The MDM Journey: 6 Lessons Learned from Mizuho and AMN Healthcare

Every successful business strategy depends on high-quality data. But we’re seeing a trend: traditional master data management approaches are failing to deliver the data companies need to achieve the outcomes they expect.

If this sounds familiar, trust us, you’re not alone. Too many times, companies bought into the promise of traditional MDM solutions which were designed to provide a system of APIs and a repository of data. But here’s the thing: though these companies were well-intentioned, they were solving the wrong problem. 

Businesses don’t actually need sophisticated tools to deliver better data repositories and APIs. What they need is better data. And when they fail to address the underlying data quality issues, their MDM implementations will fall short on delivering the trustworthy golden records their business needs to drive better decisions and achieve better outcomes. 

That’s where AI-native MDM comes in. AI-native MDM is disrupting the status quo by offering a new approach that focuses on understanding your data, transforming it into a reliable asset, engaging users to build trust, and integrating your data into critical operational systems. And it does so by harnessing the power of AI throughout the entire journey.

While embarking on an MDM journey may sound daunting, we’re here to assure you that it’s not. In fact, during the 2025 Gartner Data & Analytics Summit, two of our customers, Mizuho Bank and AMN Healthcare, offered insight into their MDM journeys as well as six lessons they learned along the way. 

‍6 Lessons Learned on the MDM Journey

Mizuho and AMN Healthcare both faced challenges when it came to delivering clean, trustworthy data. They both embarked on an MDM journey, supported by Tamr’s AI-native MDM solution, and along the way, they gained insights and learned lessons.

1. Canvas your data estate

Mizuho’s Chief Data Officer, JC Lionti, advises companies to gain a true understanding of their sources before proceeding on the MDM journey. And he speaks from experience. After discovering that the IT team masked poor-quality data, the team faced a two-month delay in their MDM implementation caused by confusion over source data. What the team believed to be “raw” data was, in fact, modified, causing confusion over source data quality. Their advice is to first assess the data by canvasing the data estate. That way, the team can ensure they have a true understanding of all the data sources – and their quality – before proceeding. 

2. Connect data project to business strategy

AMN Healthcare’s Chief Information and Digital Officer, Mark Hagan, underscores the importance of connecting data projects to business strategy. In his words, “nobody receives kudos for an MDM project.” They get credit for adding business value. To deliver on AMN’s objective of providing a world-class clinician experience, AMN needed to provide instant/real-time experiences. But their rules-based MDM lacked scale and didn’t provide the real-time APIs needed to do so. AMN Healthcare upgraded their legacy, rules-based MDM solution to take advantage of Tamr’s real-time, AI-driven approach that enabled them to focus on data movement. Doing so allowed the organization to meet the business requirement of delivering real-time experiences for their clinicians. 

3. Data quality is a process, not a prerequisite

While it’s tempting to fix all of your data at the source, JC from Mizuho advises against this approach, saying “perfection is the enemy of progress.” Instead, his advice is to focus on the quality of the output data, as this is more effective – and faster – than validating data quality upfront. Over time, you can turn your attention to remediating source systems. This approach enables your organization to align quality checks with real-world problems, not theoretical standards. 

4. Third-party data is your secret weapon

As an organization that acquired 18 other organizations over the years, AMN Healthcare clearly understands the importance of verified matches. But getting divisions across the organization to agree on which data is right is a challenge. That’s why, to improve confidence in the data, AMN Healthcare relies on third-party data enrichment. Doing so leads to higher matching accuracy which, in turn, helps to improve user confidence in their master data. 

5. Listen to your data consumers

Another piece of advice from Mark at AMN Healthcare is to listen to your data consumers. When you engage with your business users early and often, you can ensure your solution meets their needs. Further, AMN Healthcare also recommends that you think of your MDM as a data product. When you create an MDM data product, you increase visibility into the data, which gives your users new insights and benefits. Further, using capabilities such as Tamr’s customer 360 page, AMN Healthcare made it easy to share the data with users and collect feedback, which improves confidence in the data product and drives greater adoption of it.

6. Lay the foundation for success

Mizuho’s CDO offers one final word of advice: prematurely moving to operationalization is risky. Instead, heed his advice and assess your data first (see lesson number 1!) and then fix your data quality. By going in this order, you prevent scaling issues and stop the propagation of bad data. 

In summary, it’s important to remember that MDM is a journey, not a one-time project. Attempting to “boil the ocean” by fixing all your source data or tackling every use case leads to inefficiencies. Instead, start small, show progress, and iterate constantly. And remember, using a modern, AI-native approach to MDM will provide the adaptability and flexibility you need.

Tamr’s AI-native MDM solution enables your organization to move forward in its MDM journey with confidence. Whether you're assessing the current state of your data, cleaning and improving it, or advancing toward the ultimate goal of operationalizing your data, Tamr provides the guidance, support, and AI-native tools you need to successfully move through each stage of the MDM journey. To learn more, download our ebook, The MDM Journey: From Trustworthy Data to Operationalization.