If your IT organization is typical, its analytics consists of casting a wide net to surface myriad metrics around, for example, ticket volumes, ticket duration, number of people working on tickets, and incidents by different severity levels. These “did you know?”-type data points are good for populating dashboards, but the value they provide is debatable, as they lack insight into specific business challenges such as, “Why are incidents taking so long?”
A real value-producing analytics capability possesses two key characteristics: a top-down approach, and a project orientation.
A top-down approach means always beginning with a defined business question, e.g., “How do I reduce the overall incident load to run a more efficient organization?” It then deconstructs the question to allow focus only on relevant data elements, patterns, and trends, and to enable identification of causality, relationships, and correlations that drive specific actions.
Operating in a project-like environment enables the analytics team to generate relevant findings and outcomes. Essentially as with the Agile development methodology, analysts can act on the insights they’ve unearthed from the narrower set of data indicated by the business question. For example, they may toss some information that proves irrelevant to the issue at hand, move the data into the company’s regular reporting package, or iteratively dig deeper into specific data points until they find the answer that will drive a specific action. This constant iteration provides increasingly deeper insights into trends and patterns, thereby enabling the business to ask better questions, leading to progressively more informed business decisions.
Consider this real-world example. A large financial services institution had deployed a brand new client servicing platform that was integrated with several of its legacy back-end systems. As the platform was to be used by customers, the bank expected a significant increase in the workloads on the back-end systems. It implemented a predictive performance management system to track utilization of the back-end systems at a granular level, and correlate utilization to workloads to identify potential threshold breaches and performance bottlenecks.
Upon implementation, one of the key databases in the back-end crashed intermittently. Because this was a highly customized deployment, the bank’s IT department was unable to pinpoint the cause. An analytics team stepped in and began visualizing the performance and event data captured by the performance management system as the starting point for answering the overarching question, “What is causing this database to crash?”
The analytics team first recognized that utilization associated with the crashing database spiked at regular intervals. Digging deeper, they identified that the first crash instance happened during the weekend following the initial deployment, but that the database infrastructure didn’t show any changes. Their further analysis of platform logs found a correlation between the spikes in database utilization and a particular clean-up task that ran every weekend. The analytics group was finally able to determine that a code issue – which showed up whenever the cleanup tasks were complete – was actually causing the database to be overloaded and eventually crash.
To operate a real value-producing analytics capability, you need to ask the right business questions, and continue to dig until one or more specific actionable solutions are revealed.
As this financial services firm learned, the traditional boil the ocean approach to analytics doesn’t work, as it results in shallow data that lacks root cause information with which to make informed business decisions. To operate a real value-producing analytics capability, you need to ask the right business questions, and continue to dig until one or more specific actionable solutions are revealed.
About the Authors:
Chris has over 15 years of business and advisory experience across multiple disciplines including IT, strategy, and finance. He specializes in IT strategy, cost optimization, and technology transformation.
Hari has more than 18 years of management advisory and technology experience, with particular emphasis in IT and enterprise architecture strategy and delivery.
To learn more, read KPMG’s whitepaper Running the Business of IT: Metrics That Matter, and listen to the podcast IT Analytics Shape the Future of the IT Organization.