A few years ago, I worked with IT executives at a major financial institution eager to see its data in our HP Executive Scorecard product. The company took great pride in its incident management capabilities. However, when I saw the incident data, I was very surprised. In this post, I’ll review this customer’s situation, how it compared with benchmarks, and why this matters.
The customer gave us four months’ worth of data—October through January. The data indicated 20,000 incidents per month in October. But in the months that followed, incidents grew to 65,000 incidents per month. Even worse, incidents resolving in a day or less dropped by 75 percent. Additionally, core systems failures quadrupled. In a meeting with the customer team, I asked what happened; they just smiled. When I persisted a bit, they told me they had successfully delivered their end-of-year changes. But financial institution customers like to touch their money at end of year. And let’s face it—their change process was causing systems to fail.
According to ITIL Version 3, 55 percent of incidents/outages are self-inflected. I never did get the customer’s change success rate, but I can guarantee its change success rate really went down after October. So what is a good number? We asked HP customers to share their opinion on this and other typical service management metrics. Additionally, we asked them to share (confidentially, of course) their actual performance.
What’s a good benchmark for change success rate?
By taking a weighted average of these customers’ responses, I determined that our participants felt the benchmark for change success rate should be 95 percent or better. However, only 43 percent of respondents reported having a change success rate of 90 percent or higher. To really find out what the benchmark number should be, we would like a bigger group to participate in our survey. So, we want you!
If you would like you to vote on this and other service management benchmarks, please click the link below and confidentially share what you think the benchmark performance should be for change success rate and other important service desk measures: http://svy.mk/QoeuES.
Why do benchmarks matter?
How much value can reducing outages bring to improving IT’s reputation? More important, how much cost reduction can it bring? To me, this is the most important area of improvement for IT organizations. World-class organizations reduce costs by being better. Change success rate can be improved by:
- Improving testing
- Creating a more homogeneous environment
- Driving constancy of the IT environment
- Limiting unplanned changes
- Eliminating unapproved changes
How much is this worth? It’s all in your hands. COBIT 5 says that the change management process is about enabling fast and reliable delivery of change. In one word, it is about establishing predictability. What do you think? I would love to hear back from you about this post.
Solution page: IT Performance Management