MDM: The walking dead without data governance

Understand how to master metadata in any organization today with proper master data management.

Table of contents:
· Introduction
· Challenges ahead of MDM projects
· Example: A Canadian Life Science customer
· Solution: The key role of data governance
· Conclusion

Introduction

We all have watched a fair share of zombie shows, like The Walking Dead, depicting zombies running around without any cognitive functional capacity but only knowing how to attack living humans. Unfortunately, that’s the reality of most failing master data management (MDM) projects that started without a brain akin to data governance.

Understanding how data governance and MDM complement each other requires understanding reference data. In the financial services industry, reference data covers a standard country, currency, numeric, and product code list. In the healthcare and life science industries, this could mean codes or codesets related to ICD 9, ICD10, specialty codes, medication, and more. Reference data is a standard set or list of industry-standard code values used across an organization for multiple purposes, mainly reporting and analytics.

Reference data is usually managed by the business and IT stakeholders to ensure it consistently and accurately represents core business entities of the organization, which could mean customers, products, suppliers, sites, hierarchies, charts of accounts, comp plans, job titles, and more. Proper reference data management is key to mastering data in any organization today, including benefits around business intelligence and reporting.

Challenges ahead of MDM projects

Most reference data is often siloed and spread across various IT architecture settings and applications; there is no holistic view showing the complex relationships between reference data codes and core business entities. A lack of a shared platform to collaborate on reference data presents unique challenges for users to navigate data and understand how different codes map to KPIs and reports.

In some instances, users manually hard code, i.e., link code values to business entities that end up in valuable reports, susceptible to breakage when changes are made to underlying systems, impacting the business decision-making process.

Example: A Canadian Life Science customer

I was recently engaged with a large Life Science customer in Canada; they present a unique data ecosystem comprising 50+ data domains covering customer, product, vendor, etc) and data was spread across multiple systems and applications.

There exists no common business language alignment, multiple codes representing the same concept. They lacked a formalized process for the creation and management of reference codes. For example, when a client tissue sample is collected, it is sent to multiple vendors for testing, and for each vendor and the same tissue sample, they had to generate a specific code for each vendor instead of working with just one standard code.

In addition, when new clients come for their services, they capture a “Service Date,” and when the client gets the testing results, a “Revenue Date” is captured in respective applications (e.g., client onboard and billing systems). The lack of a formal data governance process in place and getting alignment on the definition of “Service Date” vs. “Revenue Date” presented challenges.

The business intelligence and analytics team faced challenges in managing the “Customer Satisfaction Report”, which was generated after the client checkout. This was because each “date” field was captured in various systems and what it meant from a business perspective.

Solution: The key role of data governance

To resolve the inconsistencies in reference data and its relationship with core business entities (e.g., service data, revenue date, vendor codes, etc), the organization started by implementing a master data management (MDM) solution.

The idea was to rationalize reference data from across various IT architecture settings and create a single source of truth. Although most MDM solutions are good at pulling reference data into a single place, they fail to provide ongoing maintenance, control, and collaboration around evolving business changes. Many changes come through different source systems, newly added operational systems, applications, changing ETL jobs, or manual intervention.

As evidenced by the above customer story, MDM projects can be problematic without proper ownership and control, especially as organizations experience the volume of data doubling and tripling every 12–18 months. Changes made by users, systems, and applications can be hard to track and will lead to a week’s worth of remediation work to track changes made to various fields and fields that exist without proper business meaning. Manual monitoring could be a never-ending, daunting task.

Lesson learned from various organizations: Data governance presents the holy grail of making MDM projects a reality. I will go a bit further, and you can quote (or tweet!) me later: “There is no MDM without proper data governance”. Let me explain.

“Gartner estimates that as many as 85% of MDM projects fail.”

As outlined in the previous edition of the Data Workspace newsletter, Chief Data and Analytics Officers can add value to the MDM projects, wherein Data governance provides a formalized process to define stewardship and control around the data, which complements an MDM project. This way, an organization can establish a control point around reference data. Any changes around code values or code sets go through a structured process where the right people are involved when needed.

Conclusion

Governance technologies today offer a business-friendly front end for authoring and proposing codes/code sets, managing changes through approval workflows, and involving other users to collaborate, vote, and certify. This creates the trust element and, more importantly, a shared ownership framework, leading to the adoption of best practices, i.e., using well-defined codes/code sets vetted through approval by key stakeholders. Hence, a process of achieving reference data commonality is a key to success for MDM projects.

Kash Mehdi is an experienced and passionate leader working with the industry’s first and fastest-growing data knowledge governance catalog, DataGalaxy. Over the past decade, Kash has focused on scaling go-to-market functions for hyper-growth SaaS companies and guided organizations through the complexities of data orchestration to help them empower humans with trusted data.

Keep up with me on LinkedIn and Twitter, and be sure to check out my LinkedIn newsletter, Data Workspace for more expert articles.