Master data management is all about solving business issues and improving data trustworthiness through the effective and seamless integration of information with business processes.

Unfortunately, a common mistake that some organizations make is to treat MDM as a technical issue. While this approach helps an organization quick start its MDM initiative, it leaves most critical problems unattended, and dilutes the overall benefits of the MDM program.

A technology-driven approach decreases business confidence in MDM, making it difficult to sustain the solution, thus causing the premature death of the global MDM program.

However, a technical solution well integrated with business processes, along with a strong governance program, is the right way to start an MDM program. A business driven approach can ensure the success of MDM program and enable a path for further expansion.

Here are 10 best practices for MDM implementation:

1. Establish a business case

It is easy to get executive buy-in if key corporate goals are linked to the MDM project through a business case. You need to define a strong business case to get sponsorship and funding from senior managements.

As an example, a manufacturing firm in North America is in the process of transforming their business applications to improve upon the “order-to-cash” cycle. During the initial due-diligence it has been discovered that an integrated view of vendor, supplier, dealer and product is required to support the ERP application to improve upon the “order-to-cash”. The MDM program was initiated with complete executive support.

2. Get executive sponsorship

Most MDM projects are part of enterprise level data management initiatives that are driven with critical business priorities. MDM should be positioned as an enabler of the key business activities, and it should be explained to potential sponsors as to how MDM can help them, and they should be brought on board as supporters.

A strong executive sponsor will help get the right funding and support. An executive backup can help you to deal with internal politics and conflicts. Typically, a high-ranking executive needs to communicate the MDM business case, and the ongoing matrix of success as the program rolls out.

As another example, a leading gaming major from US embarked an enterprise business transformation initiative. The key processes that were identified included game life cycle management, CRM, plan to fulfill, procure to pay, plan to report ,hire to retire and business intelligence. MDM was identified as a key enabler for the processes that included product, employee and customer master data. The goals, strategic imperatives and expected outcomes were defined along with the executive sponsor for each initiative.

3. Get business Involved

MDM is all about solving business issues by effective management of master data that is critical to business operations. MDM should be driven by business needs for otherwise it will be turn out to be another isolated master data mart.

Active involvement of business groups improves the overall success and usability of the MDM Program. Various organizational groups and business lines should be brought into the loop right at the beginning of the MDM program. MDM project should therefore be jointly owned by business units and IT department.

For example, as a first step towards MDM, a manufacturing major in Europe has initiated a master data discovery and application inventory program. All the global business process owners were briefed of the initiative during the kick-off. The single point of contact for the manufacturing and regional sites was identified. The list of application and owners per site were collected and additional information was requested. For global applications spanning multiple sites/regions, the application owners were briefed. Site visits for few production plants were made further to create awareness and collect data.

4. Invest sufficient time in planning and evaluation

MDM implementation is more complex than people realize. Investing sufficient time in planning and evaluation is vital as this helps getting the bigger picture of the master data landscape.

Start with the MDM Blue Print definition, a thorough understanding of the commonality and divergence of the master information is critical success factor for MDM program. A feasibility and option analysis for MDM solution should be performed. A “Build Vs Buy” analysis is recommended.

It is also recommended you perform a detail tool evaluation and proof-of-concept (POC). It is good to talk to reference customers and learn from their mistakes.

As an example, a manufacturing firm in North America embarked an internal six-sigma project to select the right MDM tool. Post tool selection, they executed a poof-of-concept (POC) project taking one of their business subject area.

5. Institute MDM governance and stewardship

Do not underestimate the complexities and politics involved in executing MDM project. There may be strong resistance from some of the groups in releasing data ownership. It may be hard to motivate individual groups to support the centralized MDM initiative.

It is advisable to define and establish the MDM governance committee and stewardship well in advance before the start of MDM project. The governance committee helps increase the acceptance of MDM program with business groups, and MDM stewards help in the efficient execution of the day-to-day MDM activities.

6. Adopt the right topology and architecture

MDM is unique to every organization, as it is a collection of processes, technology and governance practices specific to the enterprise. Organizations should adopt the right architecture for MDM environment to support the long term functional and non-functional requirements.

The architecture should define to accommodate the data coming from big data Sources. Parameters like “Integration point between MDM and other sources”, “Management of Master Data from different external & internal sources”, “Definition and classification of Master data”, “Process of unstructured and semi-structured Master Data”, “system performance”, “data growth”, “concurrence usages”, etc should be considered to define the topology of the MDM environment.

It is good to be cautious as real-time data integration strategy may add complexities to the architecture. It is therefore best to adopt real-time architecture only if necessary. It is recommended that you define a flexible data model and process architecture to adopt the future changes without much of re-work. It is suggested that SOA architecture be adopted with the re-usable components strategy.

7. Define the data quality strategy

A quality data is a prerequisite for the success of MDM. One of the key expectations from the MDM program is to improve data transparency and provide a single version of truth of master information. Organizations should define the data quality strategy to correct and enhance the quality of master data before it gets incorporated into the MDM Hub. It is suggested that detail analysis and assessment of the enterprise data be performed before embarking on an MDM program.

As an example, a heavy equipment manufacturing firm in Europe embarked on a data discovery and analysis program to analyze the health and quality of the master information. The data discovery findings are to be considered to define the roadmap and strategy for the MDM program.


8. Get the right staffing and SI partner - Carefully chose the team

MDM is a relatively new technology with a shortage of experienced resources in the market. Carefully select the SI partner and create a team mix of SI consultant and internal staff. The team should constitute of the mix of business and technical experts. Tools and technology training should be organized for the team as and when needed.

Sub-contacting key staff from vendors during the initial phase of implementation may be needed. However, it is recommended that you build a Master Data Competency Center (MDCC) to support the long term multi-generation MDM program.

MDM being a niche technology, it commands premium rates for MDM consultant so the budget has to be planned accordingly. The team should have the complete attention of the HR department, to avoid any staff attrition and poaching. To keep the team motivated, special allowances, bonuses and recognition schemes should be taken into account.

9. Adopt phased implementation approach-Think big, build step-by-step

Organizations should define the blueprint of enterprise-wide MDM, but limit the scope of initial implementation. An iterative or spiral implementation approach is suggested over the waterfall model.

MDM implementation should be grouped into small and logical projects and this grouping can be done based on subject area, business unit, geography etc.

It is important to define the long term master data management vision, and outline the activities in tune to achieving the long term goal. Define each implementation piece of puzzle rightly and avoid building a stovepipe MDM application.

10. Start with a quick win project first- Show a quick ROI realization

It is important to start by first identifying the common and critical business problems, and knowing which one is to be tackled first. Executive sponsors and business owners will be eager to see the ROI from the MDM implementation, and may get impatient.

A quick win project with tangible ROI should be identified, and implemented first. The first MDM implementation should be quick and easy be completed within a six month time frame and provide tangible ROI and significant business value.
By adopting this strategy it is possible to ensure the success of the initial MDM implementation, and pave the way for further expansion.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access

Sanjay Kumar

Sanjay Kumar

Sanjay Kumar is a data scientist and freelance consultant. He has over 20 years of experience in data and analytics, with specialized skills in information governance, master data management, reference data management, data quality, metadata management, business intelligence and data warehousing.