In this episode of CDO Matters, Malcolm is a guest on the Data Engineering Podcast with Tobias Lacey. This episode is ideal for any non-technical data leader who wants to gain a deeper understanding of some of the technical dependencies and concepts required for successful Master Data Management (MDM) and data governance – but without getting too deep into technical jargon or software engineering concepts.

If you're a business-focused CDO with limited technical experience or background, this podcast will help you build your data literacy and have deeper and more compelling conversations with your technical staff – and it will help you make more informed technology-focused decisions. Malcolm and Tobias cover some of the technical concepts associated with MDM and governance programs, including:

• MDM system architecture and typical implementation patterns
• The connection between MDM, data technology and system architecture
• Data modeling for MDM and data management
• The processes used in MDM platforms to support data quality requirements
• Entity resolution, i.e. matching or deduplication
• MDM team dynamics and roles, the role of data management.

After listening to this podcast, any data leader who is new to the concepts of MDM or data management will understand why their organizations need these foundational elements and will have a better understanding of how they can leverage them to drive business value.

Key moments

3:14–6:50 Identifying “who a customer is” to model and manage data
7:11–9:27 What is MDM and what added value does it offer?
10:27–14:57 Who needs MDM and how does new technology solve the problem of data quality?
15:11-17:53 Limitations and considerations in the search for a “single source of truth”
18:15–21:45 Who is responsible for MDM in an organization and who trains it?
22:16–26:59 What are the differences between analytical and operational MDM?
29:15–31:50 The 4 main reasons why so many MDM implementations fail?
32:45–36:40 Determining the right results from a business perspective
37:40–42:25 How MDM is evolving and using graph capabilities in addition to relational databases 42:32–43:15 Why Customer Data Platforms (CDPs) are not enough for enterprise-level management 43:36–49:51 Insights into new use cases of MDM: data sharing, graph databases, data structures
50:08-53:53 3 “Precautions” we have learned from years of working in data management
54:36-57:26 How small businesses can implement MDM principles
57:38-1:00:14 The gap between data software and real business results

The central theses

When is MDM relevant for an organization? (10:22–11:34)
“The larger and more complex you are and the more decentralized you are… companies struggle to have a unified view of the customer… the larger the company, the more the need for MDM tends to be.” – Malcolm Hawker

Cloud-native data warehouses vs. MDM software (15:11-16:33)
"There are many cloud-based data warehouse technologies that claim we can enable a single version of the truth, and they can… but do they have the flexibility and reconfigurability to enable everything that MDM software can do? That's usually not the case." – Malcolm Hawker

What are the differences between analytical and operational MDM? (23:51–26:22)
“In an analytical style of MDM, the flow [of data] is one-way… [operational MDM] can actually turn around and push the data back to the consumer systems.” – Malcolm Hawker

4 MDM pitfalls to avoid in your implementation (29:15–31:01)
“If you need MDM and have been mandated by your management to create a single version of the truth, then you will avoid the key pitfalls that cause so many MDM programs to fail.” – Malcolm Hawker

Companies of all sizes can benefit from MDM principles (57:05-57:26)
“I would argue that most organizations need MDM as a discipline… But there are probably still some use cases that require this consistent approach to data management…” – Malcolm Hawker

About the guest

Tobias Macey is a dedicated engineer with many years of experience and even more areas. He currently leads the Technical Operations Team at MIT Open Learning, where he designs and builds cloud infrastructures to provide online access to education to the global MIT community. He also owns and operates Boundless Notions, LLC, where he provides design, review, and implementation consulting on data infrastructures and cloud automation. In addition to the Data Engineering Podcast, he hosts Podcast.init, where he explores the many uses of the Python language. By applying his experience building and scaling data infrastructures and processing workflows, he helps audiences explore and understand the challenges associated with data management.

Please take the opportunity to connect with your friends and family and share this video with them if you find it useful.