In a nutshell, data lineage provides a two-dimensional roadmap which helps the management to ensure the authenticity of data by understanding how the data is produced and consumed in the organisation.
Building data lineage
According to Kumar, data lineage is built on metadata, which refers to data elements and attributes that businesses store in various applications. "You cannot build a data lineage if you don't have a metadata repository because you have to store all the elements in one place for us to build the lineage across all of those elements," he said.
However, establishing a metadata repository is no easy task and entails a high cost. "I can confidently say that only a few banks worldwide have a full-pledge metadata repository with end-to-end lineage built-in because it's very, very hard. Moreover, the cost [of building a metadata repository] can range anywhere from a quarter of a million dollars to more than that, depending on the [willingness of the bank to invest in it]. So what banks tend to do is to [build a metadata repository only for] their most critical processes like risk and compliance, exposures, capital risk," he shared.
Kumar added that establishing a "decent" size metadata repository will take about three to seven months, and another two months is needed to build in lineage.
Despite the challenges of building a metadata repository, data lineage can help banks save money such as by being able to trace data errors and identify discrepancies.
Assigning a data owner
Creating a data governance model also requires designating a data owner or a Chief Data Officer (CDO). According to Kumar, the CDO will bridge the gap between the business and IT, as well as determine which area of the business a particular data element belongs to. "The [CDO] is a very powerful function, and that function will talk to both the business and IT and bring them together for each data element," he explained.
While mature banks abroad already have CDOs, financial institutions in the Philippines have yet to assign one, he added. "[Philippine banks] need someone who can manage both the business and the IT entities, and start defining a landscape as to who owns what," said Kumar.
Driving data governance
According to Kumar, the data governance model is currently self-adopted by banking and financial institutions, and is not pervasive as it entails a high amount of investment. "As of today, that behaviour is self-adopted by banks that feel they have a gap to fill. However, it's not easy for banks to justify the investment required to have a holistic data governance model in place," he explained.
In line, Kumar noted that the adoption of data governance model will only be driven if there is a regulation in place.
Sign up for Computerworld eNewsletters.