Every process should have a systematic set of measurements (Key Performance Indicators – KPIs) against which its performance can be tracked, communicated and improved.
The set of measurements may include key performance results, customer results, people results and society results. The measurements you select should be based on the purpose of the process and what you are trying to achieve.
You will probably only need 4-7 KPIs in order to manage and continuously improve your process. If you have too many, it probably means you don’t understand what is really important about your process’ performance.
Checklist of possible measurements:
Which of the following are relevant?
|Internal Measures||Output Measures||Satisfaction Measures|
|– Processing time (work time in process steps)||– Error Rate or Accuracy (Right First Time)||– Perceptions of reliability, assurance, tangibles, empathy, responsiveness|
|– Cycle-time (end-to-end, elapsed time)||– Timeliness (delivery vs. deadline/requirement)|
|– Delay or Waiting time (e.g. between steps)||– Completeness||– Any “objective” measures gathered by customer(s) or stakeholder(s)|
|– Volume (input)||– Conformance to Standard|
|– Cost (direct cost per transaction)||– Success Rate/Attrition Rate/Output Volume||– Compliments|
|– Overhead cost||– Complaints||– Awards|
Internal measures enable you to assess the basic performance of the process itself. Output measures enable you to assess the quality of the intermediate or final outputs. Both can be measured without involving the customer(s) of the process. Satisfaction measures are direct assessments of the customer’s view of the process and can only be gathered by asking the customer.
For each measurement you select, you need to define:
- what it is (a precise definition)
- how you will gather the data (including sample sizes)
- how often you will gather the data
- how often you will report and review the data (including the format in which the data will be presented)
- any targeted levels of performance (if known)
- who is responsible for measurement
More on Process Measurement.