 Live from the campus of MIT in Cambridge, Massachusetts, it's theCUBE, covering the MIT Chief Data Officer and the Information Quality Symposium. Now here are your hosts, Stu Miniman and Paul Gillin. We're back, Paul Gillin here with George Gilbert at the MIT CDO IQ Symposium. This is theCUBE, our traveling video platform and we are winding down our two days of wall-to-wall coverage here at MIT with Steve Todd, excuse me, who is the VP EMC fellow, Vice President of Strategy and Innovation at EMC and joining him is Doug Laney of Gartner, who heads up the, believe it, heads up the big data practice at Gartner. Not heading, but part of our new Chief Data Officer research. But you're a key guy there. And Doug is a specialist in Infonomics, which I'd like you to explain, what is it? Yeah, so Infonomics is an idea we came up with, I don't know, almost 15 years ago, the idea that information actually behaves increasingly as an asset itself, as a business asset, and that it's increasingly incumbent upon organizations to manage it, to monetize it, to measure it with the same kind of discipline as other assets, even though it's not a balance sheet asset. Is this a sell job you have to do? I mean, do you have to convince organizations that this is true? Well, at some level, everybody gives lip service to the idea that information is an asset, right? You hardly ever come across an executive, right? See, who doesn't say information is an asset. But then when it comes down to it, do they actually treat it with the same discipline, the same principles, the same practices as their physical or financial assets or their human capital? Not at all. But their systems aren't designed to capture and track it and measure it. Correct, that's part of the problem, but I think it's more procedural and a bit attitudinal as well, so. One thing, we've been talking a lot about the role of the CDO here and a year ago and the discussion has changed quite a lot. Last year we were talking about the CDO versus the CIO and the clash of the titans. That didn't happen. What we're hearing now is that the CDO is maybe much more of a relationship manager than a technology manager. That the task of getting the information out of the organization, creating the relationships, twisting the arms to get data, is really job one for the CDO. Would you say you're seeing that? Yeah, absolutely. Part of it is bridging that gap between information, between IT and the business. They just don't communicate, they don't collaborate particularly well and more and more we'll see that as the CDO's kind of primary job. Part of it is a language. There's a language gap between the language that IT speaks and the language that business speaks and CDOs are increasingly concerned with creating a vernacular that can be used to bridge that gap. Steve, you work with a lot of CIOs. I'm sure, do they see this as a problem? They do. And one of the reasons is they see their business model shifting where companies that are traditionally just selling products are realizing that in order to grow their revenues, they have to make money off of data. And they don't have approaches, business processes, roles, most importantly, algorithms to attach value metadata to their content. So we see the CIOs increasingly looking to CDOs and saying, help me assess the value of the data. I'll try and put the IT infrastructures in place, but what are the business processes I need to plug into? Are there any formulas or rules of thumb that are emerging in terms of how to associate data as value with products or services? In other words, we hear from GE we can save x% on fuel consumption or energy consumption or increased uptime. Are there any of these rules of thumb that are emerging that we see widely? Well, it's interesting that you ask because in a year and a half of doing research with San Diego Supercomputer Center, Dr. Jim Short, who's kind of an expert in the data evaluation area, as we've talked to people, they're hungry. They're hungry for equations and they're hungry for some hard approaches. And that's when we intersected with the infonomics approach, which actually is equations that may be right or may be wrong, but they're useful. And perhaps Doug can explain what some of these equations are. Doug, you can give us some examples. How do you give us some examples about data? Over the years, we've worked with our clients and valuation experts and economists and some accountants who derive some formulas, some models that are useful in different contexts. Now, there are different reasons why you might want to measure information's value. You might want to just kind of prove or justify some kind of IT or information management related initiative. On one side, on the other side, you might want to justify a business initiative through the actual deployment of information and the economic value that it generates. So we've come up with several different models. Some of them are more what we consider to be foundational, non-financial in nature, aggregates of data quality characteristics, indices of business relevance, measures of impact on key performance indicators. And then on the financial side, we've really just borrowed from the way that valuation experts value any kind of asset using the cost approach, the market approach and the income approach, but we've tweaked the models a little bit for some of the nuances of data. I was gonna say, can you drill down a little bit into that cost, financial, different approaches? Sure, so valuation experts will tell you that for any kind of intangible asset and arguably information meets not only the criteria of an asset, it's exchangeable for cash, it generates probable future economic benefit and it's owned ostensibly by an organization. That's really the three criteria for an asset. But it also meets the criteria of an intangible asset and valuation experts say that for any intangible asset, you should initially value it at its cost, whatever it costs you to generate or produce or acquire that asset. So that's a fairly easy calculation for information. But then you start to think about how are you gonna use that data and then you get into either the market approach, if you're gonna productize it or sell it, or the income approach, if you're going to attribute top or bottom line, financial gains to the information itself. When you get into the question of how to monetize data, I can monetize it by selling it, but I can also monetize it by giving it away and thereby improving customer loyalty, upsell, cross sell potential. What are some of the, how do you figure that into making that decision? Or you can even use it internally and monetize it that way as well. You can again attribute some top or bottom line, financial gains, some process improvements, some relationship improvements. Sometimes it's difficult to measure, but some people are a bit doctrinaire about what monetization means, and they say, well data monetization only means selling data. I think that's kind of a limited perspective in that we advise our clients to think a little more broadly about it, because you suggested that monetization means everything from licensing data, remember we don't sell data because that would imply the transfer of ownership, you're actually licensing it. All the way down to some indirect methods, including maybe bartering or exchanging the information, or again using it to improve business processes in some economically measurable way. So we argue that you're monetizing information if you can measure the economic benefits of it in some way. Are there accounting implications to this? I mean, do you know of any companies that carry data on their balance sheet? No, in fact, in the U.S. you're not allowed to, according to international financial standards, you're not allowed to anymore. Hold on, now isn't a company like a credit reporting agency? I mean, that's all they have is data. You would think that it was on the balance sheet and it's not. So they can carry a chair as an asset, but not a database. Correct, that's a great example. Here we are in the midst of the information age, and yet the thing that gives it that moniker, and that is carried by so many, not only information product companies, but is the source of value for just regular old companies, is something that you're not allowed to put on the balance sheet, according to financial standards. Now there's a working group at FASB now that's talking about it, but we're still probably years away from anything from FASB. But what it creates is both a problem in measuring the volatility of companies, but also in companies not being as transparent as they probably could be or should be with the marketplace. So I think we heard on an earlier segment that IP in the form of patents can be represented as an asset. They can, I guess you can amortize them if you acquire them. Right, copyrights, trademarks. Okay, but then other forms of information that don't or data that don't fall into those categories can't. Correct, and yet 80% of the executives that we survey believe that the value of their information is represented on their balance sheets. So there's an extreme level of executive ignorance about the realities of information as an asset. Again, they're all giving it lip service, but it's not something that is really measured in any way. And so we're trying to help give our clients some tools to do that. The, our keynote speaker this morning, whose name is suddenly, Tom Davenport, has had a great quote. He said that when clients come to him and say, help me, what do I do? My executives don't get the value of data and analytics. What should I do? He says, I tell them to resign. The, in fact, do you see many companies that don't appreciate the value of data and analytics anymore? Well, probably that would argue that there are very few organizations that completely value data and analytics. I don't know what you think about it, but we didn't think. Yeah, well, yeah, the ones that I'm talking about to focus less on the models that Doug just mentioned about economic and assigning a dollar value. And they're more interested in signing use cases and also these metrics that Doug talked about, such as, are you keeping track of data quality for a given data set? Are you keeping track of how much that data set is relevant to different lines of businesses? So as we've been exposed to the Gartner research, we at EMC have begun to create a metadata repository alongside of our data lake and expose some of those key metrics. And what that'll allow us to do is to be able to score different data sets in our data lake and start to understand, well, why is this data set more important than another or a higher priority and should we treat it differently and maybe we could monetize it? So this is along the lines of, I've got a metadata catalog that has like a business glossary, so terminology that's sort of English language and it's got sort of information as to what data gets used most and I can use that as a proxy for the data value. It doesn't show up on my financial statements, but there's implicit ranking and scoring and valuation going on. And we're learning how to manage it and measure it, which is a key part of understanding the value. We just can't report it. Right, but that's a good proxy. If you're trying to prioritize, as most organizations are trying to prioritize data governance initiatives, which data to focus on governing, that kind of measure is a great proxy. It's not a financial measure, but it's a great indicator of where you should be prioritizing. What if you were in the business of selling, just to pick an example, high frequency trading algorithms, not that you would sell the same to everyone, but they would take kernels and either assemble them or they would add their own secret sauce. Of the algorithms themselves? Yeah, you sell the kernel of, you go back 50, 100 years and you have a couple different perspectives, but of course, they can't run the same models as everyone else, but they can build on yours. Or they can feed different information to them. So how would that get accounted for? Well, algorithms are a different animal altogether. Algorithms are effectively a business process and business processes can be patented. So you see all the time, algorithms being patented. So that's different than the data that actually feeds them. So an algorithm devoid of the data is something that is a formal type of IP. So it's just anachronism, essentially that we don't have a similar accounting rules for data. Now, what's gonna be interesting is as machine learning takes hold, the algorithms that are generated by machine learning can't be described, right? Because they have their own networks and so forth. And so they're gonna be more difficult to patent. Are there, in particular industries that you work with, Doug, that you think are further ahead of the curve in understanding the value of data as an asset? And conversely, are there industries that just don't get it at all? Probably those that monetize it, I would say. The retail industry, telcos are pretty sharp about the value of the data that they collect. There are several retailers that directly sell their data to their CPG partners. And in that way, they're actually able to capture the market value of that data. Telcos do much the same thing, but they do it very quietly because we're all very sensitive to having our- Regulation. Having said that, one interesting thing about that is even the companies that are selling data sets, when you go to a negotiated price, there's no formulas or there's no standards. And so we're finding that the price is negotiated based on what you paid for the last data set that you bought from the vendor and what was in it and how big it was. And then you just barter from there and you dicker from there. So there's a real need for formal methods and standardization to negotiate for data sets. And I think you'll see them emerging. No, the topic we can get into, but we are out of time. Okay, okay. Like, thank you, Doug Laney from Gartner. Steve Todd from EMC, fascinating discussion that we can go on and on. But this is the end of this segment. This is theCUBE. We will be back with our next guest in just a minute. Stay tuned. Thank you.