IT Service Management Metrics in Action, Average Handling Time
Goodbye Metric of the Month, hello new format. Yes, for those of you who have followed Northcraft for some time, you’ve probably noticed that we gave Metric of the Month the boot some time ago. The problem was that it was too abstract… and a bit detached from reality. We realize that you want to know how <em>actual</em> customers use our metrics catalog in the field, not just how big and fancy our metrics catalog is, so we’ve finally arrived at our final destination… <strong><em>Metrics in Action</em></strong>
This story comes to us from one of our most recent new customers. Here’s a bit of a backdrop:
<strong>Company Size:</strong> 3,000 employees
<strong>IT Department</strong>: 1,000 employees (how important is IT to them?)
<strong>ITSM Solution SaaS or on-Premise</strong>: SaaS
<strong>Process areas implemented</strong>: Incident Management, Change Management, Configuration Management, Service Level Management and Problem Management
<strong>Metric in Focus</strong>: <em>Average Handling Time, Incident Management. </em> This is not to be confused with an ACD metric. The measurement is based on the time it takes to fill out all of the necessary fields on an Incident <em>initially</em>. Also, this particular Incident Management system stores all requests & incidents as <em>incidents</em>. There are ways to filter out desktop support related issues and requests by using various fields, so please don’t go ITIL-crazy in the blog comments.
So, here’s the situation… in week 1, we install our solution and begin digging through the data for the first time (this is the most fun part for everyone, metrics discovery), looking for abnormal data, inconsistencies, patterns, etc. We always begin with the simple, most widely-known metrics. So, we pull up Incident count… 3,000/month… FCR for the Desktop support Groups… 88%? Wow! <a href=”http://www.linkedin.com/profile/view?id=177205&authType=OPENLINK&authToken=bUB9&trk=hb_upphoto”>Eric Zbikowski</a> will tell you that’s world class… nice! Average Handling Time… 6 Minutes. WAIT. What is that? It takes 6 minutes <em>on average</em> for the Incident form to be <strong><em>initially</em></strong> filled out? Our other customers are always between 2-3 minutes across the board. There are occasions where handling time goes up IF agents are resolving the issue on the call (which is not to be discouraged), but that wouldn’t be experienced across all incidents. So, this is definitely odd.
<strong>Pause for a moment</strong>… Aren’t we currently hearing the theme in social media that we need to focus only on the metrics that matter? Do you see Average Handling Time mentioned as a KPI for Service Desk? Lee – Don’t you have a presentation coming up at Fusion called, “Metrics that Matter, A Better Balanced Scorecard for IT”?
Yes, that’s true.
Let me ask you this – Just because we provide 1,100 metrics out of the box, do you have to drag all of them into a report?
So, please don’t go crazy on me with metrics methodology right now either.
<strong>Un-Pause</strong>– So what’s the deal with the 6-minute average? It turns out that since this was a SaaS-based service desk, MacAfee virus scan was treating pop-up windows in a more hostile fashion. So, when a user was clicking on the category link pop-up, the virus scan was ensuring that nothing bad was going on. Moreover, the desktop support people had been complaining for weeks that the performance of the SaaS ITSM solution was incredibly slow. So, we were presenting this in a meeting to the Director of Change Management whose response was… “known issue, talk to the Director of Network Security”. So, we do that (whilst create a known error record)… and find out that it’s impossible to configure MacAfee not to scan the pop-up (which I’m still not sure about, but I’m not a netsec guy).
Bottom Line: Their team re-configures ITSM application to use a tab instead of a pop-up… boom, problem solved. Desktop Support totally loves us! They have 3 minutes of their life back on every single incident. Yes, 3 minutes/agent. So, which metrics matter most?
<strong>Not a bad day.</strong>