.S federal and state authorities also won a $327 million settlement from Standard Chartered PLC. The bank agreed to a larger settlement with New York’s banking regulator, admitting that it has processed thousands of transactions for Iranian and Sudanese clients through its American subsidiaries.

Clearly, banks need to invest in analytics that address the Anti-Money Laundering laws.

Know Your Customer – KYC Analytics

In the retail and capital markets, banks are being buffeted on many fronts. They face expanded and increasingly stringent regulatory requirements that are driving up compliance costs and, in many cases, restricting fee-based revenue. Advances in technology have enabled competitors to launch competing offers in a shorter timeframe, thereby curtailing product differentiation and eroding many institutions’ competitive edge.

At the same time, there’s competition from new kids on the block, including non-bank players offering alternative products, especially in the payments sector. The cost of doing business and acquiring customers is also escalating, spurring a renewed focus on customer relationship management (CRM) and retention, especially for the “right” customers. The challenge for many institutions is how to identify those customers.

To that end, banks are focused on achieving a new 360-degree view of their customers from a CRM perspective. This usually involves gaining insight about customers across various product silos; what’s more, that insight is necessary in order to deliver:

  • Engagement across channels and lines of business
  • Profitability based on multiple dimensions, such as by product, industry, geography, and other segmentations
  • Expense management
  • Risk across many dimensions

Despite an intense focus on the customer, many financial services organizations are struggling to extend customer insight. In most cases, it is not for a lack of data. Organizations are collecting more data than ever before, but they lack the ability to deal with this data deluge.

Moving to AML or KYC Analytics – A Simple Roadmap

Banks and financial institutions face multiple challenges in putting their data to work to build stronger relationships, improve return, and reduce risk. But not all institutions face the same challenges. Below are six typical challenges that must be addressed first:

Who’s in charge?

To sustain a “single version of the truth,” it’s necessary to document, understand, and actively manage the flow of (master) data across an organization and its systems. To enable this process, many organizations are setting up new teams or re-fashioning existing ones. Either way, new roles, responsibilities, and structures are still required. Identifying key resources, aligning them to a strategy, and evolving critical roles over time will enable long-term success with enterprise data management. Why do people-related issues become the biggest challenges in data management and analytics? What key roles must be formalized and how do they inter-relate? Which stakeholder management tactics are most effective? The behavioral and political issues around data require special attention.

Data silos still proliferate

The industry has been battling siloed data for decades, and the problem persists. In some cases, siloed environments preclude the creation of even a foundational aggregate customer view. The issue continues to proliferate with the emergence of new channels, as well as growth in cross-channel experiences. Likewise, disparate datasets can lead to multiple versions of the truth, depending on which department (finance, risk, line of business, product/marketing etc.) is looking at the data and via which system.

For example, the view of a customer from the CRM system would not typically incorporate a risk profile, performance history or regulatory data associated with Know Your Customer (KYC) requirements. The result is an incomplete – and possibly inaccurate – view of a customer, which can lead to inadequate decision-making.

Data is inconsistent

Expanding on the point of multiple versions of the truth, metrics across today’s financial institutions are rarely uniform. It’s common to find that behaviors and performance are not always tracked across all channels, let alone tracked consistently across the enterprise – a situation that limits accurate insight.

Disparate data sets that exist within the bank might not all be refreshed at the same frequency or using data from the same source systems. Some may completely ignore a few data sources leading to inconsistency at a given point in time. For example, finance may have an accurate cost of fund projections on a daily basis, while the sales system refreshes this information every month. The front office might be making decision based on stale data between the two refreshes, while the finance team is looking at these same decisions through a different lens.

Business processes remain disconnected from analytical insight

Institutional and experiential knowledge – much like the data in today’s FSIs – is siloed in departments, such as finance, risk or the front-office. For example, many front-office business processes continue to be based on “old knowledge” and “old data.” We see very little, if any, integration of front-office and middle-office systems, which could provide the most recent knowledge to support credit, pricing and offer decisions at the point of customer interaction.

Business effects are not timely

Many FSIs are focused on capturing customer interactions in a timely manner. The real hurdle lies in making these customer interactions quickly known and understood across the enterprise, so they can be leveraged in operational decisions. For example, during the recent financial downturn in which conditions changed rapidly, managers in the front-office were often left to make critical decisions based only on experience and their gut, instead of insight based on science and data. In many areas, this deficiency continues today. At the most basic level, transactional behavior and impact are not rapidly and widely disseminated to all decision points in most institutions today.

Lack of execution talent

The lack of project management and business analyst manpower, and the limited deployment of tools that put insight directly in the hands of those who need it are not the only reasons that financial institutions can’t glean timely insight. Maintenance in the analytical environment can also present challenges; rules around scoring and modeling are hard to maintain, and people-dependent and predictive models aren’t always continuously refreshed.

Data Management – CRAWL, WALK, RUN

As a precursor to AML or KYC Analytics, most financial institutions are racing to improve enterprise data-management capabilities.

The concept of data management as an essential component of business operations is gaining traction in the wake of the 2008 credit crisis; it also supports the transparency and systemic risk objectives contained within the Dodd–Frank Wall Street Reform Act and similar international directives, such as the European Market Infrastructure Regulation, Solvency II directives and the Basel Accords. All of these legislative initiatives, which require companies to comply with standards, are dependent on the availability of accurate and comparable data from many diverse sources.

One of the outcomes of the financial crisis is a strong and growing recognition by both financial institutions and regulators of the importance of monitoring risk with accurate, comprehensive and aligned data – and of sharing it across functions without the need for manual reconciliation or imprecise cross-referencing.

While the need for effective data management is clear, a comprehensive and standardized mechanism doesn’t exist yet. New frameworks from EDM Council like the DMM model are aimed at filling this gap. They provide a framework and assessment methodology for evaluating the effectiveness of data-management practices and a clear evolutionary path to establish a data management culture.

Looking ahead

“We always overestimate the change that will occur in the next two years and underestimate the change that will occur in the next 10.” Bill Gates

If you were to believe the hype, every industry is on the verge of an analytical revolution. But we still have a long way to go. This is especially the case in financial services (and retail banking). As we move from “the brink of financial Armageddon” to some form of health, banks are getting savvy with their data, especially customer data. Hopefully that’s what the Big Data revolution will be about.

Leave a Reply