In Stay Sharp Episode 136, What is Master Data Management and Who Owns It?, co-hosts Juliann Grant and Jonathan Scott tackle a new topic that’s critical to the world of digital transformations: master data management. They’re joined by Ann Garcia, Razorleaf Project Manager and self-proclaimed “data enthusiast” with 30+ years’ experience in data management and governance. Together, they discuss what master data is, how it’s managed, and why it’s important.
Ann Garcia, Razorleaf Project Manager
The Fourth Pillar of Digital Engineering
Data is a cornerstone of digital engineering—along with people, processes, and technology—but Jonathan acknowledges it’s easy to get focused on one of the others and forget that data is foundational to your business even when it’s not part of a PLM or ERP system. Master data, Ann explains, is your organization’s unique set of business-critical information. She differentiates it from transactional data, reference data (all the standardized lists or controlled values in your systems), metadata, and analytical data. “Master data is the data in your business that you really need to curate, that’s specific to your business, how you operate,” Jonathan says. “You need to be careful with it because you’ve got to use it again and again in all your business processes. But it’s not the transient stuff that is, oh, it’s an order, it just flows through, it’s here and it’s gone.”
Master Data Management (MDM)
Ann makes the point that there are many different approaches to MDM, and her own has evolved over her career. She notes, “I have found the best approach is the KISS method. Just keep it as simple as possible when we’re talking managing data.” She views MDM as encompassing the work activities of collecting, cleaning, confirming accuracy, and entering the data. Jonathan immediately recognizes that mean the data in question isn’t always digital. “You’re thinking about the whole life of data, you’re doing activities like collecting data,” he says. “That might be writing it down on a piece of paper on the shop floor as you’re recording some process. It’s cleaning the data. It’s confirming that it’s right, and then it’s capturing it in a digital system or some permanent system, and then governing it.”
What can get especially controversial in the data realm is the idea of data governance. Ann was taught, and believes, that data governance (lowercase) is embedded in the management activity, in the data collection and data entry. “At the time of entry,” she says, “you’re doing proper governance to go, ‘Okay, is this approved? Were there some validation checks or controls?’” Data Governance (uppercase) is a different beast. It’s more formal, multi-departmental, and done after the data has been ingested into a digital engineering system.
What’s unfortunate is that governance (lowercase) isn’t always practiced. Ann says, “Some companies don’t realize it could be a cost savings as part of their effort for managing data. You don’t have to always wait till the end of the line. You can embed in your business process of managing data some of that governance activity.” Further, when you’re migrating data, it’s critical to define and standardize data between systems, regardless of whether you’re moving to a new PLM or an existing system. She adds, “Sometimes people forget there could be quite a bit of data cleanup and remediation just in understanding what the old definitions were and the old rules were compared to the new rules and guidelines in the new PLM system.”
Software implementations are a good point at which to define, evaluate, and clean up your master data, Jonathan suggests. He says, “Take the time and figure it out instead of ending up with another little silo of PLM data that’s not linked to your master data.” Ann makes the point that effort is required even to simply understand your master data, but it needs to be done, because master data can help solve some of the challenges companies have with integration, digital threads, and even AI.
Too often, the “same” data in different systems means different things because of how each is set up, but you’ve got to ensure what you’re putting into a system is as accurate and as clean as possible to make life easier for the data scientists, enterprise systems, AI, and data lakes or warehouses. “You really want that good foundational data accurate and built-in governance and topics up front, so then you’re right,” Ann says. “It doesn’t matter, twenty years, five years, thirty years, you have a very good infrastructure and process in place to manage data.”
Learn More About MDM
The full podcast contains more discussion and details on a variety of topics, including why Ann says PLM and other enterprise systems can still be very siloed, who’s typically involved in different types of data governance at different size companies, what the different domains of master data are, and more.
Be sure to check out Stay Sharp Episode 136, What is Master Data Management and Who Owns It? and join us each week for a new podcast.
Follow the Razorleaf Podcast, Stay Sharp in Digital Engineering on:
