Custom metrics

Custom metrics are a feature in IBM Process Mining that helps you to apply a custom SQL query at the case level of a project to create a business metric. You can create the SQL query by using the available data in IBM Process Mining to compute a user-defined result. Therefore, you must have a good understanding of the SQL and the data structures in IBM Process Mining to create a custom metric.

You can use custom metrics for the following activities in IBM Process Mining:

You can use custom metric created with SQL based on MonetDB taxonomy to compute your own metrics and then leverage them in the widget and dashboard filters. For more information, see MonetDB taxonomy.

Using custom metrics in a project might impact the performance of process analysis. To maintain optimum performance, limit the number of custom metrics and ensure their SQL query is optimized. If the custom metric is invalid, you see an Error icon Error icon next to it and its Show statistics toggle is disabled.

The following sections explain the tasks that you can do in the Custom Metric section in IBM Process Mining:

Accessing the Custom metric dashboard

To access the Custom metric dashboard, do the following steps:

  1. Select a process-mining project from the Projects table in the Projects tab.

  2. Go to Data & Settings > KPIs > Custom metrics.

Creating a custom metric

To create a custom metric in IBM Process Mining Legacy, do the following steps:

  1. On the Custom Metric dashboard, click Create a custom metric.

  2. Enter a name for the custom metric in the Name field.

  3. Select one of the following data types from the Data Type dropdown:

    • Boolean To get True or False as output based on the query.

    • Date To get the date as output in the <YYYY-MM-DDTHH:mm:ss.sssZ> format.

    • Numeric To get an integer or decimal number as output.

    • String To get the text string as output.

    • Duration To get the duration (in milliseconds) as output. When you select this data type, IBM Process Mining generates the custom metric result as a duration in milliseconds.

  4. Optional: Add a description if you need to give any details about the metric.

  5. Optional: Switch on the Show custom metric statistics if you want to view the statistics of the custom metric in the Model tab and Statistics tab.

    The switch of the Show custom metric statistics does not impact the use of the custom metrics in the Analytics dashboard.

  6. Enter the SQL code in the Custom metric measure field.

    1. Select a value from Group by dropdown to complete the SQL query.

      You can group by Case ID, Event, Project, by any event log column, as well as Resource and Role.
      When you create custom metrics that are grouped by any event log column or the Event value, a case distinct function will be applied in the background after the creation of the custom metric only if the metric's measure includes a function at case-level.

    2. Optional: Enter a clause in the Where field.

  7. Optional: Turn on the Activity interval switch.

    1. Select the interval settings by using dropdowns.

  8. Optional: Turn on the Keep last event for each case.

  9. Click the Test Code button to validate the custom SQL query.

  10. Click Next to go to the Additional configuration step.

  11. Optional: Enter the unit of the custom metric result in the Unit field if you have any specific unit to display in the custom metric statistics.

  12. In the Assign to categories section, you have two options: Select Existing category, and the Create New category.

    By default, IBM Process Mining categorizes the custom metrics in the General metrics category.

  13. To select an existing category, do the following steps:

    1. Choose the Select Existing radio button in the Assign to categories section.

    2. Choose a category by clicking a related category tile.

  14. To create a new category to link custom metric, do the following steps:

    1. Choose the Create New radio button in the Assign to categories section.

    2. Enter a name in the Name field.

    3. Enter a description in the Description field.

    4. Select an existing icon or add an icon in the Icon field.

    The maximum file size of the icon that you can upload is 200 KB.
    You can upload only the SVG, JPG, JPEG, or PNG format.

  15. Click the Create button to complete the creation of a custom metric.

You can see details of the created custom metric under the assigned category. To do so, go to the Custom metrics table in the Custom metrics page.

Editing a custom metric

To edit a custom metric, do the following steps:

  1. Go to Data & Settings > KPIs > Custom metrics.

  2. Click the Edit icon.

  3. In the Edit custom metric dialog, you can change the following fields in the Details step:

    • Description field

    • Show statistics switch

    • Custom metric measure field

  4. Click Next to go to the Additional configuration step.

  5. In the Additional configuration step, apply the changes that you want in the following fields:

    • Unit

    • Category

  6. Save your changes.

Deleting a custom metric

When you delete a custom metric, the output of the existing Analytics widgets, which apply to the deleted custom metric becomes invalid. Also, the associated project filters are deleted.

To delete a custom metric, do the following steps:

  1. Go to Data & Settings > KPIs > Custom metrics.

  2. In the Custom Metric page, search for the custom metric that you want to delete.

  3. Click the Delete icon to see the Delete custom metric window.

  4. Delete the custom metric.

When you restore a custom-metric based backup of a process model, IBM Process Mining re-creates only the custom metrics that are relevant to the selected project.

Filtering custom metrics by category

To filter custom metrics by category, do the following steps:

  1. In the Custom metric dashboard, click the All dropdown on the upper left of the window to see the list of categories.

  2. Select a category from the All dropdown.

Learn more

Example: Creating a custom metric to analyze the opportunity management process