Improving Data Quality - The Stream Financial Approach

Improving Data Quality - The Stream Financial Approach

Improving Data Quality - The Stream Financial Approach

Business Challenge Post Financial Crisis

A key lesson from the recent financial crisis is that management at many levels within large banks often lacked high quality information, and as a result the banks were often difficult to manage. As well as a greater management awareness of the importance of information quality, regulatory pressure since the crisis has resulted in an issue that is both more visible and more critical to address.

Business, Data and IT Architecture

In January 2013 a regulatory paper (known as BCBS239) attributed the senior management information gap to inadequate IT and data architectures. We think the problem is more fundamental. We believe it is caused by bad user discipline in the processes that deliver this information - for example, many banks use adjacent and inconsistent data adjustment processes rather than fixing initial data problems at source. The regulatory paper does, however, provide an opportunity to address these inadequacies.

IT systems have historically been delivered in business and functional silos that reinforce this lack of discipline and create communication barriers that slow the feedback loop involved in the production of management information in large organisations. This speed of operation of this feedback loop is critical to produce high quality information.

Feedback Loop and Partitioning

The business process can be considered as a feedback loop as can the management information production process. A short feedback loop is key for the production of high quality management information which then allows for a short business feedback loop.

Partitioning of the process design is key. Business, data and systems need to be partitioned independently but in a way that connects them together. Historically, data and systems have been partitioned along business lines which leads to inefficiency and slow feedback loops.

Analogy to Control Systems and Military Strategy

A useful analogy for the business process is the science of control systems. In this case, a system (such as industrial equipment) is measured and the measurement data is used by a controller to alter the inputs to the system in some way. The procedure repeats itself, thus creating a "feedback loop".

A second analogy is military strategy. Following the Korean War, US Air Force Colonel John Boyd, came up with the concept of using a feedback loop to describe the process a fighter pilot operates while in combat with an enemy. Boyd described the recursive loop as a OODA loop - Observe, Orient, Decide, Act. An entity (whether an individual or an organization) that can process this cycle (i.e. observing and reacting to unfolding events) more quickly than an opponent can gain the advantage.

In the case of business, the "observing" part of the feedback loop involves the obtaining or creation of information. High quality information is, thus, fundamental to any business to support good decision-making. By information we mean both data and the context of the data - for example, a bank may publish a profit figure (i.e. data) and the period over which the profit is for (i.e. part of the data context). High quality information requires both data and data context to be of high quality. By high quality, we mean data that is timely, accurate, complete and consistent across organisational boundaries and appropriate for both everyday use and times of crisis.

Data Production Process

The data production process often has it's own feedback loop, whereby "operations staff" check the data quality and if an exception has been detected re-run the process or adjust the data before releasing the data set to business users. This separation of duties (i.e. a partitioning of the business) may lead to inefficient process if the data and IT systems are partitioned in the same way.

For example, business users may only have access to "high-level" data and are required to contact operations staff in order to gain detailed data. It would be much more efficient if the business users could access the data (and its context) themselves. A single IT solution should be able to give both business and operations users access to all information (data and data context), with appropriate security features such that data is available as "read-only" or on a "need to know basis" as appropriate while giving "full drill-down" capability.

An Opportunity

There is an opportunity to fundamentally change the quality of management information by changing the culture of decision-making to reinforce collaborative behaviour. Extending this change across the entire organisation, cultural change can become the catalyst to deliver transformational IT change to produce high quality data.

A governance structure is required, to achieve success, with the following features:

  • a review across all current and proposed IT projects that looks at both "what is being delivered" but also "how it is being delivered" with a focus on architecture that recognises a separation between business, data and IT systems and is able to deliver quicker feedback loops to produce the high quality data.
  • a link of the low-level robust challenge of individual projects to top-level strategic decision-making about IT discretionary spend.
  • a process that takes decision making out of the traditional silos and reinforces collaborative behaviour.

This will facilitate more effective IT investment decisions that will enable future IT infrastructure to deliver high quality information, be cost effective, and be adaptable to change under both normal and stressed conditions. This is especially required in the current business environment given the budget imbalance where demand exceeds supply due to lower revenues, increased regulation and the need for cost savings.

Stream Financial Approach

At Stream Financial, we have worked with clients to develop a framework to support the governance of discretionary IT development budgets to achieve the long-term goal of high quality information. The framework provides a robust, independent and objective review of current infrastructure as well as proposed projects and produces a consistent set of management governance information to support both horizontal review across organisational boundaries and objective decision-making at all levels of the organisational hierarchy.

The framework is based upon a set of principles, which lead to a high-level long-term business architecture. This is then used to derive a project review score-card which is used to evaluate each proposed project to allow an organisation to focus resources on projects which deliver both short-term business benefit and long-term architecture benefit.

The primary benefit of this more structured approach to governance is improved confidence in the quality of the data used to support management decisions.

About the Author

  •    28 Marshalsea Road
         London
         SE1 1HF
         United Kingdom
  •   +44 (0)20 3627 2908
  •    info@stream-financial.com

About Stream Financial

We enjoy those intractable business problems you hate. We employ our experience and leading edge software solutions to help your organisation realise it’s full potential. Our approach is widely effective for business processes across all organisations and business sectors.


Read more

Request a Demo


Request a Demo