In the course of the past years, the increased importance of digitization and digital transformation prompted almost every single business to adjust its data management strategy. Especially, financial institutions are induced to restructure their data governance to meet the new regulatory requirements and manage the exponentially growing amounts of data. Reorganizing corporate data management frameworks demands the overcoming of various functional and technical challenges. Successfully tackling these challenges comprises the improvement of quality and harmonization of data.
Data Trees as an instrument to develop data lineage
Tracing the data’s origins and its patterns of movement serve as ways to create transparency among an organization’s data records. Thus, the identification of errors back to the root cause is enabled. Regulatory authorities demand of financial institutions that reporting data is of high quality. To achieve this, a comprehensive and comprehensible business data lineage necessary.
One approach to trace back the origin of information is the setting up of data trees. They describe the lineage from a key figure to its data source. By achieving a broad data lineage, data trees allow financial institutions and other organizations to prioritize the most relevant data and create transparency of the origins of data.
Setting up Data Trees
A stepwise approach is needed: the source of the data is traced backwards in an organized manner beginning with analyzing the information given in the reports.
Initially, all data conveyed by the reports is prioritized. Information which is considered the most relevant is referred to as critical data items (CDIs). One of these CDIs such as economic capital, for instance, is picked to serve as a prototype for further steps. The data used to calculate the economic capital of a business is subsequently determined and the most critical information for this calculation is selected. Various criteria can be used to opt the most critical data items. For example, the data items which have the highest impact, or which are known for frequent data quality issues can be considered most critical. In the context of economic capital, market price risk constitutes a significant information for calculation. Therefore, it is selected as the most critical data item. To assure a transparent prioritization, the criteria have to be documented in detail. The remaining and uncritical data elements are parked and left for later consideration.
This process is continued until a sufficient level of detail is reached for the purpose of the prototype. It generates a comprehensive data tree, displaying the full data lineage of the most critical data items with their root values from the reports, their according subtrees comprising all critical data elements, and their source data. The data tree of economic capital shows its CDIs, namely market price risk, interest risk, sensitivities, and cash flows. It eventually indicates the master data and positions used as source data to calculate economic capital reports.
Depending on the demands resulting out of particular projects and data management roadmaps, the data tree model and the CDI universe can be extended.
Data Trees as an outcome out of the DMF methodology
Capgemini Invent develops Data Trees in the course of the implementation of its data management flywheel methodology. This structured approach allows a facilitated data sourcing and a consistent data collection process, by combining our best practices in different areas of data management.
The DMF implementation is a three-phases transformation journey, beginning with a pilot phase followed by a transition, then a business as usual phase. It allows the adaption to new data environments, the driving of data-based innovations, and the optimization of data management processes. In the pilot phase, maturity assessments serve as a foundation for the setup of data trees. Questionnaires allow to evaluate sourcing processes and the maturity of an organization’s data management. The DMF is embedded in these processes. With the help of a modern data dictionary tool creating the data trees is supported by user friendly capturing and transparent rendering. This way, the data trees become part of the Data Management Flywheel and data dictionary tool set.