Matt Aslett's Analyst Perspectives

Improve Trust in Data with Master Data Management

Written by Matt Aslett | May 18, 2023 10:00:00 AM

Master data management may not attract the same level of excitement as fashionable topics such as DataOps or Data Platforms, but it remains one of the most significant aspects of an organization’s strategic approach to data management. Having trust in data is critical to the ability of an organization to make data-driven business decisions. Along with data quality, MDM enables organizations to ensure data is accurate, complete and consistent to fulfill operational business objectives.

While it is an established and mature sector of the market, MDM is also a primary focus for innovation in data management. The next generation of MDM products incorporates artificial intelligence and machine learning to facilitate improvement in operational efficiency and time-to-value from data-driven initiatives by automating approaches to mastering data that have traditionally been manual and time-consuming.

Organizations must be able to trust the data to deliver operational efficiency and analytics insight. Ensuring the integrity of data used for business decision-making can be difficult given that organizations have an increasing volume and range of data sources to contend with. This fuels investment in ongoing improvement. I assert that through 2025, 7 in 10 organizations will be engaged in data integrity initiatives to increase trust in data processes using data quality and master data management tools.

“Master data” is the term used for an organization’s foundational reference data. It provides an agreed list of entities that can be shared throughout the organization, including categories such as parties (customers or workers), places (addresses or regions) and things (products, assets, financial instruments). Much of this data will be internally generated and specific to the organization and its business relationships, but some comes from external sources and reflects international or national standards (such as country codes, industry classifications or stock symbols).

MDM is the practice of managing the organization’s master data. It encompasses processes such as data validation, matching and merging duplicate records and enriching data with related information. Another important component of MDM is data modeling, which documents the relationships between data elements. This results in the generation of a data catalog or enterprise glossary that can be shared across the organization as well as with partners and suppliers. MDM is an important aspect of a larger data governance strategy that includes policies and rules to govern accessing and editing master data.

More than 8 in 10 participants (82%) in Ventana Research’s Data Governance Benchmark Research are using MDM technologies for data governance, with more than one-half (51%) using them at least once a day. Almost three-quarters of those that use MDM for data governance (73%) are confident in their organization’s ability to govern and manage data across the business, compared to only 27% of those that do not use MDM for data governance.

While MDM as a discipline has been an important aspect of data management for decades, the tools and platforms used for MDM initiatives have evolved rapidly in recent years. MDM software was initially developed to target two key domains: customer data integration and product information management. These remain natural starting points for MDM initiatives. If organizations are unable to properly track customers, customer service and retention are likely to be negatively impacted, while cross- and upselling opportunities could also be missed. Similarly, if organizations cannot properly track the bills of materials for products, the ability to produce, market and sell those products will potentially be negatively impacted, along with product maintenance and customer engagement.

While some organizations still focus MDM efforts solely on customer or product data, this could undermine the broader purpose of MDM, which is to ensure smooth and efficient operations. Data-savvy organizations seek out MDM products with multi-domain capabilities, providing the functionality to address customer and product data alongside data about employees, assets, suppliers, locations and any other pertinent business data. Managing data from across multiple domains can be easier said than done, given the increasing range of data sources and formats as well as growing data volumes.

While MDM has traditionally involved complex manual processes and expert users, the use of AI/ML enables automation to improve efficiency and lower barriers to collaboration across domains. I assert that by 2026, more than three-quarters of organizations’ data management processes will be enhanced with artificial intelligence and machine learning to increase automation, accuracy, agility and speed.

Utilizing AI/ML in MDM software can make data more accessible and usable in several ways. For example, AI/ML can support personalization by identifying and providing access to information most likely to be relevant to a specific user and their role. AI/ML-guided authoring and assistance, including usage recommendations, can automate data profiling processes. Recommendations may also highlight related information from multiple domains in the data governance process.

The core processes involved in master data management can also be enhanced with AI/ML. Multiple matching algorithms combined with ML scoring capabilities can help improve accuracy, while AI/ML can also accelerate dynamic data classification, data profiling and in-line data enrichment. ML techniques can also be used to automatically identify missing or inaccurate relationships in data that might otherwise have been overlooked in manual processes. Examples include identifying whether individual customers are members of the same household or whether businesses are related entities. AI/ML can also be used to automatically identify rules for data quality, standardization, enrichment and matching based on previous processing outcomes as well as facilitating automated enforcement as data is processed.

These are not just theoretical examples of how AI/ML could be applied to MDM, but practical examples of how AI/ML is employed in the new generation of MDM products, lowering the barriers to successful adoption and accelerating time to value. MDM is not a new concept, but while it does not get the same attention as other aspects of data management and data operations, it is also a hotbed of innovation. I recommend that organizations looking to make more data-driven decisions evaluate the new breed of MDM products with a view to increasing trust in data and data management processes. Organizations with higher levels of confidence in data can move more quickly to make data-driven decisions, responding faster to worker and customer demands for more innovative, data-rich applications and personalized experiences and gaining competitive advantage.

Regards,

Matt Aslett