IT operations is under increasing pressure to provide meaningful key performance indicators to business managers. Generic CPU utilization and uptime figures aren't good enough. Business wants more.
Modern KPIs for IT focus on the end-user experience first, rather than infrastructure metrics. Organizations can also tailor KPIs around the digitization of business operations, even into specializations that come with industry verticals.
Is IT doing a good job?
IT must have end-to-end metrics for itself that provide meaningful information. "IT needs to get out of the shoemaker's children syndrome," said Greg Schulz, analyst at StorageIO, referring to the proverb wherein a shoemaker's family goes barefoot because he is too focused on his customers. IT mustn't focus exclusively on business KPIs to the neglect of its own.
End-to-end KPIs for IT should still have business context, but IT must see the big picture: User or client experiences exist on one end, while resource use, consumption and performance exist on the other. Customer experience equates generally to application latency and throughput, followed by performance-related metrics, such as mean time to detect and resolve issues.
IT needs to get out of the shoemaker's children syndrome.
Greg SchulzAnalyst, StorageIO
"It's not just usage or utilization, but also responsiveness, errors or retransmissions [when network transfer fails at least one time], how productive the user is and happy with the service, as well as how effectively resources are consumed to deliver that service experience," Schulz said.
The business side wants to know the costs associated with IT activity and whether those costs yield good value in business outcomes. Take the time to meet with business management and determine the metrics and KPIs that matter to both IT and non-IT departments, Schulz said.
KPIs for IT-enabled digital businesses
The narrative around KPIs for IT has shifted from solely traditional metrics to data that supports broader business outcomes, said Kristi Lamar, managing director of the CIO program at Deloitte, a multinational accounting and professional services organization.
Beyond individual application performance, IT organizations should measure and improve the speed of application development and deployment pipelines, as well as how closely those applications match customer needs. In addition, IT organizations must create technological solutions that support changes to the business's operating model, she said.
Senior-level business managers -- and even C-level officers -- must communicate what their business objectives are, and senior-level IT directors and managers should offer ideas on how IT can help. "IT and the business have to agree on what to measure: If there's a mismatch, it's not time well-spent," Lamar added.
For example, IT knows the cost of database transactions but must also translate that expenditure to the business's productivity initiatives. "[This information either] impresses the business or makes them ask what it would cost to double the [database's transactional] performance," Schulz said.
Tools to improve IT performance
Incidents that take down an application are rare, especially in distributed environments, wherein there are fewer, if any, single points of failure. However, these complex architecture layers make issues more difficult to locate, said Rich Lane, analyst at Forrester Research. The IT monitoring tool space is making strides toward a dependency mapping model, which shows the active infrastructure and application layers in use. Rather than rely on distinct application and infrastructure performance monitoring tools, organizations seek out tools that do both simultaneously to determine both end-user experience and performance bottlenecks that cause problems.
General and specialized industry KPIs
In the past, IT's role to provide stability and keep the lights on didn't change much from one business to the next. Now, the pressure is on in modern organizations for innovation, new service development and business-focused technological operations. Those aspects of digital transformation start a shared discussion between business and IT, according to Dennis Drogseth, analyst at Enterprise Management Associates, into a realm beyond generic KPIs.
Every industry vertical has a specific set of relevant KPIs related to IT infrastructure and services. For example, agricultural companies that use IoT devices rely on IT to track IP addresses associated with cattle locations on a farm. Automotive manufacturers require IT KPIs around customer-facing digital operations within the vehicle, such as Bluetooth functionality, as well as a wholly different set of KPIs for part tracking and assembly sequences on the manufacturing floor, Drogseth explained.
With the advent of IoT and digital business enablement, as shown in the retail space, IT's responsibilities cover customer-facing as well as business operations technologies, each with specialized KPIs.
But organizations don't have to reinvent the wheel to evaluate IT against each business requirement. Generally, use KPIs that connect the dots of application performance. For example, track metrics for application uptime, combined with those for incident detection and resolution of system errors, and use those KPIs to set service-level agreements or compare IT's performance to established ones. This information gives the business a sense of whether IT meets customer expectations in a given area.
