Meeting the data challenge

Data is getting bigger and more important every day – so why is it still being ignored and/or poorly treated? Here Carsten Kraus of Omikron Data Quality, Steve Tuck of Datanomic and Asset Control’s Martijn Groot explain how to meet the challenge head on.

FST. Financial institutions are swimming in data. What are their key data management challenges?
ST. Whilst it’s easy for organisations to measure the quantity of data, few have a firm grip on this valuable asset; with data often split across silos, financial institutions are often working with an incomplete or inaccurate view of their business when making critical business decisions.

Financial institutions must instate a clear data management strategy that guarantees data security but also ensures that key knowledge workers have access to the information they need to successfully do their jobs. The data they use to make informed decisions must be complete, consistent and coherent.

CK. For both BI and CRM, it is vital to create a unified customer view throughout all parts of the organisation.

AC. Implementing change involving data management requires a combination of reliable technology and good processes. The more challenging of these two often relates to business processes. Success with these business processes requires a holistic view of data management and requires a complete understanding of existing processes and any regulatory, productivity, and strategic issues involving these often diverse needs.

Clients are exposed to various types of risk as a result of the vast torrent of data they are forced to confront. Simultaneously, they are finding that regulations have become much stricter. Trading volumes and increasingly complex products create operational challenges.

Another emerging issue relates to the speed at which the data is processed and concerns of latency – the lag time between the time it takes to transmit and receive data – that can no longer be measured in milliseconds but microseconds. Just a few milliseconds can mean a huge difference in the price of a stock and the multimillions of dollars in lost opportunity costs. This highlights the need for quickly normalised, reliable data. The key challenges for a bank when thinking about data management include:

• How will we to collect it?
• What will go into its maintenance and cleansing?
• How will the data be standardised?
• How will we provide adequate levels of service in our agreements for our internal business lines?
• What processes are we going to put in place to ensure that the different business lines, such as the equities and fixed income departments, are getting their information in a timely manner and without redundancy in technology and effort.

These business processes that surround data management concerns require the anticipation of future business needs to determine how to put in place data management architecture that both satisfies today’s pressing issues, such as regulations, new market entry, and positions businesses for future challenges and growth.

FST. A centralised approach to data management makes intuitive sense in many ways – and indeed some financial institutions are now introducing Chief Data Officers. What are the advantages of this avenue, and what drawbacks are there?
CK. The drawback is that in some cases nothing gets done. In my opinion it is important to have a centralised vision, but to start with single initiatives – if the CDO is guiding and coordinating efforts rather than planning and authorising them, results will be achieved more quickly and often more appropriately.

ST. The arguments for having a centralised approach to data management are compelling; it provides an organisation with a clear focus on information management and recognises the value of data as a corporate asset. Preferably, this should be supported by the appointment of a senior executive with responsibility for data management within the organisation, such as a Chief Data Officer who should ideally report directly to the CEO. Their key responsibilities are to define the strategic priorities for data systems and processes and to identify new opportunities to exploit the institution’s information to deliver business benefit.

Without executive sponsorship any data management strategy is likely to fail. A centralised approach may face resistance from pockets of workers who fear a lack of ownership of data, but a strong executive mandate can combat this and deliver greater benefit for the organisation as a whole.

Rather than ad hoc standards and processes, a coherent approach to managing information ensures that data is seen as a corporate asset and the responsibilities of everybody collecting and using it are clearly defined. Data governance should be supported by the appointment throughout the business of Data Stewards, who have the responsibility for maintaining data standards.

AC. We are strong proponents of a centralised approached to data management, including the role of Chief Data Officer, which is also sometimes informally referred to as a “Data Czar.” There is much overlap between the tools and processes used to successfully manage financial data. Employing this centralised approach enables efficiency gains and allows for the exploitation of a myriad of profit-making opportunities both within business units and across them.

FST. Data quality is essential to get any value out of your data management operations. How can this be ensured?
AC. Ensuring data quality requires the appropriate operational tools, applications, and business processes within the organisation. Data management operations require applications that enable data analysts to compare, investigate, and fix reference and market data. This toolset coupled with workflow enables the data management operations to erect a structure in place that defines how data is collected, viewed, repaired, and maintained for downstream business applications. Furthermore, it has been our experience that it is necessary to provide a complete, multi-dimensional audit trail.

An audit trail should enable review of the content each provider contributed to the golden copy view and the original format in which the information was delivered. At the end of this process, untreated vendor or internal data, a normalised or client-enriched view, and a final approved “golden copy” version of the instrument, corporate action, SSI and counterparty data can be viewed side-by-side for review by members of an organisation, in addition to auditors and regulators.

ST. Poor data quality undermines the value of information; it leads users to distrust the applications they use, which often results in more garbage being entered. It is tempting for organisations to address this challenge by implementing stringent data entry standards, but if these are poorly conceived they too can lead to the creation of more erroneous information as users become ever more creative in their efforts to circumvent the imposed checks.

Best practice is to address data quality holistically and systematically: which requires appropriate technology and processes. Organisations must 1) understand what data they have and how it is used, 2) improve the quality of critical business information to ensure that it is fit for purpose, 3) protect applications from the entry of flawed or duplicated data, and 4) control data quality by continuously monitoring performance against clear objectives.

CK. For some aspects, DQ can be ensured by software modules that can be integrated into IT systems, such as postal validation or duplicate checking. For all other aspects, the most important factor is awareness. In my opinion, making everyone aware of the necessity for correct, updated, and complete data is best achieved by showing them the impact of what is done with the data, preferably in their own operational surroundings. To find out who is still ignorant of DQ is the second step. For this, you need DQ monitoring software.

FST. With the focus on compliance and AML in the financial world in recent years, how can better data management help companies meet some of the requirements of regulators?
AC. Regulations such as Sarbanes Oxley and Basel II are the key drivers here. Many of our customers are large banks who use the advanced measurement approach (AMA) for the calculation of their operational risk charge under Basel II. Consequently, they need to get their data processes under control. Workflow tools enable online event resolution and detailed logging and can be applied to cleansing and reconciliation of custodial data, new instrument set-up, price discrepancies, and new-client set up processes. For root cause analysis of loss events, the ability to drill down into the various sources that make up a golden copy view is essential. This ensures data quality and minimises fails and claims in the settlement process.

In addition, the forthcoming MiFID regulation in the European Union implies that institutions have to classify their clients according to a firm’s specific regulatory obligations for the business they conduct with a certain client category. This will cause financial institutions to reclassify their current client data, which in many cases is largely scattered over different business lines and various technical architectures.

ST. Compliance with regulations such as the EU 3rd Money Laundering Directive (EU3MLD), MiFID and Basel II can only be assured with good data management practice. For instance, EU3MLD requires every financial institution in Europe to act to further tighten controls to prevent access to the financial system by terrorists, money launderers and other criminals. Worryingly though, there’s a huge gulf between those exemplifying best practice and the rest.

Whilst a growing number of financial institutions are approaching EU3MLD by employing best practice business rules and sophisticated matching technology, too many are falling behind and some are taking a laissez-faire attitude, which may land them in court. They risk damage to the reputation as well as financial and possibly custodial penalties if they are found to have been negligent in their approach to risk-based anti-money laundering.

In the case of Basel II, whilst the UK’s Financial Services Authority has decided that it is not appropriate for it to directly audit the quality of an organisation’s data it has stated the majority of firms would “have difficulty certifying the quality standard they apply to their data”. Meanwhile, delays in ratifying MiFID across Europe are being used by many organisations to turn a blind eye to the data standards it requires.

FST. Having lots of information stored is one thing. Extracting the insight and value from that data is another – how can banks meet the challenge of getting the most value from their data?
CK. Once a unified customer view is achieved, many analyses are possible with BI tools.

ST. Data has a limited shelf life. For example financial institutions’ customer data generally degrades at a rate of 20-30 percent every year. You cannot rely on your customers to tell you about changes in their employment status or personal circumstances unless they think it has direct relevance to the services you provide them with. Indeed, you can’t even rely on them to always tell you about simple facts like a change of address; a situation that has led the Royal Bank of Scotland to introduce a charge for those credit card customers that neglect to inform them.

Making best use of data through business intelligence and actively engaging with customers through focused campaigns not only optimises the value you can extract from your data. It also provides the best route to maintain and enhance the knowledge you have about your customer base. Individuals and organisations are, more often than not, willing to share more information with a financial institution that responds to their needs and provides them with appropriate offers in a timely manner.

AC. The ability to have all data cleansed and available in one place prepares the groundwork; disclosure of that information to decision-making points downstream lends it value. Any solution should be tailored for different kinds of use, such as ad-hoc queries and bulk processing, and should also be able to integrate technology standards for the dissemination of information. Functionally, it should always be clear who owns which piece of information, its quality grading, and how it was derived.

FST. Could you outline some of the solutions your company provides?
AC. Our flagship product AC Plus for in-house data management consists of various components, including AC Server, the backbone server that gathers, integrates, cleanses, standardises and distributes investment data. A second component is AC Data Analyzer, a modelling and development rules engine used throughout the entire solution, which stores, maintains and extends the business rules and workflow for data mapping and cleansing. It flags suspect data, automatically corrects data based on business rules, and develops enriched and complex derived data.

Additionally, our TAPMaster Data Access provides automated data acquisition for 22 different vendor sources, contains open platform technology providing access to the central data repository using common tools, uses automated data deployment and distribution via SQL, ODBC, APIs, and add-on modules like TAPMaster OnDemand and Publisher and contains a push and pull capability using TAPMaster connectors to common portfolio management, trade order management, and general ledger systems.

ST. Datanomic provides a unique and innovative data quality software system, dn:Director. Together with our structured methodology, this allows organisations to understand, improve, protect and control the quality and value of their information assets. It is used to deliver data quality and data migration projects and is a critical enabler for Master Data Management and Customer Data Integration initiatives and is described by Bloor Research as “arguably, the most flexible data quality product” available.

Datanomic has extensive experience in the Financial Services sector and, through its work with clients such as Barclaycard, Alliance & Leicester and Global Asset Management, has developed a range of template solutions based around the dn:Director platform. One area of particular focus is compliance solutions, with solutions integrating best business practice and state-of-the-art technology to address EU3MLD, MiFID and Basel II.

CK. One of our core competencies is matching. Our new matching algorithm for international data is able to consider the optical similarity between Chinese characters, as well as the endings of feminine names in Slavic languages, or optional vowels in Arabic languages – all at the same time, allowing a truly worldwide matching.

The Omikron Data Quality Server is a platform solution designed to ensure data quality in an SOA environment. As a platform, it can encompass multiple modules including the matching mentioned above, but also data structuring, postal validation, reference data addition, etc. It can connect directly to various data sources in your organisation, in addition to being addressed as a web service via SOAP, and thus is able to work as a customer data hub ensuring a unified customer view between various operational systems.