Data lineage is critical to regulatory compliance, cost-effective data management, and arming the business with accurate and timely information on which to make decisions. It is also challenging to implement, difficult to sustain and often suffers from a lack of management buy-in and funding.
A recent webinar hosted by A-Team Group on how to get data lineage right, explored the challenges and described approaches to successful implementation. While an early audience poll suggested data lineage is not yet mature, with 38% of respondents starting to build a solution, 28% in the planning stage and just 13% with a complete solution, the benefits of data lineage are enticing, with 58% of respondents to another poll question expecting to gain significant business benefits and 42% significant operational benefits.
Commenting on the progress of data lineage adoption, Dan Bucosky, a data quality and governance consultant at a Tier 1 institution, said: “Large institutions are well aware of the need for data lineage and many are using metadata to collect required data. I have also seen firms using Excel, but this is not the best approach.”
Regulatory drivers behind adoption include BCBS 239, Dodd-Frank, Markets in Financial Instruments Directive II (MiFID II) and, to a limited extent, Fundamental Review of the Trading Book (FRTB) regulation. Data lineage can also reduce the risks of non-compliance with General Data Protection Regulation (GDPR). Considering these regulatory requirements, Philip Miller, co-founder of Solidatus, commented: “Not having data lineage for regulatory compliance could be hugely expensive.”
That said, getting data lineage right comes with numerous challenges. Stephen Veasey, CEO at 3d Innovations, noted problems around managing large volumes of data, legacy environments, disparate systems, mixed data formats, data quality, data ownership and extracting useful information. Responsibility for these challenges and solutions, he said, should lie with the chief data officer. Bucosky suggested that with components of lineage in various parts of large organisations, a centralised metadata group could provide expertise and guidance.
Vendor solutions are available to support data lineage – an audience poll showed 42% of respondents using a mix of in-house and vendor solutions – and provide automation. Miller noted: “Most firms are retrofitting data lineage, but with good data access they should be able to automate about 70% of the process.”
To achieve this, data governance is a necessary partner to lineage, providing an underlying understanding of data for both technology and business teams, and ensuring data is used appropriately. Business buy-in is best achieved by demonstrating the benefits of applying lineage to an important business project, gaining a beachhead and moving on to develop a broader and sustainable solution. As Bucosky concluded: “Understand business drivers and start small with a focussed dataset.”