So far in this series, we discussed what “master data” is and the people and processes necessary to provide enterprise commerce and transformation initiatives with a solid foundation of data that will lead to success. As you know, the three keys of MDM are People, Processes, and Technology. So, this time, we will take a deep dive into the technology options upon which your MDM initiative will rely. And, how you might begin to determine which technologies are the best fit for your efforts based on your specific business requirements.
As a reminder, master data exists in multiple data domains within every enterprise, chief among them Persons (Customer, Patient, Citizen, Supplier / Vendor); Products, Locations, and Assets (Physical and Digital). Most MDM projects begin by addressing one or more of these domains (and in a later series, we will take a deep dive into each). Some organizations will plan for addressing multiple domains of data over the course of the project, but all will start by implementing the correct technology for one domain of data before they move on to the next domain.
Master Data Management (MDM)
While many companies, analysts, and thought leaders will use the term Master Data Management to describe a suite of potential offerings to address enterprise data needs, it’s becoming increasingly popular to refer to mastering “party” data (customers, patients, members, citizens, etc) as Master Data Management (MDM). Rarely, you might see a reference to Customer Data Integration (CDI), but that term isn’t nearly as popular as it was 10+ years ago. Do not confuse MDM for party data with technologies that pre-dated MDM and remain quite popular (such as CRM, ERP, Data Warehouse, or BI) but don’t address the same use cases as MDM.
To properly manage party data, there are four main actions that an MDM technology must address:
- Resolve – This action refers to the process of locating unique records about the person from systems across the enterprise and matching those records to create a single source of truth about that person, or “Golden Record.” This requires the capability to integrate with disparate source systems, such as billing or marketing systems, and may even include tough to integrate legacy systems. Most MDM platforms will offer multiple methods of data integration to address all use cases and, as part of the data integration process, most will provide the ability to transform and normalize key pieces of data upon load using simple ETL tools. The platform will then use a powerful algorithm to determine which records should be matched as the same person according to pre-set rules that can be probabilistic, deterministic, or a combination of both. Consultants, like those here at Amplifi, can help your organization tune the algorithm to ensure that the algorithm is matching and linking records most effectively and efficiently possible. This is important because the rest of the actions rely on a high-quality Golden Record.
- Relate – During this step of managing party data, an MDM platform will map the network of relationships within and between resolved customers. This can provide useful insights into your B2C and B2B customer base by visualizing households (B2C) and corporate hierarchies (B2B). This information can be used to make decisions about marketing, credit, market penetration, and more.
- Enhance – As you create a Golden Record from multiple source systems, you will likely begin to see a view of your customer that provides more detail than in any single source system. However, there may be additional information you may want to assign to each Golden Record from third-party reference data sets, credit agencies, and the like, to personalize the Golden Record for specific use cases.
- Extend – Once a Golden Record has been created, verified, and enhanced, your organization can now share those customer insights with any number of departments, business systems, social networks, or BI analytics tools.
Product Information Management (PIM)
Products are the cornerstone of your business – whether you are a retailer, distributor, manufacturer or somewhere in between. Managing the Information Lifecycle for rich, accurate content about products is key, and great care must be taken to ensure that the “supply-side” can receive and collaborate on product content creation to satisfy the “demand-side,” which is looking for accurate and engaging content on every channel.
Similar to managing party data, there are a few actions that are crucial for any PIM system to handle:
- Acquire – Just like MDM, a PIM system must acquire data from various source systems, such as ERP or PLM systems, and just like MDM systems, most PIM platforms offer multiple means of data integration and the ability to perform normalization techniques as data is loaded into the platform. One crucial difference is the need to upload and manage information stored in spreadsheets and other similar tools. Most PIM platforms will provide varying degrees of flexibility and customization with regards to integrating data from spreadsheets, including creating a product catalog directly from a spreadsheet to managing “hot folders” for updating products already within a catalog.
- Manage – The management of product data is very different from managing party data. The critical difference is that a PIM system will allow for product managers and other stakeholders within the enterprise to collaboratively author and update product content. (You can see why this would not be ideal for customer data.) Some of the typical activities during the stage would be:
Defining product attributes
Analyzing data completeness
Creating hierarchies and assortments
Classifying products according to hierarchies and assortments
Assigning pricing information
Performing localization or any other types of translation efforts
- Publish – Once rich content has been created, it needs to be shared with all commerce channels and syndicated to trading partners. Most PIM platforms operate on a “write once, publish everywhere” concept to ensure that all consumers, be they B2B or B2C customers, will receive the same rich product content.
MultiDomain MDM (MMDM)
As we mentioned earlier, many organizations will embark on an MDM project with an eye towards mastering multiple domains of data – and importantly, exposing the crucial insights about the relationships between your customers and products, products and suppliers, suppliers and locations, etc. An MMDM system will provide all of the capabilities defined in the previous two topics, but will do it within the same architecture and should allow the ability to view the relationships between data domains and use that information to make high-quality business decisions.
Digital Asset Management (DAM)
While not technically an MDM technology, many platforms will include the ability to manage digital assets, such as images, videos, and documentation, and associate that information directly with a product, customer, location or supplier record. This can be important to offer a superior shopping experience, manage supplier contracts and agreements, or recognize top customers from social or review data.
In a future post in this series, we will discuss deployment models, including Cloud vs. onPrem, but, in the meantime, contact us if you would like to discuss any of your current MDM needs.