Chat with us, powered by LiveChat

The ugly truth about MDM Governance

Nikhil Bhatia

By Nikhil Bhatia

On August 18, 2016

Sometime in 500 B.C. Heraclitus stated “The Only Thing That Is Constant Is Change”. More than 2 millennia later this still holds true for everything that we do with management and governance of data in our organizations. Things are constantly changing, be it business objectives, business processes, governance policies and the data itself.

As a pure-play MDM innovator we have faced some questions from our customers such as:

  1. How can I empower my business users to change my data and governance policies to accommodate my changing business needs with minimal help from the IT organization or my application vendor?
  2. How do I enforce governance of both policies and data per my business relevance?
  3. If business relevance changes how do I quickly adapt my governance policies?
  4. What is the state of my data right now?
  5. How can I see what are the bottlenecks for my speed-to-market?
  6. How can I see the impact of changing governance policies on my processes and organization?

“Without effective governance, an MDM initiative will probably fail, so it is vital that program managers create the MDM governance framework early.”(1). It is an industry best practice to implement a set of data and policy governance principles while implementing an MDM system to achieve the best possible value for the enterprise. Most MDM systems will support setting up this governance with varied levels of complexity in implementation. The challenge that faces the MDM industry is that traditional data governance methodologies, processes and technologies are inherently very strict by nature. One of the reasons for this was the extremely high amount of importance put on quality of master data, so that rest of the organization can trust this data. All the focus was put on having a robust governance process, which manifested itself into a variety of implementation models and approaches, and even significant feature functions in MDM applications. There is a delicate balancing act that ends up being played between quality and velocity (or throughput), with no simple way to achieve both.

Say you are a retailer, whose business objective changes to become a marketplace as well, hence creating a need to increase product assortment by 1 Million items in 6 months. If it takes you even 10 minutes of total time to go through a specific workflow and set up an item from ingestion to go-live, it would take you approximately 27,777 person-hours or 3472 person-days to set up all the additional 1 Million items. If you want to onboard these items through the same governance process in 6 months, it would take a whopping 700 member team to do so. To put things in perspective Amazon’s product selection has expanded by 235 million in the past 16 months. That is an extraordinary average of 485 thousand new products per day.

Here are a few other examples that reflect the changing needs in some other industries. As a manufacturer who acquired another company, how easy (or difficult) will it be to consolidate the data of the combined assortment and syndicate to your existing customers? What if the FDA passed a new regulation which requires you as a food manufacturer to start capturing a whole new set of attributes. Can your MDM system allow for this change, incorporate these missing attribute data into new validations and report on it? If your customer acquisition process changes, what does it take for your MDM system to change your current workflow process?

“As firms become more insights-driven, the stakeholders in data governance expand and objectives become more complex.” (2) What I am trying to hit home is that without automation and without having relevant and insight-driven governance processes, no enterprise can achieve the true value of MDM Governance. If your business objectives or business processes change, you need a MDM system which business users can dynamically change to configure the governance that is relevant.

This brings me to my next important point. Although most MDM initiatives are business driven, the implementation and subsequent maintenance of MDM applications is considered an IT driven initiative. For a MDM system to be truly nimble and effective this dynamism in governance should be controllable by business users without much assistance from IT organizations. This will also require additional ability for the system to measure and report on impacts of these governance changes. Only then business users (and even IT users) would be able to truly fathom the consequences of their changes.

Key Takeaways

  • At no point in time should the focus deviate away from business objectives and goals. MDM implementations should strive to increase enterprise’s trust in the governance process and data, help business focus resources on the important and relevant things, and provide insights on what else should be or can be done.
  • It is not enough to just have a data governance program. The governance program and the resultant MDM implementation have to be future proof for any new curve ball future may bring upon the business.
  • Businesses should be able to pivot to incorporate any change that may happen to their business objectives, processes, organization structure and data in their MDM system. They should be able to do this without any/minimal help from the IT organization.
  • Business should be able to measure the impact of existing data and policy governance and also any changes to this governance structure to incorporate this feedback in setting up the governance correctly.
  • Business should re-look at the current processes and implementations to see how much of dynamic governance thought process is already incorporated, how Adaptive are the current Workflows. And finally design and establish an enhanced dynamic governance program.


  1. Gartner: The Seven Building Blocks of MDM: A Framework for Success; 02 August 2016, Analyst(s):Bill O’Kane | Michael Patrick Moran
  2. Forrester: Establish A Data Governance Journey Toward Data Citizenship; 27 April 2016, By Henry Peyret with Alex Cullen, Gene Leganza, Michele Goetz, Shaun McGovern, Diane Lynch