Stewart Bond is VP of IDC’s Data Intelligence and Integration Software service. With over 30 years of IT experience, he researches trends in data movement, ingestion, transformation, and cleansing in the digital business era. Stewart is a recognized industry analyst, valued by leading software vendors and peers in enterprise data management.
As the Co-founder and CEO of Alation, Satyen lives his passion of empowering a curious and rational world by fundamentally improving the way data consumers, creators, and stewards find, understand, and trust data. Industry insiders call him a visionary entrepreneur. Those who meet him call him warm and down-to-earth. His kids call him “Dad.”
0:00:03.6 Satyen Sangani: Welcome back to Data Radicals. On today's episode, I'm joined by Stewart Bond, Research Vice President at IDC. I have enormous respect for Stewart. Not only did he launch his career as a data practitioner, he also coined the term data intelligence. In this conversation, we dive deep into data intelligence. What is it? Why does it matter? And how will the rise of AI impact this large market? Stay tuned to hear Stewart's insights into the past, present, and future of data intelligence.
0:00:31.6 Producer: This podcast is brought to you by Alation, a platform that delivers trusted data. AI creators know you can't have trusted AI without trusted data. Today our customers use Alation to build game-changing AI solutions that streamline productivity and improve the customer experience. Learn more about Alation at A-L-A-T-I-O-N.com.
0:00:57.2 Satyen Sangani: Today we're excited to have Stewart Bond on the show. Stewart is the vice president of IDC's Data Intelligence and Integration software service. His research includes emerging trends that are shaping and changing data movement, ingestion, transformation, mastering, cleansing and consumption in the era of digital business. Stewart has over 30 years of IT industry experience in database and application development. He's a recognized and valued industry analyst by leading software vendors, of course, including Alation, consumers and peers in the enterprise data management today. Stewart, welcome to the show.
0:01:28.8 Stewart Bond: It's great to be here, Satyen. Thank you for having me.
0:01:31.1 Satyen Sangani: So, we've known each other for many years and you know, have followed each other through the space, but I think many of our listeners might not know you outside of your current context in IDC. So I'd love for you to start by just telling us kind of about your career, what you've done, how you've done it, because I think that that knowledge obviously informs so much of what you are and who you are as an analyst.
0:01:51.8 Stewart Bond: Yeah. No, appreciate that, Satyen. It's been a long career for sure. My career goes back actually to my post-secondary education. When I got out of high school, I was really good in math, really enjoyed math, and I went to a very prominent Canadian university to study mathematics and computer science. That didn't go so well. And I found myself at a community college studying to be a transportation engineering technologist. I did great in that program and every work term that I had as a part of that co-op program, I did something with computers. When I graduated, the first company that took interest in me was a transportation technology consulting firm, that was using computers to solve transportation problems. So, they snatched me up and I think for the first year and a half, maybe two years of my career there, I spent more of my time building a management information system using a database software program, than I actually spent doing anything related to transportation engineering technology.
0:02:56.1 Stewart Bond: I found my way out of that organization, into a local municipality, thinking if I was going to be successful in my career, I should get some experience in what I went to college for. While I was in the transportation planning department, and I really did not enjoy my job. I did not enjoy it so much that I actually figured out how to get back into computers. And I spent quite a bit of time in the industry, in the municipal government setting, working with computers and doing a lot of programming, doing a lot of integration work, application integration, really at the data level. Then I found my way into being a consultant. I was a consultant with a services firm in the late '90s. And then from 2000 until 2010, I was with IBM, I was a certified IT architect with the Open Group working at IBM, and I was doing everything from implementing integration solutions for customers all around the globe, to helping them set their integration strategies.
0:03:57.2 Stewart Bond: And so I was working with integration technologies, both data and application integration. And then I found my way into becoming an analyst in 2011. And it's been great. It's been my journey ever since. I've really enjoyed it and I've really enjoyed this space of the market that I'm looking at, which is data integration and data intelligence. Hopefully that gives you a little bit more clarity and background on who I am and where I came from.
0:04:22.2 Satyen Sangani: Yeah. It's, I think really important because despite the fact that you spent 10 years as an analyst, where you are looking across companies, across technologies, a lot of your writing and sort of your thinking is informed by some of the basic problems that you faced as a practitioner. And I think you in particular, like at least in my experience, that comes out pretty deeply in how you articulate the market, think about the market. And I think it's really great because it tends to be less, in many cases, theoretical than it does practical. And I think you speak about it in a very pragmatic way, which is what I've always appreciated.
So, you mentioned the word data intelligence, and it's certainly that's historically how we at Alation have as described ourselves and the market that we sit in and we really aligned to... Have aligned historically to that definition. What made you come up with that term and what is data intelligence, and what forced the need for it to be something that was new, relative to what existed at the time?
0:05:15.9 Stewart Bond: Yeah. Absolutely. I actually went through some of my notes, went back and looked at some of the things that I published not long after I got to IDC. And back in 2016, I discovered that I first started using the term data intelligence. In late 2016, getting into 2017 as GDPR was gonna be coming out in 2018, I had companies calling me and asking me where could they buy a data governance solution? And my response was, you can't buy a data governance solution. It is an organizational discipline. And that discipline requires people, process, policy... Oh yeah, it requires technology. But to me, the technology that you use to govern your data is data intelligence technology because at the heart of it, at the core of it, data intelligence software, data intelligence technology really gives you all of the intelligence about your data. And you need that intelligence about your data if you're going to have any hope in taking control of it and governing it. And the other benefit of the data intelligence terminology is that data governance always have kind of a negative connotation to it.
0:06:30.3 Stewart Bond: Governance was always taking control and no, you can't have access to this data. No, you can't use this data. So governance was not really looked upon in a favorable context by people in the organizations. And so intelligence was a little bit different term that allowed me to talk about it in a more positive light and help those organizations that were trying to do data governance bring it back to the basics of, well, we're really just trying to gather all the intelligence about the data so that we can then take control of it. And in some cases, some of those organizations actually spoke about, not data governance, but data enablement, which is about getting the right data to the right people at the right time, which is data governance, but it's just the reverse of that. You're using a positive term instead of a negative term. So that's the genesis of that. That's where that came from.
0:07:23.7 Satyen Sangani: Got it. So, data governance was the why and data intelligence is, in essence, the how. And what does data intelligence consist of? Like is it one type of software or when you started the market, was it multiple types of software? What are the different components of the data intelligence stack?
0:07:40.6 Stewart Bond: Yeah. So, we actually define that IDC. There's a formal definition. There's a blog post and there's documents at IDC that talk about this, but we essentially talk about it as, data intelligence leverages business, technical, relational and operational metadata to provide transparency of data profiles, classification, quality, location, lineage, and context. Which helps organizations and people build processes and leverage technology with trustworthy and reliable data to answer fundamental questions about data usage, governance, access, necessity, and relationships.
So the technology stack that goes into that, at the core, we actually see the data catalog as being almost the cornerstone of that stack of data intelligence software, because it provides all that necessary location information, the context information, it gives you the classification, it gives you potentially the profiles of the data.
0:08:40.9 Stewart Bond: So there's the data cataloging. Data quality fits into this bucket where it's not only the data profiling capabilities that give you an idea of what the shape of the data is? What does it look like? How many nulls do I have? What's the min and max values? What are my smallest strings, my largest strings, all the different ways that you can look at data, understand the cardinality, all that sort of thing. But it's also the technologies. Here's all the rules that make up good data and it'll spit back, well, here's all the data that didn't meet those rules. That's probably bad data. But in addition to that, we also put matching and cleansing software in there. So those domain specific data quality solutions that know how to work with address data. They know how to work with contact information. They know how to work with location information and really cleanse that information. Add to that, we also put in that buckets kinda the more general metadata management capabilities that have been around for quite some time.
0:09:38.4 Stewart Bond: And there's a number of different metadata technologies that exist out there. But this goes way back to just basic schema definitions and basic catalogs that exist in some of these technologies.
Then we add lineage in there. That's part of catalogs typically as well. And increasingly we're seeing this whole new thing of data marketplace or a term we're settling on more and more inside of IDC is the concept of a data product hub. Because data marketplace gets really confused with external data marketplaces, those companies that are buying, selling, trading data. That gets a little confusing when you start talking about marketplace. You're talking about internal or external. So from an internal perspective, we're starting to sort of settle on this data product hub terminology. At at a core, that's what we put in data intelligence. Now at at IDC, we like to count things and one of the things that we do is we count how big the market is.
0:10:34.5 Stewart Bond: And in terms of the market definition, in terms of how we count the market at IDC, I also include what I call master data intelligence software in the core data intelligence market. And that is software that tells you, where's my system of record? Where's my system of reference? Where are my systems of entry and what are all the reconciliation and match survivorship rules that exist for all of the people, places and things that I care about the most in my business? All of that master data. So that's also a part of data intelligence in terms of how we count how big that market is. It is not the full blown master data management software market because we believe that is a much larger market. And that brings in things like product information management. It brings in potentially some customer data platform capabilities. There's a lot of additional components to master data management that is outside of what I would consider core data intelligence.
0:11:37.4 Satyen Sangani: Got it. So the MDM market is not part of data intelligence in your mind. It's a separate market, but the master data intelligence market, which might be sort of a Venn diagram between this MDM market and the data intelligence market could be included or maybe sits in both domains.
0:11:51.4 Satyen Sangani: Yeah. So, it's not necessarily all the applications to keep that fundamental customer record, fundamental product record, because often those things sit in ERP or CRM systems.
0:12:01.4 Stewart Bond: Exactly.
0:12:02.2 Satyen Sangani: But it is, on the other hand, just the notion of like, oh, I've got customer data in 14 different places, and oh, by the way, these things map to each other in this particular way. So, it does solve and address some of the fundamental MDM problems. I mean, I think it's really important, especially in your work, because you guys go out of your way to quantify these markets in a very thoughtful way, and I think bring it back to some form of empirical truth, where I think a lot of analyst firms rely upon the blurred lines to sort of, in some sense, gloss over where people sit. So, I do appreciate that about the work that you do. And I think it's important because all data people have to deal with this sort of definitional vagueness in the categorization, and of course the world isn't obviously messy in the same way that we all describe it.
0:12:47.9 Satyen Sangani: By building this hierarchy, you're being at least honest to the scale of the firms and the relative growth rates, which does tell you something in retrospect about where things are going, what's doing well, what's doing poorly.
But of course, these markets are changing constantly. You've got new entrants obviously that come in. Now there's this world of AI. How do you see the world of data intelligence changing with the advent of AI? Do you see the category changing or evolving? Does it stay the same and just serve a different master? How do you see that change?
0:13:19.9 Stewart Bond: Absolutely, I see it changing. I see it's gonna evolve. Fundamentally, data intelligence has not changed, but we're seeing the types of intelligence that are available changing, and we're seeing how that intelligence is being used changing. There's more semantic capabilities that are emerging, graph is becoming more prominent in helping to understand the relationships within and across data, and that really helps provide the context, which is so critical for AI. Clean, relevant, timely data is used at time of inference, whether it's predictive, interpretive, or generative AI.
0:13:57.7 Stewart Bond: And the only way really to improve the accuracy and relevancy of what you get from those outcomes, is to make sure you have not only that good data to feed into the model, but you also have all the appropriate intelligence about that data to feed into the model. But the other aspect of this is how is AI changing, how the data intelligence is being used? Absolutely.
0:14:19.8 Stewart Bond: When you think about as we're moving towards this sort of agent, copilot, assistant, advisor types of architectures, and you're leveraging these natural language interfaces in analytics, for example, to ask a question of the data that you have in your organization, the only way that agent or that advisor or that assistant is gonna be able to help you is if they understand the context of the question you're asking and they understand where in the organization the data that maps in that context exists so that it can put that data together and make that data available for those analytics to happen. And that all comes back to the five W's of data. Where is my data? What does it mean? How is it being used? Who's using it? And what are the relationships that are in it? So for me, that all comes back to that data intelligence as a core enabler of those technologies.
0:15:25.0 Satyen Sangani: Yeah. And many of those technologies, and generative AI in particular, talk a lot about this idea of context. And so, in some sense, all of this context is getting stored in a central place, and in the world of structured data, it does feel like the data intelligence layer is almost even synonymous with the context layer. Do you agree with that, or would you see subtle differences in those definitions, or any differences in those two ideas?
0:15:50.8 Stewart Bond: Yeah. It's a great question. I would say that the context of structured data is not defined by the schema or defined by the table within the database in which it's being used, because the data can be used in so many different ways. And the context of the data really has to be looked at in the context... From the perspective of how that data is being used.
0:16:17.2 Stewart Bond: And data intelligence can help answer that question, because it can, for example, offer a business glossary, for example, that says that customer to marketing means this to the marketing department, but customer to sales means something different, and customer in terms of order fulfillment means something completely different. So, data intelligence business glossaries that exist within or alongside data catalogs can help you answer those questions. And then if you overlay a graph on that, you can start to understand those relationships, and you get more about that context, about that field called customer that exists in a table. The point is that that data is being used in so many different ways, in so many different places in the business, in the organization, in the applications.
0:17:10.8 Stewart Bond: You need to have that understanding so that when you are building a prompt using natural language or feeding it into any kind of model, then the models also have an understanding of context. Providing that information in terms of what that data is gonna be used for or what kind of outcome you're expecting from that question you're asking can help put it into the right context.
0:17:33.2 Satyen Sangani: I think the example of customer meaning multiple things in multiple different settings, I think is really to the point of, if you think about something like ChatGPT, very practically you can say, how many customers do I have? And there can be five different answers depending on which bit of context you're referring to. And that is why these generative models don't necessarily work super well on these precision database systems, because one is fundamentally about unstructured text and predicting the next token, and the other one is about really pulling the right semantics out of a database. And those are two different problems that this world of data intelligence helps you bridge. You do this great job of speaking about the market in very grounded, practical terms.
0:18:15.0 Satyen Sangani: That's why I like, frankly, your work.
One of the things that I think is somewhat interesting is that there's a lot of messiness, though, because lots of people say they have a data catalog. Lots of companies are coming in and saying that they do intelligence. And so you see the hyperscalers do it. You see some of the major platforms like certainly Snowflake and Databricks do it. I think even Databricks talks about itself as a data intelligence platform. How do you see this category evolving? Will it stay as a standalone category? Does it get enveloped in other categories? Where do you see this moving?
0:18:47.3 Stewart Bond: I think I get that question every week from someone. I think from the perspective of, is data intelligence going to remain its own thing, or is it gonna be subsumed by a bigger platform or by something else? At the end of the day, I think it's only going to be subsumed into a bigger platform or something else if that is the only platform that an organization is using in their environment. And the likelihood of that happening is slim, very slim. Some organizations will choose to be all in on one hyperscaler, or will be all in on one data platform. But I know the surveys that I do, the survey that I just did this past summer, I think we had 35 different sources of data on average, data was being pulled from to be put into an analytical repository. And it wasn't one analytical repository it was going into, there were 18 different analytical repositories that the data was going into.
0:19:48.4 Stewart Bond: So we talk about the modern data environment as being highly distributed. Data's all over the place. It's very diverse. There's so many different kinds of data that we're dealing with today. It's also very dynamic. That data is always moving, and it's always changing. So I think data intelligence as a category, as a capability, there's always gonna be that need to have the intelligence about the data that the organization manages in the modern data environment available that is visible across all the different places the data lives in that modern data environment.
0:20:25.0 Stewart Bond: Now, whether that is in one data platform that's able to federate all that intelligence together, and also operate on that data, which is where you get the likes of the cloud data warehouses and the cloud data lakes and cloud data lake houses, I'll just leave it at that. It's where you potentially, you could also have a data intelligence software vendor that is independent of any data processing, but they have that intelligence about all that data that exists in the organization, have visibility into where all that data is and what organizations are doing with that data. In fact, there's such a focus, I think there's a huge focus in that data modernization initiatives to get all the data into, at this point, I'm gonna say the modern data platform, which is different from the modern data environment that I just talked about.
0:21:21.5 Stewart Bond: There's so much focus in getting all that data into one place. I think the first move that organizations should take is to get all of the intelligence about their data into one place. And then depending on the problem they're trying to solve, they can then make that decision on where that data needs to be. But for a particular business use case where you need access to that real-time live data, you might be able to leverage data federation capabilities and use that data in situ and give you what you need. But if you don't know where that data is and what that data looks like and how that data is changing, which you get in your data intelligence software, you might just think, oh, I have to wait for it to get into my modern data stack before I can leverage it. So I think there's places for both, but I think data intelligence as a thing, I think will continue to need to be independent of where the data is.
0:22:16.3 Satyen Sangani: Yeah, it's funny. I get that question sometimes, not often, but sometimes from CIOs. And sometimes they're saying, well, why can't I just do this with the core data platform that I support? And I have an answer for that question. I'd be curious to know what your answer to that question would be. Like if a CIO were to come to you and say, well, why don't I just do this as a part of Megascale or X 'cause I'm putting all my data compute there, what would be your advice to them?
0:22:40.4 Stewart Bond: My advice to them would typically be that is probably good for a time, but there will likely come a time when you will need capabilities that go over and above what you have. In a lot of cases, what you'll find in the hyperscaler platforms is that their data intelligence capabilities are still somewhat limited to the data that's in the scope of their platform. That is changing. That is changing as we're seeing some of them talk about being able to federate intelligence about data and being able to federate data that exists outside of that platform. And I would question whether or not all of your data compute is really going to be in that location. Perhaps for analytics, if you're just looking at analytics, that might be the case. But you also need to look at what's happening in all the applications, what's happening operationally, what's happening from an event streaming perspective, what's happening with all the APIs that I work with. So there's a lot of other places that data exists in the organization that is not just in that data compute.
0:23:50.5 Satyen Sangani: Is there any effective lock-in to the compute layer do you think that folks can build in? If the compute and the intelligence stack are united, does that help you as a compute vendor or a processing vendor have more control over a customer?
0:24:07.7 Stewart Bond: Satyen, I think that's a great question. And absolutely. There is a desire to have all of those capabilities. And it's really funny. I think if we think back to when my career started, right around the time that PCs first started to emerge. And way back then, it was all about compute and data. Everything was on one platform until those PCs emerged. And now it's almost coming back to that one platform, but it's so much more diverse than it ever was before. And there's a lot of the survey work that we do what we know that it's a multi-cloud world. It's a hybrid world. There are very few organizations out there that have chosen and are staying on one single cloud hyperscale platform.
0:25:00.6 Stewart Bond: And there are very few that have all of their compute on that one platform. There's absolutely a benefit for vendors with the larger platforms to have the capabilities, the software to give you the data intelligence, to give you the analytics, in addition to giving you the compute, in one way, because for the customer, they're paying for that in the same way that they pay for everything else. So the customer that is all in on a particular cloud platform, yeah, they're gonna look for capabilities from that cloud provider, probably first. And if that cloud provider doesn't have it, or it doesn't meet their needs, they'll then look in the marketplace. But if they can pay for that technology, if they pay for that capability, the same way that they pay for their compute and their storage, it's gonna be easier for them to work with.
0:25:54.2 Satyen Sangani: Maybe switching gears a bit, if you think about the world of intelligence, there's been some interesting writings recently about the CDO, and there's been a couple of CDOs, I think they were based in the UK, who came out and said the CDO is effectively dying as a role of the chief data officer, and they gave a whole bunch of really interesting reasons for it. They talked about how the CDO is not necessarily a profit center, but a cost center, that it's hard to get success in the role, they're pulled in a lots of different directions, that they might favor technology over solving business problems. So it was an interesting discussion. How do you see that impacting the market? First of all, do you see the same phenomena? Do you think this is a thing? Or is it just like a casual observation? And how do you see this evolution of the budgetary role impact the evolution of this market?
0:26:43.0 Stewart Bond: Well, certainly the role of the chief data officer is growing from what we've seen. And the number of chief data officers that exist out there in the world is growing. And we actually have a completely opposite perspective. Every year IDC, publishes our FutureScapes. And our FutureScapes are essentially 10 predictions for different areas of the market. Going forward into the following year. So, these are our predictions in 2025 and beyond. But we actually believe that by 2028, 60% of chief data and analytics officers in the G2000 companies will rival the CIO in terms of influence on enterprise spending and technology. I'm not saying they're gonna have the same budget, but they're going to have the same influence.
0:27:36.7 Satyen Sangani: So on the marginal incremental spend, they might be impacting that equally or differentially greater than they otherwise are today?
0:27:44.4 Stewart Bond: Absolutely. And the big reason they believe that is because of what is happening with AI. There is no AI without data. And there are very few organizations that have their data AI ready. And what the CDO can do, what the office of the CDO can do with the processes, the policies, the technology that it can bring to bear can really help the organization get their data AI ready. We know there's all kinds of concerns about privacy, concerns about sensitive information getting out and being used with external foundational models. There's issues about the quality of the data that's being used.
0:28:30.4 Stewart Bond: Data quality has always been a problem. And yet it's not until, I'd say, the last 12 to 18 months that people have started finally paying attention to it. And it's because even when we had predictive AI and there's a data scientist that was looking at that data and they were spending all of their time or 80% of the time preparing that data, cleaning that data, getting that data ready to use with their models. Why? Because if the data wasn't clean, the models didn't work. And so you always had this human in the loop when the data was being used with AI. Fast forward to where we are now in gen AI and foundation models, large language models, data is being used at the time of inference with these models without a person being in the loop.
0:29:23.7 Stewart Bond: In fact, at IDC, we're saying that people are going to be on the loop. They are no longer going to be in the loop. That's a huge distinction. And so people will be evaluating the outcomes of these models more so than being involved in making sure and providing the input to generate those outcomes. So there's some really interesting things that are going on. But for me, that comes back to, you have to use timely, quality data and intelligence about that data at the time of inference to reduce those hallucinations, to improve the accuracy and relevancy of what's getting generated by that model. And this kind of brings it back full circle to the importance of data intelligence and leveraging data in any business outcome, in any business activity, as more AI is being used.
0:30:24.3 Satyen Sangani: Makes sense. You mentioned this concept of a data product hub. You're saying that that's a new sort of category that seems to be emerging. Talk a little bit about that trend. What is it? What does it mean? What are you seeing in the market that's new and different in this world?
0:30:38.5 Stewart Bond: Yeah, absolutely. So we actually believe that data products are a key to getting your data AI ready, getting that data to be useful with AI. But this whole concept of data products has been around a long time, certainly before the AI that we've seen in the last two years. This whole concept of treating data as a product is critical. And so at IDC, we actually define... We don't define what a data product is. We say that a data product is defined by three things. By access, it has to be accessible within all the appropriate controls that are required to protect that data.
0:31:13.4 Stewart Bond: It needs to have business value. It has to have real business value. What you do with that data product has to generate business value in some way, shape, or form. And it has to have accountability. There needs to be an owner of that data product. That's great. That's really theoretical. I understand that. The idea of a data product hub or a data market place is that this data products are pre-assembled and they're put together and they're made available in a marketplace or a product hub. It's not unlike your favorite e-commerce technology. I'm gonna go and shop for the data that I need. Tell me what are all the customer data products that are available for use in my organization. They can find all those different customer data products that exist out there. Perhaps they're working on... They're building this customer churn model. [laughter] That they're building their own data product, but they go and find all the different data products that are relevant to the customer.
0:32:11.9 Stewart Bond: They go and look at all the sales data that's relevant to customers. They go and look at perhaps all of the interactions data that are relevant to customers. Those are the different data products that they want to put together to build that churn model. So they find the products, they put them into their shopping cart, they go and they check out, and when they check out, oh, it looks like one of those data products they don't have access to. So that kicks off an entire workflow that goes to the owner of that data product. That data product can approve or deny the request to use that data product because along with that request is, here's how I want to use that data. So it all comes full circle back to all of the metadata about how that data product is being accessed, how that data product is being used and leveraged in the organization is being captured as another point of intelligence about that data.
0:33:05.2 Satyen Sangani: Makes sense. Do you feel as if the data product movement is something that will transform the intelligence landscape itself? Like on some level it almost seems like the job of CDO is to basically build and build great data products. I mean, on some level that could be the summary of the organization and what they're supposed to do. Do you think that that's where the, the market is going on some level? Or do you feel like that might be an oversimplification?
0:33:31.3 Stewart Bond: It could be an over simplification, that there will be data products. The chief data officer absolutely plays a significant role in those data products being created. But the chief data officer is not the owner of the data product. The data products are typically owned by the data owners, which should be people on the line of business. But the chief data officer is the one account we're responsible for providing all the technologies and personnel and people to help those product owners create those data products. Data products are going to be... Absolutely, we believe... We do believe that is fundamental shift in thinking and a fundamental difference in approach to getting that data AI ready.
0:34:13.4 Stewart Bond: But not every data product is being used with AI. Data products are still gonna be used with analytics. They're still gonna be used for business intelligence. It's gonna be used for many different... Many other things in the organization. But you're also still going to have those people in the organizations, the data engineers, the data scientists, the application developers, the machine learning people, the prompt engineers. They still need access to that data at a lower level that maybe is below the level of a specific data product. They need access to that raw data. They need access to that raw intelligence about the data so that they can then assemble the appropriate products for people to leverage and use in the business.
0:34:59.0 Satyen Sangani: In the context of data product, you mentioned that there's this idea of sort of three things that sort of bring it forward. One of them is a product owner, one of them is access, and if you can use it and access it. And then the third is of course this idea of business value. And it seems like every data firm now is talking about the concept of business value, and that's quite in vogue. How do you measure business value and how should one think about that in the context of these data products?
0:35:23.3 Stewart Bond: Yeah. Every data project, every thing that you do with data in an organization needs to have some correlation to business value. In fact, this maybe comes back to the whole demise of the CDO. If the CDO cannot prove what they do with data has business value, has a relationship to how the business is performing in the market, then yeah, the CDO is a cost center. CDO really needs to be able to prove that business value. And I'll be honest, I think this is still an area that continues to evolve. It's really hard to do, but it really needs to come back to measurement for me. If you're not measuring what you do with data, if you're not measuring how much better you're getting with data, you can't even begin to make that connection between what you're doing with data and what the business outcomes are.
0:36:23.7 Stewart Bond: If, for example, you can show that you've had a marked improvement in data quality because of the data quality work that you've been doing. And let's say that data quality work has been focused in on data that's related to customer experience, customer retention, and oh, look, your customer experience and customer retention metrics are much higher today than they were before you started your data quality cleanup initiatives. Then those two things go together. You may not be able to tie the actual data quality improvement to a percentage point in your customer experience improvement, but there's a definite correlation in there. In fact, in a lot of the survey work that I do at IDC, I will ask how good are you at an organization at all these different capabilities and data intelligence, data management, data integration?
0:37:24.0 Stewart Bond: And we'll measure them on... Put them on a scorecard in terms of how well they do on those different things. And we'll ask them how much better they've gotten in those different things over the last three years. And we'll also ask them, what are the top business metrics that you use to measure your company and how your company performs in the market? And we do that both from a financial perspective and also an operational perspective. And again, how have these metrics changed or how have these metrics improved over the last three years?
0:37:54.2 Stewart Bond: And if we can find a correlation between how well they do data management and data intelligence and how well they've improved on those things over the past three years with an improvement in business metrics, we can draw a correlation. And we can say maybe there's some causation here. If it's customer experience, maybe they had an entire customer experience campaign that improved their customers experience scores, and it had nothing to do with the data. But outside of that, we do have this correlation between, hey, the data has gotten better, and the customer experience has gotten better. So we can point to, yeah, we've got some business value here in what we're doing.
0:38:36.3 Satyen Sangani: Yeah. And then there's that old adage, if you can't measure it, you can't manage it. So.
0:38:39.5 Stewart Bond: Absolutely.
0:38:40.4 Satyen Sangani: You are about to publish a MarketScape. A MarketScape is sort of a market evaluation mechanism that IDC uses. Lots of firms have different flavors of these things. What is a MarketScape, and how does it sort of differ from what might happen at some of your peer firms, and what does it weight, and how do you think about it being useful and different relative to what else is out there in these market evaluation guides?
0:39:06.4 Stewart Bond: Yeah, absolutely. So IDC uses this MarketScape market evaluation method that we have, and it really assesses vendors across two primary axes. We come up with a scoring rubric, and we score vendors on how well they do on capabilities, and how well they do on strategy. And so strategy is essentially your go-to-market, your market presence, your growth initiatives, your partnering initiatives, all of those things that go into how you go to market and how you're positioning the company for the future in terms of those strategic initiatives.
0:39:43.9 Stewart Bond: Then on the capabilities axis, we're really looking at what does the technology do? So what does the technology do. In terms of the data intelligence MarketScape, we looked at what is the functionality of data cataloging? What is the functionality of data governance, data stewardship? What is the functionality of data quality? What is the functionality of data lineage? What is the functionality of the data product hubs or the data marketplaces? And we score each vendor on those two axes, and we score them against that rubric. It's a placement in XY. It's not an evaluation of how one vendor compares to another vendor. It's how that vendor compares to that scoring rubric.
0:40:26.3 Stewart Bond: And so we have leaders, which is the top tier. We have major players, which is the next one down from that. And then we have market players. And typically, most of the vendors that we evaluate end up in that major players and that leaders spot. We don't often have folks in that lower tier. Sometimes we do.
0:40:49.2 Satyen Sangani: Makes sense. Well, Stewart, this has been amazing, as I suspected it would. Because I feel like every time you speak about the space, rather than speaking in sort of abstractions, you tend to be very clear about the mapping and the language. And I think that's what makes you awesome at what you do. So thank you very much for taking the time to speak with us and do look forward to catching up again soon.
0:41:08.9 Stewart Bond: Absolutely. It was a pleasure to talk to you today, Satyen.
0:41:14.2 Satyen Sangani: That was such an insightful conversation. Data intelligence is key to modern enterprises, driving transparency, quality, and governance. As AI continues to evolve, so too must data intelligence, leading to the rise of data product hubs, making AI-ready data more accessible. And for CDOs, big changes are coming. According to IDC, by 2028, your role could rival the CIOs in shaping technology investment. So stay agile, embrace the change, and keep up with the pace of innovation. Thanks for listening, Data Radicals. Until next time.
0:41:45.1 Producer: This podcast is brought to you by Alation. Your boss may be AI-ready, but is your data? Learn how to prepare your data for a range of AI use cases. This white paper will show you how to build an AI success strategy and avoid common pitfalls. Visit alation.com/AI-ready. That's alation.com/AI-ready.
Season 2 Episode 11
Generative AI is so new — and there are so many ways to leverage it and misuse it — that it can feel like you’ll need a separate AI to figure it all out. Fortunately, Frank Farrall, who leads data and AI alliances at Deloitte, is here to tell you about the decisions, variables, and risks that companies need to consider before they invest in AI.
Season 1 Episode 23
Centralizing data was supposed to produce a single source of truth. Why did it fail? Zhamak Dehghani shares why she created the data mesh, and reveals how this socio-technical approach decentralizes data ownership to address growing complexity at large organizations.
Season 1 Episode 13
Growth in any industry usually requires innovation. But when you challenge the status quo, you encounter different levels of risk. Bigeye CEO and former Uber data scientist Kyle Kirwan details his experiences on finding the balance between innovation and risk.