Master Data Lifecycle

From Glitchdata
Jump to navigation Jump to search

Master Data Management is a complex field often involving multi-domains, multiple-stakeholders, multiple-systems. The approach to Master Data Management can be framed as part of a Data Lifecycle. This involves:

  • Acquisition/Creation of masterdata
  • Inspection/read/analysis of masterdata
  • Updates/Triage of masterdata
  • Destruction/Archive of masterdata

Inspection/Read/Analysis of Masterdata

This involves the aggregation and alignment of data from multiple sources. Using metadata, and data tools, it is possible to align and analyse it for Data Quality, and masterdata anomalies.

In particular, the business analysis of identities, and fraud is possible using various mechanisms like Search-Indexing, Network-Graph analytics, Duplicate scoring. This capability provides the first step towards Masterdata Triage.

Updates / Triage of Masterdata

Triage and reconciling Masterdata is significantly more complex. This involves Data Governance initiatives in establishing sources (and priorities) of truth, understanding triage process-flow, and the sequential updates that need to occur to successfully apply a reconciled Golden Record. These updates often tend to include source-data systems.

  • The complexity of applying changes to source-data systems involves trust, and approvals to provide a source-system update interface. The process (& timing) of change needs to be calibrated so that the Golden Record is fully established, old records are decommissioned, and end-to-end systems are functional.

Solutions to facilitate Golden Record updates can vary in approach. These could be:

  • Use of Robotic Process Automation (RPA) to apply updates via the GUI interface of source systems.
  • Use of back-end event processing mechanisms to REST API data interfaces and microservices.

The optimal method depends on multiple things like the maturity of the environment, available skillsets, and forward perspectives of the client. It is likely that RPA presents an optimal method with a “Manual Merge human-intervention step.

  • ServiceNow or ITIL-based service management can support Masterdata Triage by managing the service queue and service levels.
  • RPA is limited by automation. As “Manual Merge” is a probable requirement, it is unlikely that an RPA server/s would be used. RPA also presents additional cyber attack vectors.
  • So a Triage process would involve a combination of analytic tools, RPA process tools, and Service Management tools.
  • Back-end updates via say REST APIs can be challenging as it takes effort to build a catalog of suitable APIs.