Master data is increasingly becoming the most important data companies have. It’s the key to managing, organizing and transporting data to make it relevant across the business. NetApp recently shared five tips for how they implemented a master data program within their organization. Speaking from practical experiences, they revealed that companies making a foray into master data need to be nimble, on the same page, thinking about governance, aligning with the C-suite, and continually articulating the value. Now, we’re turning to the experts helping companies like NetApp make their master data goals a reality. We sat down with Dun & Bradstreet’s Distinguished Architect of Master Data, Elizabeth Barrette, to learn more about how companies should approach executing a master data strategy.
Master data is not easy; it’s a time-consuming, labor-intensive process. In fact, it’s often been referred to as a grand journey. That said, can you share some of the things you’ve seen that have failed? What are some of the roads companies shouldn’t be taking on this journey?
Barrette: It always surprises me how often companies are blinded by their own data quality. They look across all their disparate systems and choose one where they think the data is the best and they choose the data from that system as their primary source. From there they take the data from all their other systems and match it to their primary source. It could be bad data matched to bad data and this could end up creating bad decisions. A gold standard is very difficult to create internally. The most successful master data programs look to third parties to help in the mastering of their data. Using a third party helps ensure data quality.
Another pitfall is organizations that allow the continuation of data silos. Making sure your teams work together is really one of the only ways that you can see success happening and ultimately gain value within your master data program. If somebody's building it, using it, improving it, if someone is getting additional value from it, it's a real relationship. Siloes can’t be tolerated.
What are some examples where companies are doing the right thing?
Barrette: As a solution architect on the data side, I've seen a reoccurring change with technology businesses who were some of the first to implement master data programs 10 to 15 years ago. They are sunsetting what they previously built and reengineering their master data strategies. They are thinking more holistically and futuristically on the changing times, especially on how to match the influx of data, and the speed of data. They are engineering processes that are mastering data on the fly. The best practice is to master data as soon as new data is created. It’s an interesting transition to see how these changes are driving greater value and insights faster than ever before. Essentially, it’s okay to start from scratch.
Is it unfair to expect to start seeing ROI from a master data strategy right out of the gate?
Barrette: Actually, you want to show value as early as possible! It comes back to the idea of being on a journey, not working on a one-off project. Businesses can get frustrated if they are not seeing continuous value. You hear a lot more about agile environments. You don't have runway to show the value of your program, you don't have years to put together your master data program. You want to make sure that your master data program is at the forefront of your business strategy, that you're driving value throughout the entire time.
Let’s take a minute to talk about data governance. Many companies don’t equate that to master data. Is it possible to have a master data strategy without data governance?
Barrette: No. If we consider master data as a lifestyle and not a project, you must look at governance throughout the course of your master data initiative; otherwise the data that you took a snapshot of in the very beginning is going to decay overtime. It will lose value continuously throughout that process.
What challenges do you see arise when companies don’t employ data governance in convert with master data?
Barrette: Data governance is about adhering to definitions and standards. If you can approach data governance proactively, for example a search before creating for a new entity, you are more likely to get it right from the start. Employing a pre-emptive approach is the best approach to ensuring data quality from the start.
It's very difficult to get it all right the first time and adhering to a standard can be very tough. We've seen a lot of successful companies put off the data governance portion of the master data process. They start on their data governance journey as a part 2 by understanding what's going right and what's not going right. No matter where you are in your master data journey, if you haven’t employed data governance you need to do so now or you won’t be successful.
As a suggestion to those already in the process; take a step back and look across your business. Look at the different business entities that you have and make sure you understand what the use cases are. What's the business value you're trying to bring to that portion of the business? Look at the data, understand what their data definitions are, and look at how that bridges across the overall entity. When you come up with that standard definition take into consideration data management. What's it going to take to get all the businesses to adhere to that same standard? Remember you've got legacy systems that you may not be able to change or upgrade to meet this standard or even incorporate it. Some of those newer technologies lend to change more readily than your legacy systems.
So, would you put data governance and master data management under the same leadership, or do you break that out into different groups?
Barrette: Data governance is an enterprise wide initiative that should include the most senior executives all the way down to the data stewards, data creators and data users. Anyone who touches the data should be versed and included in the governance process. Master Data Management is a business/IT program that is run and managed typically by a team of the business and IT.
Where should you incorporate data governance?
Barrette: Where shouldn’t you incorporate data governance! Unfortunately, all too often we see companies choosing either data governance to clean up their data reactively or integrating APIs for entity creation to start cleaning the data that it's coming in proactively. Even if data governance is used to clean up your existing data, you're going to lose the value without ensuring governance stops the bleeding on the new data coming in. Make sure you are incorporating governance proactively and reactively to ensure you maximize the value of your master data program and solidify its continued success.
You mentioned the importance of working with third parties to help companies master their data. Companies may be wary of external help when it comes to their data. How does this play into master data environments?
Barrette: Okay, you asked about my favorite subject. Before working at Dun & Bradstreet I was a solution architect for about 8 years making my living answering that question for other businesses. “Do we really need D&B as part of our master data solution? Do we need to take on that expense?” The answer is yes. Think of what D&B does today. We have created the largest business database in the world. It consists of over 265 million businesses. We make 5 million changes a day to the data; we spend a quarter of a billion dollars on data management, on data governance, and updates. We suggest you incorporate D&B data as your referential master data database as part of your ecosystem and let us continuously master the business data and keep up with all those changes.
Also, we help enrich that information, giving you an additional trusted source to expand your knowledge of business entities. Not only will D&B help in the mastering of your data but you will also have access to our firmographic data so that you can understand family trees, and employees, and things like that right out of the gate.
What’s the one thing you tell your clients before going on this master data journey with them?
Barrette: Managing change is often harder than mastering the data itself. Understanding that your data is a critical component to your businesses success makes any changes needed a requirement, not a nice to have.