Artificial intelligence (AI) and cloud are top trends impacting master data management (MDM) now and into the year ahead, as organisations turn their attention to data quality and management to support areas such as digital transformation and analytics.
While AI and cloud are certainly noteworthy trends in MDM, my advice − as always − is to ensure there's a solid business case for MDM as a first priority.
More South African organisations have turned their attention to getting data management right as a foundation for analytics and data science.
Reliable analytics is a solid business objective, but MDM is not just about analytics, it's also about operational efficiencies: helping you ‘make more money, spend less money, or stay out of jail’. It's about raising revenues, cutting costs and compliance.
AI may not yet have transformed MDM, but there’s a great deal of interest and a lot of progress being made in deploying AI to improve data governance, metadata management and MDM.
The application of AI to MDM is a natural progression that will accelerate rapidly in the short-term, and significantly enhance the value and efficiencies of MDM in organisations.
Reliable analytics is a solid business objective, but MDM is not just about analytics, it's also about operational efficiencies
As data volumes continue to grow and data sources become more and more diverse, organisations will find that traditional and manually configured MDM approaches are no longer sufficient to meet their rapidly-evolving data management needs.
This will require the evolution of MDM to include augmentation through AI and ML, and there are many aspects of MDM that can benefit from this.
As an example, automated identification of data elements based on actual content, and then instant application of pre-configured data quality validation and remediation rules pertaining to that data, without human developers first having to profile data and then manually map rules.
In addition to this, such AI algorithms continuously learn and improve, thus automatically reducing errors and enhancing overall data quality on a continuous improvement basis.
Another example of AI application to MDM would be the replacement of static and rigid data governance processes with adaptive governance frameworks that automatically evolve with changing needs. Overall, AI-driven MDM promises to be a powerful solution to modern data management challenges.
Cloud-based MDM has all the benefits of cloud models for other applications, such as rapid deployment, improved scalability, lower costs associated with on-premises IT infrastructure and operations, and improved disaster recovery, to name a few.
Other benefits provided specifically by cloud-based MDM could be the enablement of easier and greater collaboration across larger audiences (human or machine) spread over multiple divisions, organisations, geographic regions, industries or a combination thereof, together with the required seamless scalability that such complexity would demand.
Additionally, data governance for global organisations could be significantly enhanced by cloud-based MDM, as well as reference data management for such organisations, provided, for example, by a hierarchical MDM architecture supporting global, regional and local master and reference data management schemes.
Of course, whether it is practical to have a cloud-based MDM solution with data provider or consumer applications on-premises, in the cloud or a combination of both, would require well-considered architectural frameworks and thorough piloting, in particular around the inherent data integration requirements of MDM.
Another trend highlighted by some international industry stakeholders is data fabric architecture to provide a unified view of data across different systems and sources. This has been spoken about for some years, but is only now really starting to be considered seriously by some South African companies.
As organisations continue to explore the application of data fabric to their analytics, reporting and associated data integration needs, the question that arises is where MDM fits into a data fabric strategy, or vice versa.
It must be noted that, while there is some overlap between data fabric and MDM, they are different, sometimes subtly, in many regards. Both share a common theme of the need to share data, but the scope of data fabric is all data, while that of MDM is generally master data, and often also reference data.
Data fabric has a strong data integration focus, while integration for MDM is more a means to an end, with MDM having a stronger focus on data quality and a persisted, standardised best of breed system-of-record, than what data fabric has.
Ultimately, data fabric benefits from MDM being a supplier of harmonised, quality, master (and reference) data, enabling fabric to deliver improved reliability and much greater value when combined with other data, such as transactional data, from multiple sources. Therefore, a move to data fabric may turn more of a spotlight on MDM in future.
Share