Whether data are used for risk, marketing or other purposes, data matters are often seen as solely technical matters, with business teams focusing on the consumption of data for reporting, modelling, or operational purposes. While this assumption is correct regarding the infrastructure and ‘plumbing’ of data systems, a business interest in, and involvement with data and associated systems should be encouraged. Those who rely on data can be found in all parts of an organisation and they need confidence in what they are using.
What follows is especially relevant to data stores that are consolidated from operational systems, such as those used for analysis and reporting, where data are many steps removed from their sources. Because these stores are regularly used for business decision-making, quality and trustworthiness is essential.
When considering your organisation’s data and associated systems and processes, you can ask yourself three questions:
1. Does your business have confidence in both your data infrastructure and ‘plumbing’, and the data presented to them?
An internal assessment of data and systems against existing regulatory standards that relate to data (such as BCBS239, EBA RTS, and elements of PRA CP6/22) is a useful benchmark. It can determine where improvements can be made and confirm where best practice has already been adopted. This will improve business confidence by either establishing that data and systems are of good quality and/or that issues are known and will be addressed.
2. Is metadata available to all data users across your organisation?
A wealth of information about data is available, whether an integral part of a data system or as the hard-won knowledge of in-house SMEs. This information should be made available to all data users to facilitate use and ownership.
The ‘gold standard’ is to publish detailed metadata that fully describe the data, including source to target lineage and data quality metrics. Traditionally viewed as requiring significant time and resource to collect and maintain, technical developments have removed some of these obstacles. Making metadata collection and maintenance a fundamental part of data management requires additional effort, but this is easily offset by the benefits for data owners and users.
One of these benefits relates to regulators’ growing interest in data used by organisations. Occasionally, organisations are asked to provide evidence that data are accurate, whether this is for purposes such as IRB, IFRS9, or SARs under GDPR. Establishing source to target lineage allows this evidence to be gathered and informs an organisation’s data consumers.
3. Do your organisation’s data consumers own the data?
To own the data, senior executives, managers and analysts need to understand and trust the data. While senior users may not directly access detailed information about data, they will benefit from summary information on data quality, the confidence of analysts using data on their behalf and evidence that data issues can be promptly identified and addressed. Without these, it is unlikely that those who should own the data will accept the responsibility.
The above can be thought of as a cycle where benchmarking your data and systems leads to increased knowledge about data and systems, enhancing business involvement and driving improvements in quality and confidence.