Develop a Master Data Management Practice and Platform

Are you sure you have a master data problem?

Analyst Perspective

Executive Summary

Info-Tech Insight

Everybody has master data (e.g. customer, product) but not a master data problem (e.g. duplicate customers and products). MDM is complex in practice and requires investments in data governance, data architecture, and data strategy. Identifying business outcomes based on quality master data is essential before you pull the trigger on an MDM solution.

What is master data and master data management?

• Master data domains include the most important data assets of an organization. For this data to be used across an enterprise in consistent and value-added ways, the data must be properly managed. Some common master data entities include customer, product, and employees.
• Master data management (MDM) is the control over master data values to enable consistent, shared, contextual use across systems, of the most accurate, timely, and relevant version of truth about essential business entities (DAMA DMBOK).
• The fundamental objective of MDM is to enable the business to see one view of critical data elements across the organization.
• MDM systems will detect and declare relationships between data, resolve duplicate records, and make data available to the people, processes, and applications that need it. The end goal of an MDM implementation is to make sure your investment in MDM technology delivers the promised business results. By supplementing the technology with rules, guidelines, and standards around enterprise data you will ensure data continues to be synchronized across data sources on an ongoing basis.

Info-Tech’s Data Management Framework Adapted from DAMA-DMBOK and Advanced Knowledge Innovations Global Solutions. See Create a Data Management Roadmap blueprint for more information.

Why manage master data?

Master data drives practical insights that arise from key aspects of the business.

Your Challenge

Modern organizations have unprecedented data challenges.

• The volume of enterprise data is growing rapidly and comes from a wide variety of internal and external data sources (e.g. ERP, CRM). When data is located in different systems and applications, coupled with degradation and proliferation, this can lead to inaccurate, inconsistent, and redundant data being shared across departments within an organization.
• For example, customer information may not be identical in the customer service system, shipping system, and marketing management platform because of manual errors or different name usage (e.g. GE or General Electric) when input by different business units.
• Data kept in separate soiled sources can also result in poor stakeholder decision making and inefficient business processes. Some issues include:
  • The lack of clean customer list results in poor customer service.
  • Hindering good analytics and business predictions, such as incorrect supply chain decision when having duplicate product and vendor data between plants.
  • Creating cross-group consolidated reports from duplicate and inconsistent local data requires too much manual effort and resources.

On average, 25 different data sources are used for generating customer insights and engagement.

On average, 16 different technology applications are used to leverage customer data.

Source: Deloitte Digital, 2020

Common Obstacles

Finding a single source of truth throughout the organization can be difficult.

Changes in business process often come with challenges for CIOs and IT leaders. From an IT perspective, there are several common business operating models that can result in multiple sets of master data being created and held in various locations. Some examples could be:

• Integrate systems following corporate mergers and acquisitions
• Enterprise with multi-product line
• Multinational company or multi-geographic operations with various ERP systems
• Digital transformation projects such as omnichannel

In such situations, implementing an MDM solution helps achieve harmonization and synchronization of master data and provide a single, reliable, and precise view of the organization. However, MDM is a complex system that requires more than just a technical solution. An organization might experience the following pain points:

• Failure to identify master data problem and organization’s data needs.
• Conflicting viewpoints and definitions of data assets that should reside in MDM across business units.

Building a successful MDM initiative can be a large undertaking that takes some preparation before starting. Understanding the fundamental roles that data governance, data architecture, and data strategy play in MDM is essential before the implementation.

“Only 3 in 10 of respondents are completely confident in their company's ability to deliver a consistent omnichannel experience.”

Source: Dun & Bradstreet, 2022

Insight summary

Overarching insight

Everybody has master data (e.g. customer, product) but not a master data problem (e.g. duplicate customers and products). MDM is complex in practice and requires investments in data governance, data architecture, and data strategy. Figuring out what the organization needs out of its master data is essential before you pull the trigger on an MDM solution.

Phase 1 insight

A master data management solution will assist you in solving master data challenges if your organization is large or complex, such as a multinational corporation or a company with multiple product lines, with frequent mergers and acquisitions, or adopting a digital transformation strategy such as omnichannel.

Organizations often have trouble getting started because of the difficulty of agreeing on the definition of master data within the enterprise. Reference data is an easy place to find that common ground.

While the organization may have data that fits into more than one master data domain, it does not necessarily need to be mastered. Determine what master data entities your organization needs.

Although it is easy to get distracted by the technical aspects of the MDM project – such as extraction and consolidation rules – the true goal of MDM is to make sure that the consumers of master data (such as business units, sales) have access to consistent, relevant, and trusted shared data.

Phase 2 insight

An organization with activities such as mergers and acquisitions or multi-ERP systems poses a significant master data challenge. Prioritize your master data practice based on your organization’s ability to locate and maintain a single source of master data.

Leverage modern capabilities such as artificial intelligence or machine learning to support large and complex MDM deployments.

Blueprint Overview

Blueprint deliverables

Each step of this blueprint is accompanied by supporting deliverables to help you accomplish your goals:

1. MDM Readiness Assessment Tool	2. Business Needs Assessment Tool
3. Business Case Presentation Template	4. Project Charter Template
5. Architecture Design Template

Key deliverable:

6. MDM Practice Pattern Template

7. MDM Platform Template

Define the intentional relationships between the business and the master data through a well-thought-out master data platform and practice.

Measure the value of this blueprint

Refine the metrics for the overall Master Data Management Practice and Platform.

In phase 1 of this blueprint, we will help you establish the business context and master data needs.

In phase 2, we will help you document the current and target state of your organization and develop a practice and platform so that master data is well managed to deliver on those defined metrics.

Info-Tech offers various levels of support to best suit your needs

DIY Toolkit

“Our team has already made this critical project a priority, and we have the time and capability, but some guidance along the way would be helpful.”

Guided Implementation

“Our team knows that we need to fix a process, but we need assistance to determine where to focus. Some check-ins along the way would help keep us on track.”

Workshop

“We need to hit the ground running and get this project kicked off immediately. Our team has the ability to take this over once we get a framework and strategy in place.”

Consulting

“Our team does not have the time or the knowledge to take this project on. We need assistance through the entirety of this project.”

Diagnostics and consistent frameworks used throughout all four options

Guided Implementation

What does a typical GI on this topic look like?

A Guided Implementation (GI) is a series of calls with an Info-Tech analyst to help implement our best practices in your organization.

A typical GI is 8 to 12 calls over the course of 4 to 6 months.

Workshop Overview

Contact your account representative for more information.
workshops@infotech.com 1-888-670-8889

Phase 1: Build a Vision for MDM

Develop a Master Data Management Practice and Platform

Step 1.1

Assess Your Master Data Problem

Objectives

1. Build a solid foundation of knowledge surrounding MDM.

2. Recognize MDM problems that the organization faces in the areas of mergers and acquisitions, omnichannel, multi-product line, and multi-ERP setups.

This step involves the following participants:

CIO, CDO, or IT Executive

Head of Information Management

Outcomes of this step

An understanding of master data, MDM, and the prerequisites necessary to create an MDM program.

Determine if there is a need for MDM in the organization.

Understand your data – it’s not all transactional

Info-Tech analyzes the value of data through the lenses of its four distinct classes: Master, Transactional, Operational, and Reference.

Recognize the fundamental prerequisites for MDM before diving into more specific readiness requirements

Organizational buy-in

• Ensure there is someone actively invested and involved in the progress of the project. Having senior management support, especially in the form of an executive sponsor or champion, is necessary to approve MDM budgets and resourcing.
• MDM changes business processes and practices that affect many departments, groups, and people – this type of change may be disruptive so sponsorship from the top ensures your project will keep moving forward even during difficulties.
• Consider developing a cross-functional master data team involving stakeholders from management, IT, and the business units. This group can ensure that the MDM initiative is aligned with and supports larger organizational needs and everyone understands their role.

Understanding the existing data environment

• Knowing the state of an organization’s data architecture, and which data sources are linked to critical business processes, is essential before starting an MDM project.
• Identify the areas of data pain within your organization and establish the root cause. Determine what impact this is having on the business.

Before starting to look at technology solutions, make sure you have organizational buy-in and an understanding of the existing data environment. These two prerequisites are the foundation for MDM success.

Master data management provides opportunities to use data for analytical and operational purposes with greater accuracy

MDM can be approached in two ways: analytical and operational.

Think of it in the context of your own organization:

• How will MDM improve the ability for accurate data to be shared across business processes (Operational MDM)?
• How will MDM improve the quality of reports for management reporting and executive decision making (Analytical MDM)?

An investment in MDM will improve the opportunities for using the organization’s most valuable data assets, including opportunities like:

• Data is more easily shared across the organization’s environment with greater accuracy and trust.
• Multiple instances of the same data are consistent.
• MDM enables the ability to find the right data more quickly.

9.5% of revenue was at risk when bad experiences were offered to customers.

Source: Qualtrics XM Institute, 2022

Master data management drives better customer experience

85% In a survey of nearly 17,000 consumers and business buyers, 85% of customers expect consistent interactions across departments.

Source: Salesforce, 2022

Yet, 60% of customer say it generally feels like sales, service, and marketing teams do not share information.

Source: Salesforce, 2022

What is a business without the customer? Positive customer service experience drives customer retention, satisfaction, and revenue growth, and ultimately, determines the success of the organization. Effective MDM can improve customer experiences by providing consistent interactions and the ability to meet customer expectations.

61% of customers say they would switch to a competitor after just one bad customer service experience.

Source: Zendesk, 2022

Common business operating models or strategies with master data problems

Mergers and acquisitions (M&A)

M&A involves activities related to the consolidation of two companies. From IT’s perspective, whether the organization maintains different IT systems and applications in parallel or undergoes data integration process, it is common to have multiple instances of the same customer or product entity across different systems between companies, leading to incomplete, duplicate, and conflicting data sets. The organization may face challenges in both operational and analytical aspects. For many, the objective is to create a list of master data to have a single view of the organization.

Multiple-instance ERP or multinational organizations

Multiple-instance ERP solutions are commonly used by businesses that operate globally to accommodate each country’s needs or financial systems (Brightwork Research). With MDM, having a single source of truth could be a great advantage in certain business units to collaborate globally, such as sharing inventory coding systems to allow common identity and productive resource allocation and shared customer information for analytical purposes.

Common business operating models or strategies with master data problems (cont.)

Multiple product lines of business

An example for firms that sells multiple product lines could be Nike’s multiple product lines including footwear, clothing, and equipment. Keeping track of many product lines is a constant challenge for organizations in terms of inventory management, vendor database, and a tracking system. The ability to track and maintain your product data accurately and consistently is crucial for a successful supply chain (whether in a warehouse, distribution center, or retail office), which leads to improved customer satisfaction and increased sales.

Info-Tech Insight
A master data management solution will assist you in solving master data challenges if your organization is large or complex such as a multinational corporation or a company with multiple product lines, frequent mergers and acquisitions, or adopting a digital transformation strategy such as omnichannel.

Omni-channel

In e-commerce and retail industry, omnichannel means a business strategy that offers seamless shopping experiences across all channels, such as in-store, mobile, and online (Oracle). This also means the company needs to provide consistent information on orders, inventory, pricing, and promotions to customers and keep the customer records up to date. The challenges of omnichannel include having to synchronize data across channels and systems such as ERP, CRM, and social media. MDM becomes a solution for the success of an omnichannel strategy that refers to the same source of truth across business functions and channels.

Assess business model using Info-Tech’s MDM Readiness Assessment Tool

30 Minutes

• The MDM Readiness Assessment Tool will help you make the decision to stop the MDM project now or to continue on the path to MDM.
• Not all organizations need MDM. Don’t waste precious IT time and resources if your organization does not have a master data problem.

Download the MDM Readiness Assessment Tool

Step 1.2

Identify the Master Data Domains

Objectives

Determine which data domain contains the most critical master data in the organization for an MDM strategy.

This step involves the following participants:

Business Domain Representatives

Data Stewards or Data Owners

Information Management Team

Outcomes of this step

Determine the ideal data domain target for the organization based on where the business is experiencing the largest pains related to master data and where it will see the most benefit from MDM.

Reference data makes tackling master data easier

Reference data serves as a great starting place for an MDM project.

• Reference data is the simple lists of data that are typically static and help categorize other data using code tables. Examples include lists of countries or states, postal codes, general ledger charts of accounts, currencies, or product codes.
• Loading information into the warehouse or an MDM hub usually requires reconciling reference data from multiple sources. By getting reference data in order first, MDM will be easier to implement.
• Reference data also requires a relatively small investment with good returns so the value of the project can easily be demonstrated to stakeholders.
• One example of how reference data makes master data easier to tackle is a master list of an organization’s customers that needs an attribute of an address. By maintaining a list of postal codes or cities as reference data, this is made much easier to manage than simply allowing free text.

Info-Tech Insight

Organizations often have trouble getting started because of the difficulty of agreeing on the definition of master data within the enterprise. Reference data is an easy place to find that common ground.

There are several key considerations when defining which data is master data in the organization

A successful implementation of MDM depends on the careful selection of the data element to be mastered. As departments often have different interests, establishing a standard set of data elements can lead to a lot of discussion. When selecting what data should be considered master data, consider the following:

• Complexity. As the number of elements in a set increases, the likelihood that the data is master data also increases.
• Volatility. Master data tends to be less volatile. The more volatile data is, the more likely it is transactional data.
• Risk. The more likely data may have a risk associated with it, the more likely it should be managed with MDM.
• Value. The more valuable a data set is to the organization, the greater the chance it is master data.
• Sharing. If the data set is used in multiple systems, it likely should be managed with an MDM system.

Begin by documenting the existing data sources within the organization.

Use Info-Tech’s Master Data Management Business Needs Assessment Tool to determine master data sources.

Info-Tech Insight

While the organization may have data that fits into more than one master data domain, it does not necessarily need to be mastered. Determine what master data entities your organization needs.

Master data also fall into these four areas

More perspectives to consider and define which data is your master data.

Master data types can be divided into four main domains

Single Domain vs. Multi-Domain

• By focusing on a single master data domain, organizations can start with smaller, more manageable steps, rather than trying to tackle everything at once.
• MDM solutions can be domain-specific or be designed to support multiple domains.
• Multi-domain MDM is a solution that manages multiple types of master data in one repository. By implementing multi-domain from the beginning, an organization is better able to support growth across all dimensions and business units.

Use Info-Tech’s Master Data Management Business Needs Assessment Tool to determine master data priorities

2 hours

Use the Master Data Management Business Needs Assessment Tool to assist you in determining the master data domains present in your organization and the suggested domain(s) for your MDM solution.

Download the MDM Business Needs Assessment Tool

Step 1.3

Create a Strategic Vision for Your MDM Program

Objectives

1. Understand the true goal of MDM – ensuring that the needs of the master data users in the organization are fulfilled.

2. Create a plan to obtain organizational buy-in for the MDM initiative.

3. Organize and officialize your project by documenting key metrics, responsibilities, and goals for MDM.

This step involves the following participants:

CEO, CDO, or CIO

Business Domain Representatives

Information Management Team

Outcomes of this step

Obtain business buy-in and direction for the MDM initiative.

Create the critical foundation plans that will guide you in evaluating, planning, and implementing your immediate and long-term MDM goals.

MDM is not just IT’s responsibility

Make sure the whole organization is involved throughout the project.

• Master data is created for the organization as a whole, so get business input to ensure IT decisions fit with corporate goals and objectives.
• The ownership of master data is the responsibility of the business. IT is responsible for the MDM project’s technology, support, platforms, and infrastructure; however, the ownership of business rules and standards reside with the business.
• MDM requires IT and the business to form a partnership. While IT is responsible for the technical component, the business will be key in identifying master data.
• MDM belongs to the entire organization – not a specific department – and should be created with the needs of the whole organization in mind. As such, MDM needs to be aligned with company’s overall data strategy. Data strategy planning involves identifying and translating business objectives and capability goals into strategies for improving data usage by the business and enhancing the capabilities of MDM.

Keep the priorities of the users of master data at the forefront of your MDM initiative.

• To fully satisfy the needs of the users of master data, you have to know how the data is consumed. Information managers and architects must work with business teams to determine how organizational objectives are achieved by using master data.
• Steps to understanding the users of master data and their needs:

1. Identify and document the users of master data – some examples include business units such as marketing, sales, and innovation teams.
2. Interview those identified to understand how their strategic goals can be enabled by MDM. Determine their needs and expectations.
3. Determine how changes to the master data management strategy will bring about improvements to information sharing and increase the value of this critical asset.

Info-Tech Insight

Although it is easy to get distracted by the technical aspects of the MDM project – such as extraction and consolidation rules – the true goal of MDM is to make sure that the consumers of master data (such as business units, sales reps) have access to consistent, relevant, and trusted shared data.

Interview business stakeholders to understand how IT’s implementation of MDM will enable better business decisions

1 hours

Instructions

1. Identify which members of the business you would like to interview to gather an understanding of their current data issues and desired data usage. (Recommendation: Gather a diverse set of individuals to help build a broader and more holistic knowledge of data consumption wants or requirements.)
2. Prepare your interview questions.
3. Interview the identified members of the business.
4. Debrief and document results.

Tactical Tips

• Include members of your team to help heighten their knowledge of the business.
• Identify a team member to operate as the formal scribe.
• Keep the discussion as free flowing as possible; it will likely enable the business to share more. Don’t get defensive – one of the goals of the interviews is to open communication lines and identify opportunities for change, not create tension between IT and the business.

Info-Tech Insight

Prevent the interviews from being just a venue for the business to complain about data by opening the discussion of having them share current concerns and then focus the second half on what they would like to do with data and how they see master data assets supporting their strategic plans.

Ensure buy-in for the MDM project by aligning the MDM vision and the drivers of the organization

MDM exists to enable the success of the organization as a whole, not just as a technology venture. To be successful in the MDM initiative, IT must understand how MDM will help the critical aspects of the business. Likewise, the business must understand why it is important to them to ensure long-term support of the project.

“If an organization only wants to look at MDM as a tech project, it will likely be a failure. It takes a very strong business and IT partnership to make it happen.”

– Julie Hunt, Software Industry Analyst, Hub Designs Magazine

Use Info-Tech’s Master Data Management Business Case Presentation Template to help secure business buy-in

1-2 hours

Objectives

• This presentation should be used to help obtain momentum for the ongoing master data management initiative and continued IT- business collaboration.
• Master data management and the state of processes around data can be a sensitive business topic. To overcome issues of resistance from the operational or strategic levels, create a well-crafted business case.

Download the MDM Business Case Presentation Template

Use Info-Tech’s project charter to support your team in organizing their master data management plans

Use this master document to centralize the critical information regarding the objectives, staffing, timeline, budget, and expected outcome of the project.

1. MDM Vision and Mission

Overview

Define the value proposition behind addressing master data strategies and developing the organization's master data management practice.

Consider

Why is this project critical for the business?

Why should this project be done now, instead of delayed further down the road?

2. Goals or Objectives

Overview

Your goals and objectives should be practical and measurable. Goals and objectives should be mapped back to the reasons for MDM that we identified in the Executive Brief.

Example Objectives

Align the organization’s IT and business capabilities in MDM to the requirements of the organization’s business processes and the data that supports it.

3. Expected Outcomes

Overview

Master data management as a concept can change based on the organization and with definitions and expectations varying heavily for individuals. Ensure alignment at the outset of the project by outlining and attaining agreement on the expectations and expected outcomes (deliverables) of the project.

Recommended Outcomes

Outline of an action plan

Documented data strategies

4. Outline of Action Plan

Overview

Document the plans for your project in the associated sections of the project charter to align with the outcomes and deliverables associated with the project. Use the sample material in the charter and the “Develop Your Timeline for the MDM Project” section to support developing your project plans.

Recommended Project Scope

Align master data MDM plan with the business.

Document current and future architectural state of MDM.

Download the MDM Project Charter Template

5. Identify the Resourcing Requirements

Overview

Create a project team that has representation of both IT and the business (this will help improve alignment and downstream implementation planning).

Business Roles to Engage

Data owners (for subject area data)

Data stewards who are custodians of business data (related to subject areas evaluated)

Data scientists or other power users who are heavy consumers of data

IT Roles to Engage

Data architect(s)

Any data management professionals who are involved in modeling data, managing data assets, or supporting the systems in which the data resides.

Database administrators or data warehousing architects with a deep knowledge of data operations.

Individuals responsible for data governance.

Phase 2: Build the MDM Practice and Platform

Develop a Master Data Management Practice and Platform

Step 2.1

Document the Current Data State

Objectives

1. Understand roles that data strategy, data governance, and data architecture play in MDM.

2. Document the organization’s current data state for MDM.

This step involves the following participants:

Data Stewards or Data Custodians

Data or Enterprise Architect

Information Management Team

Outcomes of this step

Document the organization’s current data state, understanding the business processes and movement of data across the company.

Effective data governance will create the necessary roles and rules within the organization to support MDM

• A major success factor for MDM falls under data governance. If you don’t establish data governance early on, be prepared to face major obstacles throughout your project. Governance includes data definitions, data standards, access rights, and quality rules and ensures that MDM continues to offer value.
• Data governance involves an organizational committee or structure that defines the rules of how data is used and managed – rules around its quality, processes to remediate data errors, data sharing, managing data changes, and compliance with internal and external regulations.
• What is required for governance of master data? Defined roles, including data stewards and data owners, that will be responsible for creating the definitions relevant to master data assets.

For more information, see Info-Tech Research Group’s Establish Data Governance blueprint.

Ensure MDM success by defining roles that represent the essential high-level aspects of MDM

Regardless of the maturity of the organization or the type of MDM project being undertaken, all three representatives must be present and independent. Effective communication between them is also necessary.

The following roles need to be created and maintained for effective MDM:

Data Owners are accountable for:

• Data created and consumed.
• Ensuring adequate data risk management is in place.

Data Stewards are responsible for:

• The daily and routine care of all aspects of data systems.
• Supporting the user community.
• Collecting, collating, and evaluating issues and problems with data.
• Managing standard business definitions and metadata for critical data elements.

Another crucial aspect of implementing MDM governance is defining match rules for master data

• Matching, merging, and linking data from multiple systems about the same item, person, group, etc. attempts to remove redundancy, improve data quality, and provide information that is more comprehensive.
• Matching is performed by applying inference rules. Data cleansing tools and MDM applications often include matching engines used to match data.
  • Engines are dependent on clearly defined matching rules, including the acceptability of matches at different confidence levels.
• Despite best efforts, match decisions sometimes prove to be incorrect. It is essential to maintain the history of matches so that matches can be undone when they are discovered to be incorrect.
• Artificial intelligence (AI) for match and merge is also an option, where the AI engine can automatically identify duplicate master data records to create a golden record.

Match-Merge Rules vs. Match-Link Rules

Match-Merge Rules

• Match records and merge the data from these records into a single, unified, reconciled, and comprehensive record. If rules apply across data sources, create a single unique and comprehensive record in each database.
• Complex due to the need to identify so many possible circumstances, with different levels of confidence and trust placed on data values in different fields from different sources.
• Challenges include the operational complexity of reconciling the data and the cost of reversing the operation if there is a false merge.

Match-Link Rules

• Identify and cross-reference records that appear to relate to a master record without updating the content of the cross-referenced record.
• Easier to implement and much easier to reverse.
• Simple operation; acts on the cross-reference table and not the individual fields of the merged master data record, even though it may be more difficult to present comprehensive information from multiple records.

Data architecture will assist in producing an effective data integration model for the technology underlying MDM

Data quality is directly impacted by architecture.

• With an MDM architecture, access, replication, and flow of data are controlled, which increases data quality and consistency.
• Without an MDM architecture, master data occurs in application silos. This can cause redundant and inconsistent data.

Before designing the MDM architecture, consider:

• How the business is going to use the master data.
• Architectural style (this is often dependent on the existing IT architecture, but generally, organizations starting with MDM find a hub architecture easiest to work with).
• Where master data is entered, updated, and stored.
• Whether transactions should be processed as batch or real-time.
• What systems will contribute to the MDM system.
• Implementation style. This will help ensure the necessary applications have access to the master data.

“Having an architectural oversight and reference model is a very important step before implementing the MDM solutions.”

– Selwyn Samuel, Director of Enterprise Architecture

Document the organization’s data architecture to generate an accurate picture of the current data state

2-3 hours

Populate the template with your current organization's data components and the business flow that forms the architecture.

Think about the source of master data and what other systems will contribute to the MDM system.

Download the MDM Architecture Design Template ArchiMate file

Step 2.2

Document the Target Data State

Objectives

1. Understand four implementation styles for MDM deployments.

2. Document target MDM implementation systems.

This step involves the following participants:

Data Stewards or Data Custodians

Data or Enterprise Architect

Information Management Team

Outcomes of this step

Document the organization’s target architectural state surrounding MDM, identifying the specific MDM implementation style.

How the organization’s data flows through IT systems is a convenient way to define your MDM state

Understanding the data sources present in the organization and how the business organizes and uses this data is critical to implementing a successful MDM strategy.

Operational MDM

• As you manage data in an operational MDM system, the data gets integrated back into the systems that were the source of the data in the first place. The “best records” are created from a combination of data elements from systems that create relevant data (e.g. billing system, call center, reservation system) and then the data is sent back to the systems to update it to the best record. This includes both batch and real-time processing data.

Analytical MDM

• Generates “best records” the same way that operational MDM does. However, the data doesn’t go back to the systems that generated the data but rather to a repository for analytics, decision management, or reporting system purposes.

Discovery of master data is the same for both approaches, but the end use is very different.

The approaches are often combined by technologically mature organizations, but analytical MDM is generally more expensive due to increased complexity.

Central to an MDM program is the implementation of an architectural framework

Info-Tech Research Group’s Reference MDM Architecture uses a top-down approach.

A top-down approach shows the interdependent relationship between layers – one layer of functionality uses services provided by the layers below, and in turn, provides services to the layers above.

Info-Tech Research Group’s Reference MDM Architecture can meet the unique needs of different organizations

The MDM service layers that make up the hub are:

• Virtual Registry. The virtual registry is used to create a virtual view of the master data (this layer is not necessary for every MDM implementation).
• Interface Services. The interface services work directly with the transport method (e.g. Web Service, Pub/Sub, Batch/FTP).
• Rules Management. The rules management layer manages business rules and match rules set by the organization.
• Lifecycle Management. This layer is responsible for managing the master data lifecycle. This includes maintaining relationships across domains, modeling classification and hierarchies within the domains, helping with master data quality through profiling rules, deduplicating and merging data to create golden records, keeping authoring logs, etc.
• Base Services. The base services are responsible for managing all data (master, history, metadata, and reference) in the MDM hub.
• Security. Security is the base layer and is responsible for protecting all layers of the MDM hub.

An important architectural decision concerns where master data should live

All MDM architectures will contain a system of entry, a system of record, and in most cases, a system of reference. Collectively, these systems identify where master data is authored and updated and which databases will serve as the authoritative source of master data records.

Central to an MDM program is the implementation of an architectural framework

These styles are complementary and see increasing functionality; however, organizations do not need to start with consolidation.

Consolidation implementation style

Master data is created and stored in application systems and then placed in a centralized MDM hub that can be used for reference and reporting.

Advantages

• Prepares master data for enterprise data warehouse and reporting by matching/merging.
• Can serve as a basis for coexistence or transactional MDM.

Disadvantages

• Does not provide real-time reference because updates are sent to the MDM system in batch mode.
• New data requirements will need to be managed at the system of entry.

Registry implementation style

Master data is created and stored in applications. Key identifiers are then linked to the MDM system and used as reference for operational systems.

Advantages

• Quick to deploy.
• Can get a complete view of key master data identifiers when needed.
• Data is always current since it is accessed from the source systems.

Disadvantages

• Depends on clean data at the source system level.
• Can be complex to manage.
• Except for the identifiers persisting in the MDM system, all master data records remain in the applications, which means there is not a complete view of all master data records.

Coexistence implementation style

Master data is created and stored in existing systems and then synced with the MDM system to create an authoritative record of master data.

Advantages

• Easier to deploy workflows for collaborative authoring.
• Creates a complete view for each master data record.
• Increased master data quality.
• Allows for data harmonization across systems.
• Provides organizations with a central reference system.

Disadvantages

• Master data is altered in both the MDM system and source systems. Data may not be up to date until synchronization takes place.
• Higher deployment costs because all master data records must be harmonized.

Transactional implementation style

All master data records are materialized in the MDM system, which provides the organization with a single, complete source of master data at all times.

Advantages

• Functions as a system of record, providing complete, consistent, accurate, and up-to-date data.
• Provides a single location for updating and managing master data.

Disadvantages

• The implementation of this style may require changes to existing systems and business processes.
• This implementation style comes with increased cost and complexity.

All organizations are different; identify the architecture and implementation needs of your organization

Architecture is not static – it must be able to adapt to changing business needs.

• The implementation style an organization chooses is dependent on organizational factors such as the purpose of MDM and method of use.
• Some master data domains may require that you start with one implementation style and later graduate to another style while retaining the existing data model, metadata, and matching rules. Select a starting implementation style that will best suit the organization.
• Organizations with multi-domain master data may have to use multiple implementation styles. For example, data domain X may require the use of a registry implementation, while domain Y requires a coexistence implementation.

Document your target data state surrounding MDM

2-3 hours

Populate the template with your target organization’s data architecture.

Highlight new capabilities and components that MDM introduced based on MDM implementation style.

Step 2.3

Develop MDM Practice and Platform

Objectives

1. Review Info-Tech’s practice pattern and design your master data management practice.

2. Design your master data management platform.

3. Consider next steps for the MDM project.

This step involves the following participants:

Data Stewards or Data Custodians

Data or Enterprise Architect

Information Management Team

Outcomes of this step

Define the key services and outputs that must be delivered by establishing core capabilities, accountabilities, roles, and governance for the practice and platform.

What does a master data management practice pattern look like?

The master data management practice pattern describes the core capabilities, accountabilities, processes, and essential roles and the elements that provide oversight or governance of the practice, all of which are required to deliver on high-value services and deliverables or output for the organization.

Download the Master Data Management Practice Pattern Template ArchiMate File

Master data management data practice setup

• Define the practice lead’s accountabilities and responsibilities.
• Assign the practice lead.
• Design the practice, defining the details of the practice (including the core capabilities, accountabilities, processes, and essential roles; the elements that provide oversight or governance of the practice; and the practice’s services and deliverables or output for the organization).
• Define services and accountabilities:

1. Define deployment and engagement model
2. Define practice governance and metrics
3. Define processes and deliverables
4. Summarize capabilities
5. Use activity slide to assign the skills to the role

General approach to setting up data practices

Guidelines for designing and establishing your various data practices.

Understand master data management practice pattern

A master data management practice pattern includes key services and outputs that must be delivered by establishing core capabilities, accountabilities, roles, and governance for the practice.

Assumption:

The accountabilities and responsibilities for the master data management practice have been established and assigned to a practice lead.

1. Download and review Master Data Management Practice Pattern (Level 1 – Master Data Management Practice Pattern).
2. Review and update master data management processes for your organization.

Download the Master Data Management Practice Pattern Template ArchiMate File

Info-Tech Insight

An organization with heavy merger and acquisition activity poses a significant master data challenge. Prioritize your master data practice based on your organization’s ability to locate and maintain a single source of master data.

Initiate your one-time master data management practice setup

1. Ensure data governance committees are established.
2. Align master data management working group responsibilities with data governance committee.
3. Download and review Master Data Management Practice Pattern Setup (Level 1 – Master Data Management Practice Setup).
4. Start establishing your master data practice:

4.1 Define services and accountabilities

4.2 Define processes and deliverables by stakeholder

4.3 Design practice operating model

4.4 Perform skills inventory and design roles

4.5 Determine practice governance and metrics

4.6 Summarize practice capabilities

5. Define key master data management deliverable and processes.

Download and Update:

Process Template: MDM Conflict Resolution

MDM operating model

The operating model is a visualization of how MDM commonly operates and the value it brings to the organization. It illustrates the master data flow, which works from left to right, from source system to consumption layer. Another important component of the model is the business data glossary, which is part of your data governance plan, to define terminology and master data’s key characteristics across business units.

Choosing the appropriate technology capabilities

An MDM platform should include certain core technical capabilities:

• Master data hub: Functions as a system of reference, providing an authoritative source of data in read-only format to systems downstream.
• Data modeling: Ability to model complex relationships between internal application sources and other parties.
• Workflow management: Ability to support flexible and comprehensive workflow-based capabilities.
• Relationship and hierarchies: Ability to determine relationships and identify hierarchies within the same domain or across different domains of master data.
• Information quality: Ability to profile, cleanse, match, link, identify, and reconcile master data in different data sources to create and maintain the “golden record.”
• Loading, integration, synchronization: Ability to load data quality tools and integrate so there is a bidirectional flow of data. Enable data migration and updates that prevent duplicates within the incoming data and data found in the hub.
• Security: Ability to control access of MDM and the ability to report on activities. Ability to configure and manage different rules and visibilities.
• Ease of use: Including different user interfaces for technical and business roles.
• Scalability and high performance/high availability: Ability to expand or shrink depending on the business needs and maintain a high service level.

Other requirements may include:

• MDM solution that can handle multiple domains on a single set of technology and hardware.
• Offers a broad set of data integration connectors out of the box.
• Offers flexible deployments (on-premises, cloud, as-a-service).
• Supports all architectural implementation styles: registry, consolidation, coexistence, and transactional.
• Data governance tools: workflow and business process management (BPM) functionality to link data governance with operational MDM.
• Uses AI to automate MDM processes.

Info-Tech Research Group’s MDM platform

Info-Tech Research Group’s MDM platform summarizes an organization’s data environment and the technical capabilities that should be taken into consideration for your organization's MDM implementation.

Design your master data management platform

2-3 hours

Instructions

Download the Master Data Management Platform Template.

The platform is not static. Adapt the template to your own needs based on your target data state, required technical capabilities, and business use cases.

Download the MDM Platform Template

Next steps for the MDM project

There are several deployment options for MDM platforms; pick the one best suited to the organization’s business needs:

Capitalizing on trends in the MDM technology space would increase your competitive edge

AI improves master data management.

• With MDM technology improving every year, there are a greater number of options to choose from than ever before. AI is one of the hottest trends in MDM.
• By using machine learning (ML) techniques, AI can automate many activities surrounding MDM to ease manual processes and improve accuracy, such as automating master data profiling, managing workflow, identifying duplication, and suggesting match and merge proposals.
• Some other powerful applications include product categorization and hierarchical management. The product is assigned to the correct level of the category hierarchy based on the probability that a block of words in a product title or description belongs to product categories (Informatica, 2021).

Info-Tech Insight

Leverage modern capabilities such as AI and ML to support large and complex MDM deployments.

Informatica, 2021

Related Info-Tech Research

Build Your Data Quality Program

• Data needs to be good, but truly spectacular data may go unnoticed. Provide the right level of data quality, with the appropriate effort, for the correct usage. This blueprint will help you determine what “the right level of data quality” means and create a plan to achieve that goal for the business.

Build a Data Architecture Roadmap

• Optimizing data architecture requires a plan, not just a data model.

Create a Data Management Roadmap

• Streamline your data management program with our simplified framework.

Related Info-Tech Research

Build a Robust and Comprehensive Data Strategy

• Formulate a data strategy that stitches all of the pieces together to better position you to unlock the value in your data.

Build Your Data Practice and Platform

• The true value of data comes from defining intentional relationships between the business and the data through a well-thought-out data platform and practice.

Establish Data Governance

• Establish data trust and accountability with strong governance.

Research Authors and Contributors

Authors:

Contributors:

Bibliography

“DAMA Guide to the Data Management Body of Knowledge (DAMA-DMBOK Guide).” First Edition. DAMA International. 2009. Digital. April 2014.
“State of the Connected Customer, Fifth Edition.” Salesforce, 2022. Accessed Jan. 2023.
“The new digital edge: Rethinking strategy for the postpandemic era.” McKinsey & Company, 26 May. 2021. Assessed Dec. 2022.
“Value and resilience through better risk management.” Mckinsey & Company, 1 Oct. 2018. Assessed Dec. 2022.
“Plotting a course through turbulent times (9TH ANNUAL B2B SALES & MARKETING DATA REPORT)” Dun & Bradstreet, 2022. Assessed Jan. 2023.
““How to Win on Customer Experience.”, Deloitte Digital, 2020. Assessed Dec. 2022.
“CX Trends 2022.”, Zendesk, 2022. Assessed Jan. 2023
.”Global consumer trends to watch out for in 2023.” Qualtrics XM Institute, 8 Nov. 2022. Assessed Dec. 2022
“How to Understand Single Versus Multiple Software Instances.” Brightwork Research & Analysis, 24 Mar. 2021. Assessed Dec. 2022
“What is omnichannel?” Oracle. Assessed Dec. 2022
“How AI Improves Master Data Management (MDM).” Informatica, 30 May. 2021. Assessed Dec. 2022