For the great majority of years in the past decade, Chief Information Officers named “business intelligence and analytics” as their top focus in Gartner Inc. annual surveys of technology priorities. That set of technologies moved to number one in the survey in 2006 and stayed there until 2009. It fell to fifth in 2010 and 2011, but was back on top in 2012 and has stayed there ever since.
One might think, then, that CIOs would put a major focus on using BI and analytics in IT operations. Nope. Many IT organizations have been enthusiastic about applying analytics in various forms to marketing, HR, supply chain, and other aspects of the business, but not to themselves. IT often strains to provide basic metrics about its own operational performance or to report them in anywhere near real time. Yes, most IT organizations report on their uptime, but that’s about it.
The rest of the business is increasingly focused on moving beyond descriptive analytics/reporting to predictive and prescriptive analytics – analyses that predict the future and tell employees how to do their jobs better. And there are a variety of applications of such advanced analytics in IT operations; there is a helpful list from Gartner on Wikipedia. They include, for example, predicting when key components of the IT infrastructure might develop problems and recommending resolution strategies for system outages. And there is clearly a great opportunity to begin using analytics to predict security breaches rather than responding to them ex post facto. However, the fact that this list exists doesn’t mean that many IT organizations have actually put those applications in place. I see very few examples of predictive and prescriptive applications in IT operations.
There are many possible explanations for why IT has become the shoemaker’s children with no analytical shoes. One would be that IT is too busy helping others with analytics to work on itself. Another would be that it doesn’t think that detailed metrics and analytics would reflect well on itself. A third explanation might be that CIOs and other senior executives feel that the payoff is greater for non-IT analytics. A fourth would be that there just isn’t enough data to analyze. This latter factor seems unlikely, as computers are pretty good about generating data on their own performance. And there are increasingly providers of software that specifically address log file analytics. In fact, I don’t see any of these as good reasons for avoiding IT operations analytics.
There are, of course, some negative implications of IT’s not taking the lead on analytics. Had it done so, the function would be much more credible in arguing that other functions should analyze and understand their performance. If business leaders see IT as “eating their own dog food” – an attribute they also like to see in vendors – they’d find IT a more credible source of help on analytics. And I can’t help but think that the lack of analytics on IT’s operational performance is one key reason why IT is hardly loved by many of the firms in which they toil. It may also be the case that if IT doesn’t measure itself well, business leaders could be more likely to do an end run around IT and build their own technological capabilities.
It seems likely that we will be seeing an even greater need for analytics on IT in the future. As every business becomes more digital, there will be more things to measure and a greater demand to know how every important system is performing right now and how it is likely to perform in the future. Focusing on IT operations analytics also could make it easier for the function to adopt new automation-oriented tools that manage mainstream operational IT tasks. As I’ve argued on this site, cognitive technologies are a straightforward step beyond traditional analytics, so it would help to have some experience with the latter before adopting the former.
Whatever the reasons and the implications, there’s not much doubt that the practice of analyzing IT performance data needs to improve. If your IT function is still just releasing monthly uptime statistics, you really need to up your game. Analytics is not just a service you provide to your internal customers, but something you need to apply to yourself.
Tom Davenport, the author of several best-selling management books on analytics and big data, is the President’s Distinguished Professor of Information Technology and Management at Babson College, a Fellow of the MIT Initiative on the Digital Economy, co-founder of the International Institute for Analytics, and an independent senior adviser to Deloitte Analytics. He also is a member of the Data Informed Board of Advisers.
Subscribe to Data Informed for the latest information and news on big data and analytics for the enterprise, plus get instant access to more than 20 eBooks.