In a recent CDO magazine interview, Sue Pittacora, Chief Strategy Officer at Wavicle Data Solutions, and Sarang Bapat, Director of Data Governance at Mitsubishi Electric Trane US (METUS), sat down for an in-depth discussion about how to demonstrate the value of initiatives related data governance and key data, analytics, and AI focus areas in 2024.
In continuation from the second part of the interview, this conversation with Sue and Sarang explores the importance of data quality and how high-quality data drives successful AI initiatives.
Throughout the discussion, you can learn why effective data governance requires a cultural shift within organizations and the importance of educating people about it in order to fuel analytics and AI projects.
Watch the full interview here or scroll down for a detailed transcript of Sarang’s insights.
Sue: Hello and welcome to the CDO Magazine interview series. I am Sue Pittacora with Wavicle Data Solutions, and I’m delighted to be joined here today by Sarang Bapat, the Data Governance Director at Mitsubishi Electric Trane US, otherwise known as METUS. Welcome, Sarang.
Sarang: Thank you, Sue.
Sarang: There is an initial excitement about data governance programs. You probably get basic funding resources, set up a couple of use cases, and you get rolling. Then it fizzles out like that mentality that it’s a project, when it is not. It needs to be culture driven.
I would borrow concepts from economics, here’s what I would borrow. I love Shark Tank. Right after the pitch, sharks ask how much it costs to create that product or make that product, which is critical for the business. So, if you think about data governance, and this goes to economies of scale, something you make 200 of, you sell at $5. If you make 200,000 of them, you sell them at 50 cents.
If I apply this to data governance and doing it at scale, my ability to search the data that I need to create digital product goes down significantly. It was taking me two or three weeks to find out where the right data is. I don’t know where the quality is; I don’t know who owns it. Now, I have a data governance platform, I go there and type what I need, and it gives five data sets as the results. It gives the quality, which is around 85% or 90%, owners of those data, and the feedback that is given and crowdsourced by other people who use that data set. It’s a five-star asset or four-star. To use that data set, you can click a button to shop for that data, like an Amazon interface, and then it goes to the owner for approval.
Now, we are making a lot of strides forward, where I found that data within a few hours versus a few weeks. The amount of time to search for that data was reduced. So the way I’m creating a new digital product and the cost of making that product is a lot less. Going back to my example from Shark Tank, if things are at economies of scale, if the governance platform is at scale, you are going to find all that information at your fingertips quickly. So that’s one metric.
How many systems do you have in terms of economies of scale on your platform? And we are data governance people, so we need metrics, right? We need to know how much? Year one, there were 10 systems. Year three, 25 systems. And in year five, 50 systems That shows how you are scaling.
Another aspect is economies of scope. For example, the iPod came in 2000, and then the iPhone came in 2007, but the iPod technology was reused in the iPhone. The marginal cost of creating technology for the iPhone was zero, pretty much. But the incremental revenue to add that feature was there. Now, all you’re doing is incorporating your current capability, adding that to your product, and increasing the price because you have a new feature since there’s a cost associated with it.
If I translate that to data governance, building a data lineage is a regulatory requirement. But you can use that same data lineage for your data change management a hundred times over. When something changes in the source, you don’t know where it’s touching because you don’t have the lineage to know the 10 other places that data is going. But if you have built that correctly, you can use it many times over and solve the data change management problem. So now, you have created two capabilities out of one capability. The lineage used for regulatory purposes is now used for data change management effectiveness. So, my metric would be how many of these capabilities we have built.
Another example would be sometimes you don’t need extreme data quality rules. A data profile would suffice. If 20 percent of the lightweight data sets, data profile is sufficient instead of data quality rules, you already took care of that just by implementing data profile. Simple capabilities like how many of them are created is another metric.
The last one is economies of learning. For example, take the components of the Tide washing soap and its interaction with water. To check this, the company has research development centers on all six continents because the quality of water is different everywhere. And the type of product they need to create varies, and they would learn from each other on how better to make that product. Likewise, in data governance, as you scale and as data stewards get better, we create communities of practice. That’s our economy of learning.
Now data stewards start talking to each other to identify how to make things better with data. They start asking questions on why the data governance platform is not doing that. For example, let’s imagine somebody has already written data quality rules in their system and does not want to recreate them in other systems. But can you actually expose the results of my data quality results into another data governance platform? That’s how the platform matures as a result of economies of learning
So economies of scale, scope, and learning; if I get metrics around that, I’m well on my way to show how I’ve made use of that investment and how this is helping in terms of productivity gains, data change management initiatives, and learning across the board. This is a framework I created and have shared in many other forums.
Sue: Yeah, you made some excellent analogies. Thank you. Great insights.
Sarang: A few years back, we were discussing how nice it would be to have a data governance framework because of defensive rationale and regulations. Now, as the gen AI debate is heating up, there is a lot of talk about good data foundation and data governance. It’s coming to the forefront right now.
Before, there was also a loose tie between data governance and business strategy. Now, people want to know how governance ties to the business outcomes, which is a big change. For example, if a mortgage business used to have 20 percent of exception reports, now that’s 20 percent of revenue lost because you can only keep that in your queue for 30 days. If a customer is applying for a mortgage loan, and if I don’t hear from the company in a week or so, I’m moving on to a new vendor. So without a good data quality matrix, the mortgage application couldn’t be processed, and the company would lose revenue opportunities. Business outcomes and data governance being tied to business strategy are the changes I see in 2024.
A couple of other thoughts are more about AI and giving everyone analytics capabilities. Before, analytics capabilities were in the hands of data scientists and people who were highly technical. Now, everybody in the organization can use analytics. From a business perspective, you need to think about data democratization. CDOs are concerned about their data literacy programs and data enablement programs because they need to educate data stewards and everybody across the company about the data governance portal, which includes all the data assets, metadata, ownership, quality of that data, and how the lineage works. So the ability to take a lot more advantage of AI depends on how good your data democratization is.
That is a big area that CDOs will focus on in 2024 and take big initiatives to help everyone understand the data. So data democratization would be definitely something CDOs would look for.
The last point I would make is that the role of the data steward is changing a lot. Before, we were talking about talent strategy and how “data steward” was kind of a casual role. And now if you look at job descriptions, there are a ton of data steward roles out there. This has been a big change in the last couple of years. The other day I met somebody, and her title was chief data steward.
Sue: Oh, my goodness. I’ve never heard that.
Sarang: This is an important activity, and getting that kind of importance is awesome because then you will make more momentum on the quality of the data, ownership, and educating people. Those are the four things I would say CDOs would focus on: tying governance to the business strategy, data democratization, seriousness around gen AI topics, and then data stewardship.
Sue: I think what we’re seeing is the importance of data, and the quality of data is only accelerating in the near future based on new ways to use data with AI. So I think you’re right about all of that.
Sarang thank you so much. It was such a pleasure to meet with you today. Great insights. And to the rest of everyone, please visit cdomagazine.tech for additional interviews. Thank you.
Sarang: Thank you so much, Sue.
This brings us to the end of the three-part interview with Sarang Bapat covering data quality and data governance strategy topics. Learn more about our data governance thought leadership here.
Ready to get started on your data governance journey? Get in touch with Wavicle’s experts.