Data governance is essential to managing not only regulatory requirements, but also client lifecycles, product innovation and cost and risk reduction programmes. While this is increasingly the case, developing data governance policies and processes that deliver tangible outcomes can be difficult. Dennis Slattery, CEO of EDMworks, set out the requirements for successful data governance – including collaboration, data ownership and senior management buy in – at last week’s A-Team Data Management Summit in London.
Slattery noted particular need for data governance to support regulatory compliance and good customer experience, but said this can only be achieved through changes to company culture, a shift from process driven data models to data ownership schemes, and the simplification of bank’s business models that, in the case of large banks, can include tens of thousands of systems.
He explained: “Large banks need to make cultural changes. To start, they need to ensure people put good information into front-end systems. Then they need to consider the pressure on people to satisfy regulations, improve client offerings and replace legacy systems with simpler systems. All these areas have interrelated datasets that must be managed, so the need is to set target states for the data environment and initiate change programmes while running the business as usual. A clear vision of data architecture is useful and must be communicated to everyone as success comes down to a culture that fosters collaboration on data and encourages people to work to a plan. Ultimately, this will deliver one view of each customer.”
Typically, an organisation will create a group policy on data governance that provides a framework in which particular aspects of the business can be prioritised, perhaps the customer experience, regulatory compliance or need for greater efficiency. Using a principles based approach to governance it is then possible to step through understanding data, not just entity data but also the data around it; data design, which will provide a vision of data architecture; and data integration, which will join up data sets, help to make data consistent and reduce numbers of legacy systems.
Data processes can then be brought into the governed environment and assessed using tools such as the Enterprise Data Management Council’s Data Management Capability Assessment Model. Sentiment based assessment is also important as people have different perspectives on processes. Assessments of data lineage and profiling can also be made and the analysis of data quality, a key element of data governance including data accuracy, completeness and timeliness, can begin.
With a clear understanding of the current state of processes and data, it is then possible to map information about the processes and gradually build up commonality of data. These steps generate an understanding of who has an interest in particular data and lead to the allocation of data ownership. Slattery explained: “Data ownership needs senior level commitment and oversight. It needs to include accountability for data in each link of an end-to-end process, data within business units and data across a bank.”
Considering the end game of data governance with tangible outcomes, Slattery concluded: “Once a data governance framework is in place, it is essential to communicate this to everyone across the bank and train everyone who will be part of the governance process. Successful data governance will make data owners proud and improve the business, but if it doesn’t work out, the dinosaur problem will appear and new companies will move into the market.”
