8-minute read

Quick summary: How master data management (MDM) offers a solution to data chaos, and how to start implementing it in your organization

In many companies, diving into legacy data can be a cross between a needle-in-a-haystack hunt and a trip through a funhouse mirror maze: it’s difficult if not impossible to find the data you need, and when you do find it, there may be no way of ensuring that it’s clean, current, accurate, and compliant with data privacy laws and other regulations.

As a result, these organizations are hampered in their ability to leverage data in making informed business decisions. Even answering basic questions such as “Which customer generates the most revenue?” or “Which products are the most profitable?” becomes problematic at best.

Fortunately, there is a solution: master data management, under which every element related to data—infrastructure, processes, and people—is analyzed, organized, and continuously managed according to established best practices.

How data misalignment happens

If you’ve ever looked at your misaligned data and wondered “How did we get here?” you’re not alone. Over time, ad hoc data creation, mergers and acquisitions, the emergence of data silos, and lack of guidelines for business users can add up to data chaos for organizations of any size.

Take the journey of a B2B customer as an example:

 At the front end of the journey is the CRM platform. One customer, one record—so far, so good.

• As the customer journey proceeds, more and more systems are added, each with its own data formats, and data misalignments begin to pop up. One-off records are created. Divisions or buying centers of the same company are entered as separate customers. Updates to a customer record in one system fail to make their way to others.

• Over time, the organization may lose the ability to make the connection between a customer and their activity—or to identify the customer altogether.

The costs of data misalignment (and why it’s so hard to fix)

Misaligned data is much more than just a nuisance. It has far-reaching repercussions that can impact an organization’s bottom line, including

• Misinformed or uninformed decision making

• Ineffective marketing

• Low customer satisfaction

• Risk of inadvertently violating data privacy laws

• Loss of competitive advantages

• Missed strategic opportunities

 

The reason why data misalignment is so hard to fix is the complex nature of the data ecosystem. Data is continuously being gathered, created, stored, and used by multiple systems, each with its own distinct methods for handling information. Integrating systems alone will not solve the problem (and could make it worse). To resolve data misalignments, you have to roll up your sleeves and dig into the data itself. 

Why MDM is the solution

First, let’s review a few definitions. Master data is the business data that an organization uses across multiple processes, systems, applications. Master data management involves the synchronization and continuous governance of the company’s most critical data, particularly information concerning customers or products.

When MDM is successful, each customer or product has a golden record, a single, comprehensive record that is continuously managed to ensure accuracy and reliability. This resource gives the whole organization a single resource that all systems can access with total confidence, knowing that the data they are using is complete, accurate, current, and compliant.

One common mistake we see in organizations is equating a “single source of truth” with a true golden record. The two are not synonymous, as records must go through a rigorous data governance process to ensure unquestionable quality and accuracy to become “golden.”

As you can imagine, successfully implementing master data management involves a considerable investment of time, effort, and resources; however, companies who have made the leap enjoy considerable benefits:

Benefit 1: More reliable data

With a single source of truth to rely on, users across the organization are able to move forward with data-driven initiatives such as machine learning, automation, using analytics for informed decision making, and building data models to prepare for new opportunities and challenges.

Benefit 2: Optimized data processes reduce costs, improve productivity, and reduce risk

With the most current, accurate data consolidated in one place, users spend less time hunting for and validating the data they need, enabling processes to be streamlined. Teams can accomplish more with less effort and fewer resources, and the risk of error due to faulty data is dramatically reduced.

Benefit 3: Improved customer experience and increased loyalty

Master data management enables customer service teams to respond to customer requests more quickly with data that is complete, accurate, and easy to locate. Businesses can also tie data sources together to create a 360-degree view of the customer, enabling them to better understand and respond to customer needs.

Benefit 4: Improved readiness for data privacy laws and other regulations

Now that GDPR, CCPA, and other data privacy regulations are fixtures in the global business landscape, protecting personal information is no longer optional. Master data management clears the path for successful data privacy program by presenting a clear picture of what data the company has, where it’s located, who has access to it, and how it is being protected against breaches and misuse. In addition, customers are more likely to remain loyal to organizations that treat their data with respect. According to a recent survey, 75 percent of consumers will not buy a product from a company if they don’t trust it to protect their data.

5 steps to successful master data management

Step 1: Build a governance program

The best time to think about data governance is before starting work on an MDM system. Well-planned governance helps ensure data accuracy and reliability on a continuous basis and creates a framework of accountability for data users across the organization. Governance programs should be sponsored by the organization’s senior leadership to ensure buy-in at all levels.

Step 2: Assess the current data ecosystem

As we explored in a recent article, any successful journey must begin with knowing where you stand today. Determine what data you have, where it’s located, and who has access to it. You’ll also want to evaluate which applications are involved, either actively (creating the data) or passively (consuming or reporting from the data).

Step 3: Establish definitions

Believe it or not, discrepancies over a question as simple as “What is a customer?” can become a root cause of data havoc. For example, the shipping department may consider each address as a separate customer, and some lines of business may define a customer as a buyer that has made a purchase in the last year (or five years, or ten years …). Gaining consensus on basic definitions lets you establish the defining data points that form the foundation of a master data model.

Step 4: Prepare the data

This is the step where we dig into the data itself, starting with removing old or inaccurate data and eliminating duplicates. At this stage it helps to nominate a data steward for each master data domain—someone who is deeply familiar with the Customer/Purchaser, Provider, or Product—who can make judgment calls when records reflect conflicting information.

Once the data is cleaned up, implementing data classification and data taxonomy will start adding a level of organization that will benefit future steps.

Data classification involves tagging data in a way that indicates its level of sensitivity and the extent to which it should be protected. Most data classification systems work on three levels: low (public), medium (internal-use only), and high (extremely sensitive). This will help ensure that security and privacy measures are applied in a way that aligns with data sensitivity.

Data taxonomy involves tagging and sorting data into practical categories and sub-categories, which facilitates the organization of data so that it’s easy to find and easy to manage. You may consider, for example, data subject area, subject, topic, or entity as part of your data taxonomy. This will help you standardize categories and sub-categories, which is critical for enterprise data modeling.

Step 5: Implement MDM

Approaches to implementing master data management vary depending on the organization, the nature of the data, and the specific goals of the implementation. The most important thing is to make sure your MDM architecture aligns with your long-term data architecture strategy.

Multidomain MDM: the next frontier for cross-functional organizations

Historically, organizations have implemented master data management one domain in isolation (e.g. customer data, product data, supplier data, etc.) at a time. Their data, however, is becoming more connected and more interdependent with every passing day. For example, information about customer preferences is a vital asset for guiding product development, which in turn impacts supplier relationships. So how do you manage various different types of master data in a single, well-managed repository? Enter multidomain MDM.

Multidomain MDM applies the principles of master data management across domains to create a centralized platform. The organization has a single, shared view of data across functions, enabling more robust analytics, easier collaboration across domains, and higher overall ROI on their data investments.

Tapping into the full power of data

Data can be one of a business’ most powerful strategic assets—if it’s set up to be used strategically. The journey to implementing multidomain MDM successfully requires serious investments of time, effort, and resources. But once complete, with a solid governance strategy to maintain it, your MDM implementation will yield an ample return on all your data investments—both today and for years to come.

Like what you see?

Paul Lee

Senior Architect Michael Ly has over 20 years of experience helping customers transform their businesses by identifying the core issues of their data problems, connecting the dots, and developing solutions.

Author