Start with the Outcome in Mind
There are many approaches to developing dashboards. As I stated in my last post, we often apply the focus on the KPIs without first talking about what we are trying to learn from those KPIs. I can’t tell you how many times I have heard the question…
“What KPIs does everyone else look at?”
Viewing commonly accepted KPIs usually gives a common result, we end up looking at measurements without context. Understanding why we measure something is as important as the measurement itself. Focusing on outcomes first will ensure the context for each measurement.
In this blog post, I will describe an approach to developing a dashboard for a specific outcome. I will follow the framework I presented in my last post.
- Critical Success Factors
- Key Performance Indicators
I will also point out a few design considerations for implementing ServiceNow to achieve an outcome.
Outcome Package: Effectively Manage Service Desk Staffing
The Service Desk is the single point of contact for users interacting with an IT organization. Staffing requirements for a service desk are impacted by a number of factors and require continual review to ensure quality delivery of service while controlling the cost of resources. The three critical success factors to be considered when determining staffing needs are;
- Managed User Expectations of Service
- Understand a Dynamic Workload
- Controlled Resource Costs
CSF: Managed User Expectations of Service
Customers will create their own expectations unless you present them with a defined policy. The most effective way to understand customer’s expectations is through surveys. The most effective method for communicating your expectations is through Service and Operational Level Agreements.
- Customer Satisfaction Survey Responses
- Incident Response Time SLA
- Incident Resolution Time SLA
- Request Fulfillment Task OLA
- Request Item Fulfillment SLA
- Incidents not updated in the last 5/30 days
- Utilization of Self Service Support Tools (KB)
CSF: Manage a Dynamic Workload
Service desk capacity should never exceed the required workload. Excess staffing is a cost that cannot be recovered. The increase of incoming work compared to the completion of existing work can be caused by the release new or changed services, planned maintenance activities or even agent sick days. Historic information is incredibly valuable to predicting the effort ahead. The mean time to resolve incidents is a simple predictor of how long it will take a backlog of incidents to be resolved. Catalog tasks that have accurately planned durations can create set an expectation of what work is to come.
- Incident Backlog Growth
- Request Backlog Growth
- Fulfillment Task Backlog task duration
- Mean Time to Resolve
- Total Planned Durations of Catalog Tasks for the next 7 days
- Total Planned Durations of Catalog Tasks for the next 30 days
CSF: Control Resource Costs
Understanding of the cost of resources required to manage the workload is not often as simple as counting the number of people at the service desk. The total cost of the resources is different than understanding the cost of completing work. It’s important to understand how service desk agents are spending their time and how much effort is spent resolving incidents or fulfilling requests.
- Number of service desk agents on duty
- Average number of Incidents resolved per agent
- Fulfillment task duration
- First Call Resolutions
- First Line Resolutions
- Total number of incidents resolved
- Mean time to resolve
- Cost of incidents resolved
Overview with Single Scores
Collections of KPIs can be modeled into indices. An index combines multiple indicators and can use weighting to allow one score to have greater influence than another.
Index of Meeting Customer Expectations
For example, a customer satisfaction rating might outweigh the frequency of meeting in incident response SLA.
Index of Meeting Customer Expectations =
((Average Customer Satisfaction on a 10 Point scale) x 10) x 0.6 +
(Percentage of Incident Response Time SLA Met) x 0.2 +
(Percentage of Incident Resolution Time SLA Met) x 0.2)
Index of Managing Dynamic Workload
In this next example, we consider the total backlog of incidents and the planned durations of catalog tasks due in the next 7 days to calculate the upcoming workload. We divide the total number of work hours available by that workload to understand how much work will be able to cover at the current staffing levels. Less than 100% and it appears understaffed, over 100% is over staffed.
Index of Managing Dynamic Workload =
(Service Desk Staff * Work Hours Next 7 Days) ÷ ((Incident Backlog * MTTR) +
Total Planned Durations of Catalog Tasks for the next 7 days)
Single Score for Managing Resource Costs
Scores related to cost make the most sense in local currency. It’s possible to use the Task Rate Cards to generate expense lines as incidents are resolved. A database view joining the expense lines to the incidents allow for building indicators showing total and average costs of incidents. This topic will be covered in future blog post.
Understanding the outcomes, CSFs and KPIs will help drive your understanding of configuration requirements for the platform.
- Incident Management is the primary responsibility of the services desk. Incidents are typically prioritized above other service desk activities. Incident management also produces service level metrics and can initiate customer satisfaction surveys.
- Customer Satisfaction Surveys must be configured
- SLA Definitions for response time and resolution time are required
- Incident metrics for first call resolutions could be useful
- Service Requests are often originated and in some cases fulfilled by the service desk. Fulfillment tasks can be a source of future work that can be planned for.
- Configure request item workflow with consistent use of planned start dates, end dates, and durations for fulfillment tasks
- Service Desk Resource Management is not a typical ITIL process but is an important factor to consider. Manual indicators can be used to store the number of employees on duty or the number of person-hours available to work.
- Day of the Period Table is something that I have found useful. Each day of the year has a position within a Month, Quarter and Year. I’m planning a separate blog post on this topic in the coming weeks.