Saving Money with IT Analytics: A Q & A with Dresner Advisory
Productized analytics have been a focus of Early Adopter Research since it was launched. In a recent conversation for the Early Adopter Podcast, Dan Woods spoke with Howard Dresner and Chris von Simpson from Dresner Advisory Services. They recently completed a report about IT analytics and their discussion covered that report. Dresner’s research sheds light on the role of productized analytics in IT in general. This is an edited version of their conversation.
Woods: Can you explain your roles at Dresner, as well as the objectives of the entire company?
Dresner: I’m the founder and chief research officer at Dresner Advisory Services and we publish a whole host of research during the course of the year focused on things like business intelligence analytics and information management.
Von Simpson: I joined Howard’s team in order to write about the analytical applications that are built using all of this various software, most particularly those applications that generate value, and particularly the horizontal applications, of which IT analytics was the first report.
I’ve analyzed different ways that analytics have been productized and I’ve noticed that there seem to be four levels. The first level I call the custom kitchen, where you can do anything you want. You get a tool kit. The second level is what I call dinner in a box, and that’s where you get like a Blue Apron where you can put together a meal that’s got some instructions, but really you’re assembling it yourself. The next level up is called artisanal brew and it’s like being at Starbucks, where you have lots of degrees of freedom and you can order and create many different types of analytics products by selecting the same way you do at Starbucks — I want a decaf skinny latte. And the final level is what I call the value meal, which is like going to McDonald’s and getting a very specific thing off the shelf. I’ve seen all of these levels represented in different kinds of analytics products. Given that framework, now let’s get to the survey research. What are the ways that increased awareness from analytics leads to better outcomes in IT?
Dresner: One of the things that I’ve noticed going back many, many years is that IT has been like the shoemaker’s children. They’ve gone barefoot for years and they haven’t had lots of great tools at their disposal. And one of the things that Chris and I have been talking about is that when you think of the CIO, you really should be thinking of them as the CEO of IT. They’ve got a very diverse and complex business that they’re running. There are financial aspects of it, there are project aspects, there are aspirational aspects, there are HR aspects, everything that you would expect to see in a complete business. But historically, they have not done a very good job of analyzing all the aspects of the business and sharing that or reporting that or conveying that back up the chain of command to senior management. And I think that increasingly the IT function, and probably all functions certainly but IT as one of them, have had more and more pressure to demonstrate value within the organization. And by leveraging business intelligence and analytics, it’s a great way for them to show the sorts of value that they really add to the organization and the many facets that they do on a recurring basis.
So by doing this you can change the conversation to be about not just the cost but the value as well?
Dresner: Absolutely. IT is a great enabler of the business, and yes, obviously there are costs associated with that, but it goes far beyond the costs. They do enable all the other functions, including all the front office functions and the revenue-generation functions to do what they do. So it’s not strictly a cost factor.
But unless you actually document it and quantify it, the conversation may be just a cost conversation?
Dresner: That’s the trick of it. You do have to measure it carefully, and these things are all related to each other, all the different functions and aspects of the business. And if you can get your arms around that, and then if you can relate those to other businesses you support, you can demonstrate far greater value.
One of the things that the survey highlighted was the importance of operational metrics. Why are these so crucial in the IT realm?
Von Simpson: The main reason is that we know what everybody’s objectives are for IT and for use of analytics. There’s some research we published earlier in the year, our BI market study, and we gather from thousands of people what their objectives are when using analytics. And everybody wants to make better decisions. That’s sort of always going to be the leader. But the second one is always to do with operational efficiency. Now, part of that, of course, is reducing costs, and operations metrics govern a significant part—not the whole part, but a significant part of cost management. And for IT worldwide, that’s a cost of around $3 trillion dollars this year. So there’s plenty of cost there to be saved. But of course, operational efficiency also is to do with how you run your business. As Howard said, a CIO can be thought of as CEO. So yes, it’s how you operate, how you run your business.
Does the focus on operational metrics as opposed to metrics that quantify value reflect in some ways the bias away from value and toward cost?
Von Simpson: Only to some extent. The reason is that in that same survey with the BI objectives, growth in revenue was the very next objective that people cited. And in fact, we know that people are quite carefully tracking the value that they’re generating as well. So we’ve got pretty good data on how much people are saving and how much people are able to track the ROI of their operations using business intelligence. So it’s a little bit of both, but it just so happens that operational efficiency comes ever so slightly above growth in revenues.
I was surprised when I read the survey and saw that 38% of the respondents said that they use a third party application designed for the purpose, which was IT analytics. What do you think they mean by this answer, and what is the range of solutions?
Von Simpson: I think what they mean by this are the IT analytics and IT performance software applications. Of course, these days many whole companies are dedicated to improving IT performance, and I was once asked is it better to try and build one of these or buy one of them. Well, the analogy I always like to give is if there’s a great car perfectly designed for exactly your needs, why would you try and build one and maintain it? And looking at software for a specific purpose is usually likely to work better than trying to make and start a whole new custom application yourself. Because of course, you don’t just build it, you also have to maintain it. So I would only build if there was absolutely no alternative. Now, thankfully, of course, that’s much rarer these days, that there are no alternatives. Thankfully these days there are a lot more of these third-party applications designed specifically for IT analytics, for managing the IT business.
Dresner: You also have to think about the platform. So given your model, Dan, the various layers and the different kinds of cuisine if you will, having the ability to customize is also important. There are so many solutions out there that are built on a platform so if they are an 85% fit, the customer or their service provider can then customize the additional 15% to make sure that it is a good fit.
It seems to me that sometimes in the IT realm, there’s a macho culture where it’s unmanly if you have to buy a product and you can’t build one yourself, because the companies that are very much admired are always doing engineering on their own and IT likes to see themselves as master engineers. Do you think that there’s a bias against buying in IT for its own purposes for these cultural reasons that I just mentioned?
Von Simpson: There can be. There’s even something called the not invented here syndrome. And I think everybody has that in them. Today, an IT department is quite a complex, large organizational entity. So although there’s always this tendency to think we can build it ourselves, we can invent it ourselves, I don’t need anybody else’s help, I think that there’s also a question about what is our value add here? Why are we, for example, trying to invent a whole new word processor? So don’t get caught up or distracted by needing to build everything yourself.
What advice would you give to people who were trying to justify the money spent on a product in terms of the ROI that potentially can come from using it?
Von Simpson: That’s actually, thankfully, very straightforward. In the report, we found that, first of all, you’re likely to be successful with IT analytics. Whatever method you use, whichever software you use, IT analytics initiatives have a success rate of 83%, which is just fantastic. Secondly, given the size of the potential cost savings, even though cost savings are harder to justify with the CFO than growth in revenue, the scale of the cost savings means that the IT analytics software or initiative is highly likely to pay for itself. Indeed, we have data that suggests that all kinds of BI initiatives generate north of 10% ROI, and we’ll have forthcoming research with a bit more detail on that. So you should plan for this and put it into your project plan if you’re an IT department.
Why do you think that IT analytics has such a high success rate?
Von Simpson: There’s two factors there. One is that there’s a lot of people still who don’t do it at all; they have no formal method. So when you start from, “We don’t do it” or “We have not formal method of doing it,” success isn’t particularly hard to achieve. I think that it’s because for most of these use cases for departmental or horizontal business function analytics are generally going to be successful. BI isn’t a brand new market. People are always demanding to run their businesses by the numbers. I think there’s a natural inclination now for people to want to embrace data and analytics and make the most of it, and I think those things combined lead to a high rate of success.
One of the great things about your research is that it was very detailed, and we talked about the operational metrics point earlier, but you also looked at the difference between operations planning, resource planning, resource management and systems development. And operations were where the metrics were considered most important and systems development was the place where the metrics were considered least important. Why do you think there’s such a difference?
Von Simpson: There’s a really wide disparity between them, but the differences come from all different places. You have to remember that companies that have IT departments are just so, so hugely varying. I’ll give you a couple of examples of the difference in operational metrics by industry. We’ve got another chart that shows how the different industries rate the importance of the different operational metrics. And telecommunications companies score outage detection, which is one of the groups of metrics of IT analytics, as 20% more critical than any other industry. Now, that’s not too surprising, because if your cellphone has no signal and it had a signal yesterday, you’re grumpy about it. But the variance in these use cases from company to company to company is enormous.
Do you think that the systems development is perhaps scored as the least important because the engineering processes themselves are not as instrumented and not as managed with metrics?
Von Simpson: That can happen of course. We have to remember that some people responding to the survey didn’t really have any kind of IT analytics to begin with. So yes, in cases that will be so. But there’s a lot of use these days of specialized tools, specialized software, just for software development, for development of systems and software and checking in and checking out of code and running analytics on the efficiency of systems development. So that does happen, but it’s maybe a smaller number of people might do it in some companies. And again, it’s just not going to be as critical as overall operations. You remember we mentioned earlier that the CIO can be thought of as the CEO of IT. Well, the people that take care of operations for the IT department can be thought of as the COO of operations, of IT.
I thought the 38% adoption of productized systems was a very high number. What do you think the forces are that are driving the increased adoption of these productized analytic systems in IT?
Von Simpson: The main force is value. As you alluded to, the thing that will drive the value is going to be making some money and actually using the software to generate value, whether it’s cost saving or revenue generation. So the cost saving piece of course, is huge. This process is purely about saving money. It’s not to do with destroying jobs or getting rid of people. In fact, this process of IT departments getting more efficient likely creates net new jobs. First of all, IT departments have no shortage of things to do. And secondly, the IT analytic software companies themselves are growing and also hiring. There are all sorts of new analytic software companies popping up. Just like IT analytics there’s going to be marketing analytics, sales analytics, etcetera, etcetera, and they’re also hiring too. In fact, if you look at the careers page of some of these sorts of companies, it shows dozens of open positions, most of which are IT positions. So the driving force is can you generate value in the use of the analytical software, either by reducing costs and/or by generating revenue. In most of these companies, IT is not just keeping the lights on but IT is actually generating the product, generating the business, keeping the business running itself. So no IT equals no business.
Dresner: You can think of it as like the early days in measuring sales, where sales used to be measured strictly on revenue and no salespeople are measured on revenue anymore. There are many measures. They’re certainly measured on things like profitability, measured on retention, customer satisfaction. And so IT has also had to evolve in how they measure themselves. In addition to that, if they want to get more budget and have more opportunities, they have to be able to demonstrate that real value to the business. And what better way to do that than by using something like IT analytics so they can share with management in a comprehensive way their contribution to the organization, both top line and bottom line?
Do you think that there’s any benefit that the companies that are doing these IT analytics products have had from observing the way BI has been productized in other realms and have perhaps been able to create a better product because they’ve learned from that?
Von Simpson: I think that’s happening all the time. If I was a product manager in an IT analytics software company, I’d always be looking at how is sales analytics doing, how are these analytical companies with products that are directly targeting a horizontal business function, how are they putting together their product. I think that all of these forces are gradually improving—ever so slowly sometimes, but are gradually improving the ways in which each horizontal business function can operate better. And I think that’s particularly the case with prioritizing those business functions with a critical operational benefit to the business, like IT.
Do you think IT analytics vendors have benefitted from observing productized analytics in other realms?
Von Simpson: Undoubtedly. I think it’s the case that.
Can you give some examples of the impact of the use of analytics in IT?
Von Simpson: Of course, there are many, many thousands of them as it’s gradually matured. The first obvious one that springs up is Netflix. Netflix of course, the product is going to show up on your device and you’re going to consume it, but it’s a service business, an IT service business where the licensed media products are going to show up across the internet on IT devices. If you think about it, it’s the Netflix IT devices coming through to the customers’ IT devices. Of course, customers don’t have IT departments, for the most part, but it’s an IT business. Running that business for such a company of course is a critical part, not just the product and the distribution, but it’s also a huge part of the customer experience. If the software isn’t working as well as it should and the video is stuttering or there’s this big buffer that you have to deal with, then that really doesn’t give customers a good experience. So running IT like a business there and running a great deal of IT analytics makes perfect sense. But you also have to think about all the other types of companies out there that are using a lot of IT to run their business too. Manufacturers didn’t have the highest score for IT analytics compared to some other industries, like for example financial services and insurance. But consider manufacturers with supply chains and the rise of the internet of things. Dresner recently published a report showing how IOT keeps finding ways of generating new value. And that value is generating additional investment, and the innovation, regardless of the industry or the use case in which it’s implemented, relies on many overlapping systems. And all of that is all IT as well, and all of that IT will benefit from real management, performance management using analytics.
It seems to me implied in what you’re saying is that you go, if you’re thinking of the CIO as the CEO, to a much more high granularity model and a high-resolution model, and it’s the awareness in that that provides the value. Are there any commonalities about what’s discovered in those early phases when they just get going?
Von Simpson: This is captured in some ways by the different use cases for IT analytics, which again relates back to those objectives. Some companies know that they are probably overpaying for some aspects of their services, whether it’s, for example, a cloud service or some other service, and they want to make sure that they’ve got the right kind of numbers and they need to reduce costs as quickly as possible. So that is always going to be a high, high priority. But again, it’s not just that. The quality of the service that IT is providing for the rest of the company is, in some other areas, going to be absolutely critical. So that would come back to that Netflix example, where you want to be able to say what is the quality of the service delivery to all of our customers across all of these different geographies. And the vast variety of companies means that even though the specific use cases are going to be rather more discreet, the overall grouping of those, the aggregation of those is going to come back to can we save money, can we generate revenue, can we run our business better by making better decisions and can we do all of this with some real confidence in the numbers?
My final question to both of you is what advice would you give for a CIO who has heard your advice about thinking of themselves as a CEO of IT about how to be successful with these productized analytics for IT?
Von Simpson: The first piece I come back to is you should do IT analytics. You’re likely to be successful. There’s consensus across the industry on the importance of doing it and you may be able to out-compete your market slightly using more advanced IT analytics than your competitors. But you also want to think about watching out for other people doing this to you too, out-competing your company on IT. So be prepared. Start doing this, prepare for success and get moving.
Dresner: CIOs do need to think of themselves as a CEO. For years, many CIOs have struggled with having parity or achieving parity with their other C level counterparts. And one of the ways to do this is to come armed not only with insight about your own business in a comprehensive way, but to be able to interrelate it and show how it supports the other businesses. So I think that helps to elevate the CIO in the discussion and make them more of an equal amongst the peers, the CFOs, the COOs, the CMOs of the organization. And I see analytics as an important component of that. If you’re going to be a serious person, you’ve got to have serious data.