The rise in the availability, and ease of collection, of data has led to the era of Big Data. Large data sets are collected from multiple unreconciled sources and stored awaiting the analysis of data scientists and other professionals. Logical data models are scorned in favour of unstructured data sets that can be combined in multiple ways, limited only by the ingenuity of the analyst.
Whilst Big Data has been a boon to many, the question of how to handle increasingly large amounts of data continues to be a source of innovation and debate amongst, and within, many large organisations. This is particularly noticeable in the financial industry where regulations are tight and the mastering, stewardship and ownership of data needs to be governed transparently. Still, whilst data sets can be combined into Big Data systems and put to innovative use for competitive advantage, questions remain over who manages the original data sets and how those data sets are merged.
The search for an answer to these questions led several years ago to the rise of the Chief Data Officer (CDO). CDOs were typically given responsibility for the governance and management of data sets and asked to coordinate the large number of staff and systems used to master the data. This focus on regulation and efficiency has yielded significant benefits to the bottom line, but has also had some interesting side effects that now require a different approach. It is now time to expand the remit of the CDO.
One of the key outcomes of the traditional CDO remit has been a focus on the upstream process of data management. Whilst this is absolutely necessary, it has in some cases led to deprioritisation of the use of data downstream in the organisation. In order to recapture this focus, we need to revisit the business case for data management:
Revenue: Drive bottom-line growth by realising first-to-market opportunities and facilitating cross-selling between business lines.
Cost: Reduce costs through commoditisation of complex, structured products and correct classification of instruments to optimise capital allocation.
Regulatory: Comply with regulatory and lineage requirements, ensure prompt and accurate responses to regulatory requests.
STP: Reduce data manipulation and manual fixes front to back.
Risk: Increased business intelligence through consistent data to understand systemic risks and opportunities.
In order to achieve these aims, data services need to be focused on data consumption, use, manipulation and propagation throughout an organisation. This involves expanding the reach of CDOs in order to achieve a Consumer-Focused Data Service. Specifically, there are four key areas of remit that require expansion:
1 Data Governance
Owners, stewards and consumers should jointly manage the Enterprise Data Governance Framework to ensure data is fit for purpose. Most organisations only give consumers a minor role (if any) in data governance. However, active consumer participation is required to both understand the relative importance of the data being managed and to ensure engagement in these areas.
2 Standard Data
Data sets are, in reality, managed based on ownership and distributed in combination based on consumer requirements. The traditional view of data is that it is distributed in the same way that it is mastered. Hence instrument/product data is separated from market data. This puts the onus on the consumer to map between data sets and combine them to meet their business needs. A consumer-focused data service would look for frequently occurring consumer use cases that can be satisfied by combining data sets upstream and distributing the data to the consumer in the format it will be used.
3 Data Interfaces
Interfaces are standardised according to technology and data architecture principles. Data model characteristics with bank-wide benefits are identified and retained front-to-back in the firm. Within each data set, there will be a limited number of data elements that it is advantageous to retain. Typical banking examples are instrument identifiers or relationships between parties and accounts. If a bank uses the same instrument identifiers in every system and these are preserved in the message flows, the allocation of collateral in the finance function will be much more efficient as well as reducing the need for manual manipulation of incoming data. In order to achieve this and other benefits, the remit of the CDO needs to be expanded to encompass every data interface in the organisation.
4 Data Aggregation
Data aggregation services are governed centrally to remove duplication and misinterpretation of data. The fourth and final characteristic of the consumer-focused model pulls Big Data and other data aggregation systems into a cohesive centre across an organisation. Big Data is often a victim of its own success and becomes a “solution looking for a problem”. By combining data aggregation tools, defining the strategic sources of data and ensuring correct use, the CDO can ensure that the management information is of the highest quality and is generated in the most efficient way.
The consumer-focused data service represents a significant departure from traditional data management thinking. It is the next evolution in the role of the Chief Data Officer and connects the stewards of data with the bottom line growth of the organisation
Big Data – systems or approaches for combining disparate data sets and using them for innovative analysis. Traditional data capture involved determining the use cases up front and then creating an entity model to store the data. Big Data allows you to capture data without minimal entity model design and then apply multiple use cases as the business need occurs. Big Data is not the same as simply having a lot of data.
Chief Data Officer (CDO) – executive traditionally responsible for leading the governance and management of data sources.
Data Owner – the person or group who can authorise or deny access to certain data and is responsible for approving the processes by which the data is maintained.
Data Steward – the person or group responsible for creating and maintaining the data as authorised by the Data Owner.
Data Mastering – the processes or systems used to manage the data stewards’ workflows and store the data