What You Should Know About Master Data Management

Today’s guest blog post is from Benjamin Cutler of Winpure. In here Benjamin goes through a few things that you in a nutshell should know about master data management.


People have multiple phone numbers and multiple email addresses and in 2022 there must be several decades of historic contact information available for any one person. Most of us move at least once, every few years. Sometimes we go by different nicknames in different situations, some people even change their names. We hold different titles throughout the course of our careers and we change companies every few years. Only a few people in our lives know exactly how to get a hold of us, at any given time. Many of us change vehicles just as often as we change our hair color. Many of us are employees, most of us are also customers, many of us are spouses and sometimes we are grandparents, parents, aunts, uncles, and children at the same time. Sometimes we’re out enjoying ourselves and sometimes we just want to be left alone. We each have unique interests and desires, but we also have many things in common with other groups of people.


Products have many different descriptions, they come in many different variations, different sizes, different colors, and different packaging materials. Similar products are often manufactured by different manufacturers, and they can be purchased from many different commercial outlets, at different price points. Any one product on the market at any one time will likely be available in several variations, but that product will also likely change over time as the manufacturer makes improvements. Products can be purchased therefore they can also be sold. They can also be returned or resold to other buyers, so there are different conditions and ways to determine product value. There are SKU and UPC numbers and other official product identification and categorization systems including UNSPSC and others, but none of them speak the same language.


Companies are made up of many different people who come and go over time. The company may change names or change ownership. It may have multiple locations which means multiple addresses and phone numbers, and they probably offer many different ways to contact them. Depending on where you look, there are probably more than a dozen different ways to find their contact information, but only some of those company listings will be correct. Companies have tax IDs and Employer IDs and DUNS IDs in the US, and there are many different systems worldwide.


Addresses are the systems we use to identify locations. Each country and territory has its own system so each system is different. In the US we use premise numbers, street names with and without street prefixes and suffixes, we use unit numbers, states, counties, cities, towns and 5 and 9 digital numerical postal codes. Addresses and address systems can change over time, and they are inherently one of the most inconsistent forms of identification. Addresses are usually riddled with errors, misspellings, different structures and formatting, and they can be very difficult to work with. What makes this even more difficult is that the same address represented in multiple internal business systems will often be represented differently, and will rarely match the way the same address is represented externally.


Data is a digital description of all of these things. Data usually comes in columns and rows and all shapes and sizes. Data about these things is captured, stored in business systems and it’s used to get work done. Need to call a contact? Check your contact data. Need to know a company’s billing address? Check your company data. Need to know something about a product? Check your product information. Need to know something about where your customers live and work or where to deliver the product? Check your address information. But here’s the thing: the information rarely matches from system to system and it’s very hard to keep up to date. This is especially difficult for a few reasons. Internally your company probably has many different business systems and many different ways of storing and representing these things, so it rarely matches internally, plus, the way that your company stores and represents this information will almost never match external information. How can you know the best way to contact your customer who has multiple phone numbers and multiple email addresses? If you’re searching some external system for updated information about some product or contact and the information doesn’t match, how do you find the new information? How can you know if your own information is correct and up to date? How can you scale your efforts to communicate with hundreds or thousands of customers at a time, communicating information that is specifically relevant for each of them? If the information doesn’t match or is not correct, how can you know who is who?


The relationships across people, other groups of people, products, other groups of products, companies, other groups of companies, addresses, and other addresses, is where the rubber hits the road. Business value comes from connecting companies and products or services with other people and companies, and other products and services, at scale. Customers purchasing products might be interested in purchasing related products. Customers often buy things based on location. Companies selling to customers might be able to sell more, if they target similar customers in similar locations. Products and services also sell well based on location, and companies can optimize sales territories and delivery routes based on the relative proximity to other locations.

People and Technology

The people and technology between all of this, finds it difficult to keep up. People do things one by one and we’re good with ambiguity. We program computers and business systems to do things faster. Computers do things programmatically and very quickly but they’re not good with ambiguity. People can see similarity between things that are similar, but computers and business systems cannot. People might be good with troubleshooting and critical thinking, but computers and business systems are not. A computer program might be able to find the same customer in multiple systems and might be able to update that customer’s information all at once, but how can you know if the new information is the best information? Knowing that your customer probably has multiple phone numbers and multiple addresses and multiple nicknames, how can you know which information is correct? Doing this at scale can be very, very difficult.

In Conclusion

Master Data Management is very difficult but it’s fundamental in scaling your business. People can sell products door-to-door, but data and technology allow us to market, sell, deliver, and service our products and services, to tens and hundreds of thousands of people in milliseconds, regardless of the distance. Most organizations still view data as a cost of doing business but with the right investments in people, process, technology, and in data management, we can scale as worldwide organizations.

Nine Essential Master Data Processes in The Data Supply Chain

Master Data Management (MDM) and the overlapping Product Information Management (PIM) discipline is the centre of which the end-to-end data supply chain revolves around in enterprises that produce and/or sell goods.

Nine essential master data processes are:

1: Onboard Customer Data

It starts and ends with the King: The Customer. Your organization will probably have several touchpoints where customer data is captured. MDM was born out of the Customer Data Integration (CDI) discipline and a main reason of being for MDM is still to be a place where all customer data is gathered as exemplified in the post Direct Customers vs Indirect Customers.

2: Onboard Vendor Data

Every organization has vendors/suppliers who delivers direct and indirect products as office supplies, Maintenance, Repair and Operation (MRO) parts, raw materials, packing materials, resell products and services as well. As told in a post on this blog, you have to Know Your Supplier.

3: Enrich Party Data

There are good options for not having to collect all data about your customers and vendors yourself, as there are 3rd party sources available for enriching these data preferable as close to capture as possible. This topic was examined in the post Third-Party Data and MDM.

4: Onboard Product Data

While a small portion of product data for a small portion of product groups can be obtained via product data pools, the predominant way is to have product data coming in as second party data from each vendor/supplier. This process is elaborated in the post 4 Supplier Product Data Onboarding Scenarios.

5: Transform Product Data

As your organization probably do not use the same standard, taxonomy, and structure for product data as all your suppliers, you have to transform the data into your standard, taxonomy, and structure. You may do the onboarding and transformation in one go as pondered in the post The Role of Product Data Syndication in Interenterprise MDM.

6: Consolidate Product Data

If your organization produce products or you combine external and internal products and services in other ways you must consolidate the data describing your finished products and services.

7: Enrich Product Data

Besides the hard facts about the products and services you sell you must also apply competitive descriptions of the products and services that makes you stand out from the crowd and ensure that the customer will buy from you when looking for products and services for a given purpose of use.

8: Customize Product Data

Product data will optimally have to be tailored for a given geography, market and/or channel. This includes language and culture considerations and adhering to relevant regulations.

9: Personalize Product Data

Personalization is one step deeper than market and channel customization. Here you at point-of-sale seek to deliver the right Customer Experience (CX) by exercising Product eXperience Management (PXM). Here you combine customer data and product data. This quest was touched in the post What is Contextual MDM?

Digital Twins in the MDM Product Domain / PIM

When working with the product domain in Master Data Management (MDM) and with Product Information Management (PIM) we have traditionally been working with the product model meaning that we manage data about how a product that can be produced many times in exactly the same way and resulting in having exactly the same features. In other words, we are creating a digital twin of the product model.

As told in the post Spectre vs James Bond and the Unique Product Identifier the next level in product data management is working with each product instance meaning each produced thing that have a set of data attached that is unique to that thing. Such data can be:

  • Serial number or other identification as for example the Unique Device Identification (UDI) known in healthcare
  • Manufacturing date and time
  • Specific configuration
  • Current and historical position
  • Current and historical owner
  • Current and historical installer, maintainer and other caretaker
  • Produced sensor data if it is a smart device / machine

There is a substantial business potential in being better than your competitor in managing product instances. This boils down to that data is power – if you use the data.

When managing this data, we are building a digital twin of the product instance.

Maintaining that digital twin is a collaborative effort involving the manufacturer, the logistic service provider, the owner, the caretaker, and other roles. For that you need some degree of Interenterprise MDM.

Unified Customer MDM and Supplier MDM

When blueprinting a Master Data Management (MDM) solution one aspect is if – or in what degree – you should unify customer MDM and supplier MDM.

In theory, you should unify the concept for these two master domains in some degree. The reasons are:

  • There is always an overlap of the real-world entities that has both a customer and a supplier role to your organization. The overlap is often bigger than you think not at least if you include the overlap of company family trees that have members in one of these roles.
  • The basic master data for these master data domains are the same: Identification numbers, names, addresses, means of communication and more.
  • The third-party enrichment opportunities are the same. The most predominant possibilities are integration with business directories (as Dun & Bradstreet and national registries) and address validation (as Loqate and national postal services).

In practice, the problem is that the business case for customer MDM and supplier MDM may not be realized at the same time. So, one domain will typically be implemented before the other depending on your organization’s business model.

Solution Considerations

Most MDM solutions must coexist with an – or several – ERP solutions. Many popular enterprise grade ERP solutions have adapted the business partner view with a common data model for basic customer and supplier data. This is the case with SAP S/4HANA and for example the address book in Microsoft Dynamics AX and Oracle JD Edwards.

MDM solutions themselves does also provide for a common structure. If you model one domain before the other, it is imperative that you consider all business partner roles in that model.

Data Governance Considerations

A data governance framework may typically be rolled out one master data domain at the time or in parallel. It is here essential that the data policies, data standards and business glossary for basic customer master data and basic supplier master data is coordinated.

Business Case Considerations

The business case for customer MDM will be stronger if the joint advantages with supplier MDM is incorporated – and vice versa.

This includes improvement in customer/supplier engagement and the derived supply/value chain effectiveness, cost sharing of third-party data enrichment service expenses and shared gains in risk assessment.  

Available Solutions

Check the list of innovative solutions in the MDM space here.

Direct Customers vs Indirect Customers

When working with Master Data Management (MDM) for the customer master data domain one of the core aspects to be aware of is the union, intersection and difference between direct customers and indirect customers.

Direct customers are basically those customers that your organization invoice.

Indirect customers are those customers that buy your organizations products and services from a reseller (or marketplace). In that case the reseller is a direct customer to your organization.

The stretch from your organization via a reseller organization to a consumer is referred to as Business-to-Business-to-Consumer (B2B2C). This topic is told about in the post B2B2C in Data Management. If the end user of the product or service is another organization the stretch is referred to as Business-to-Business-to-Business (B2B2B).

The short stretch from your organization to a consumer is referred to as Direct-to-Consumer (D2C).

It does happen, that someone is both a direct customer and an indirect customer either over time and/or over various business scenarios.

IT Systems Involved

If we look at the typical IT systems involved here direct customers are managed in an ERP system where the invoicing takes place as part of the order-to-cash (O2C) main business process. Products and services sold through resellers are part of an order-to-cash process where the reseller place an order to you when their stock is low and pays you according to the contract between them and you. In ERP lingo, someone who pays you has an account receivable.

Typically, you will also handle the relationship and engagement with a direct customer in a CRM system. However, there are often direct customers where the relationship is purely administrative with no one from the salesforce involved. Therefore, these kinds of customers are sometimes not in the CRM system. They are purely an account receivable.

More and more organizations want to have a relationship with and engage with the end customer. Therefore, these indirect customers are managed in the CRM system as well typically where the salesforce is involved and increasingly also where digital sales services are applied. However, most often there will be some indirect customers not encompassed by the CRM system.

The Role of Master Data Management (MDM) in the context of customer master data is to be the single source for all customer data. So, MDM holds the union of customer master data from the ERP world and the CRM world.

An MDM platform also has the capability of encompassing other sources both internal ones and external ones. When utilized optimally, an MDM platform will be able to paint a picture of the entire sell-side space where your direct customers and indirect customers are.

Business Opportunities

Having this picture is of course only interesting if you can use it to obtain business value. Some of the opportunities are:

  • More targeted product and service development by having more insight into the whole sell-side space leading to growth advancements
  • Optimized orchestration of supply chain activities by having complete insight into the whole sell-side space and thereby fostering cost savings
  • Improved ability to analyse the consequences of market change and changes in the economic environment in geographies and industries covered leading to better risk management.

This list has a presentation of some of the most innovative MDM and adjacent solutions for getting a combined view of your direct customers and indirect customers. Check out the list here.