Business Mentor
If You Can Measure IT, You Can Manage IT

After being forced into my first IT management role, which is a story for another column, I discovered my IT department lacked credibility. Much like the Keystone Cops, it never got anything done. I'd heard the wise adage "If you can measure it, you can manage it," and I took those words to heart. I tried to define a set of metrics we could use to measure our performance and thereby motivate us to improve.

Measurement, Not Micromanagement

But immediately, I hit a roadblock: Just how should we measure IT performance?

I started off badly by defining a set of metrics that micromanaged behavior and, ironically, discouraged accountability (such as how many lines of code a developer should complete in a day). I also defined too many metrics. I decided, for example, that the help desk -- our primary interface with our business customers -- should collect cause-and-effect relationship information to help IT become more proactive.

If 30% of help desk calls involved user problems like "I can't find the document I was using 15 minutes ago," that meant we should conduct Windows and Office training. To further improve proactivity, I made one of the help desk performance metrics the number of calls received. I thought this would encourage the help desk to use this information to ultimately reduce the number of help desk requests. And indeed, the help desk did reduce call volume by an impressive 50% -- but only by unplugging half the telephone lines.

Back to Basics

After several similar missteps, I had to identify critical IT tasks to focus on. I also wanted to ensure that our performance metrics were obvious and simple to measure. So I started over, focusing on the basics. What are our primary activities? In what areas do our customers need us to improve? I came up with three areas to measure.

Application delivery. We deliver applications to internal and external customers, and those applications should perform. For applications, our metrics were reliability (i.e., uptime) and performance (response times).

Project delivery. Almost everything IT does -- from implementing a new application to installing a patch -- is a project. So our metrics covered all aspects of proper delivery, such as on-time, on-budget and on-target measurements.

Customer service. Like it or not, we are in the human-interaction business. So we should observe basic customer service standards. We should over-communicate what is going on and why, present options, and let customers make the best choice for them. For customer service metrics, we instituted frequent customer satisfaction surveys to track how we were doing.

Measuring performance in these three areas helped turn IT into a high-performance, high-credibility machine. Taking stock of application performance eliminated the finger-pointing, allowing us to solve problems like slow response times. Paying heed to customer satisfaction helped transform IT into a group our customers want to work with. And tracking project deliverables turned everyone into rigorous project managers. We worked more closely with our project sponsors and never let projects spin out of control.

And now I have a predefined set of IT metrics that I can use in multiple situations with the same great results.

Digg This!     StumbleUpon     Del.icio.us