Types of insights
Association insights
Association insights recommends which subgroups to investigate because they are performing poorly in one of your selected metrics. Subgroups are segments of the overall data set that share one or more attributes. These subgroups are sorted by potential improvement of the global value of the metric if it is brought to the level of the corresponding population outside the subgroup.
In the sample insight below, the Invoice Creation to Post metric measures the time taken between when an invoice is first created and when it is officially posted in the accounting system. A low value means invoices are processed efficiently. In this insight, the global value of Invoice Creation to Post is 19 days. The subgroup being considered is a combination of two attributes: Posting User = USR-20218 and Vendor = 0010810981 - Company-6945. This insight shows a potential improvement in Invoice Creation to Post of 3 days, from 19 days to 16 days.
To see further explanation of an association insight in Insight Explorer, hover over the info icon.
In the explanation you see that this subgroup contains 5.41k invoices with an Invoice Creation to Post time to process of 153 days. The time to process these cases is much higher than the value of 16 days for all other Posting Users and Vendors which shows an easy area for improvement.
Process attributes
Association insights include process attributes which describe your process, do not need to be added to the Knowledge Model, and are provided out of the box. Insight Explorer uses the following process attributes:
Process attribute | Description | Visualization |
|---|---|---|
Number of Distinct Events coming soon | Number of distinct events in the process. |  Process complexity |
Total Throughput Time coming soon | Time between the first and last events. |  Long-running |
First Event | Name of the first event that occurs in the process. |  |
Last Event | Name of the last event that occurs in the process. |  |
Day of First Event | Day of the week on which the first event falls. |  High demand |
Event Occurrence | Indicates whether each event occurs in the process. |  Undesired event  Skipped event |
Event Rework | Indicates whether each event occurs more than once in the process. |  |
Edge Occurrence coming soon | Indicates whether each edge occurs in the process. |  Wrong order  Ping pong |
Edge Throughput Time coming soon | Classifies each edge as Bottleneck or Rushed based on the deviation of its throughput time from the median. |  Bottleneck  Rushed |
Event Attribute coming soon | Takes the value of each attribute in each event, such as whether an event is automated. |  |