Do you measure up?

IT departments have been measuring speeds and feeds for years now. Is that good enough?

Peter Drucker, the management consultant once said: “If you can’t measure it, you can’t manage it.” Of course, IT departments are right into the business of measurement. We track user statistics, uptime, downtime, memory use, CPU utilisation, disk space, who’s using what systems and report and myriad other metrics. But are we adding any business value? Are we measuring what’s important or what’s easy to measure?

In late 2012, HP commissioned Coleman-Parkes Research to conduct a study into the changing nature of IT performance measurement. The 600 telephone-based interviews conducted across the Asia-Pacific region amongst IT and business executives found, unsurprisingly, that “measuring IT performance is considered very important”. The field of respondents was evenly distributed across Japan, China, India, Australia, Singapore and Korea.

A couple of the key outcomes of the survey were that most organisations were focused on quality-of-service measures but that business metrics hadn’t yet found their way into IT reporting consistently.

It’s clear looking at the survey data that most IT managers are measuring the same things they always have. The two most common measures were quality of service and speed of response/ticket resolution. In other words, are the systems up and is the help desk responding quickly when there’s a problem.

What’s surprising is that almost two-thirds of the respondents are still using manual processes for their performance measurement. As one respondent put it: “Yes, some challenges do exist, mostly around the collection and presentation of the data itself. They are not so easy to analyse because so many methods and systems of collection are there for different areas and it is not easy to compile it.”

So, the question is - are IT departments letting the business down with what they measure?

Peter Lilley, Director of Presales, South Pacific, HP Software, answers: “I think the survey speaks to something of a disconnect. But it’s specifically about the idea of IT performance and how we measure performance.

“With some of the enterprise IT customers we see, this is certainly a theme we see. It’s pretty universally accepted that stronger IT performance measurement is a very important thing but that IT executives don’t necessarily have the tools or the methods to do it and the disconnect is clearly around the execs want to demonstrate their contribution and that the value that they bring to the broader business themes and business performance issues but often their performance systems are stuck in a legacy of IT metrics.”

Our IT departments are in a state of flux at the moment. With the shift away from centralised infrastructure, the focus is no longer on quality-of-service measures such as system uptime or throughput.

“The thing to aspire to is a system of record or the kind of IT performance system that is not that conceptually dissimilar to the performance systems that IT has been promoting to other lines of business for years,” Lilley added.

Finance teams have systems for closing the books each quarter. Manufacturing and logistics have systems for optimising inventory and schedules. And IT seems stuck in a previous age when it comes to measurement systems. One of the challenges is translating the traditional ‘speeds and feeds’ view of measurement into something that is presented in business, rather than technical, view.

Rhys Evans, CTO of IT consultancy firm Thomas Duryea, says: “The traditional view of uptime is that if the server is switched on, can someone ping it or access it. That’s good but what we’re seeing, for example at the Commonwealth Bank, is that focus is on uptime but around applications.”

By switching the IT measurement away from infrastructure to applications, IT can prove its value in a couple of ways. Firstly, performance is measured in terms of completion of transactions in a timely manner. In banking, the ability for a customer to access an ATM and successfully complete their transaction in a few seconds touches on almost every element of traditional IT performance measurement from both an availability and performance point of view.

“If the customer can’t complete the transaction within the allotted time, then the application is down. There’s less focus on speeds and feeds and more on the customer interaction,” Evans said.

This approach drastically reduces the disconnect between what IT measures and what is valuable to the business. In the early days of e-commerce, companies like Amazon were concerned about the number of virtual shopping carts that were abandoned by customers. Although systems were available, it was clear that customers were losing patience. By focusing on business metrics, like the time taken to complete a transaction, IT was able to better understand what it needed to measure and then act on it.

One of the shifts that is forcing businesses to consider this style of IT measurement is the move to cloud services. Cloud-based systems put businesses in the position where application availability and performance is a key metric.  Evans says: “They’re thinking about cloud-based software and applications. They are thinking that if they have a CRM, ERP or line-of-business application then I should score my providers based on the uptime and performance rather than whether it’s just operational.”

With applications now far more easily available - anyone with a credit card in the business can order an online application - IT needs to ensure that its measurement systems are in line with those of the business. The era of speeds and feeds being key metrics is over.

Image credit ©iStockphoto.com/Henrik Jonsson