Defining KPIs

The Add/Edit KPI page in Optimize enables you to create and edit KPIs. You can activate this page from the Business Data page in Optimize for Process or from the System Data page in either version of Optimize. For more information about the elements that make up KPIs and a description of how KPIs function within Optimize, see the “Business Visualization Concepts” chapter in webMethods Optimize User’s Guide. Also, when creating KPIs it is helpful to understand how event maps function in Optimize. For more information about event maps, refer to Mapping Events.

Important:
When Analytic Engine runs in Static DB Schema mode, you cannot create, edit, or delete dimensions, event maps, dimension hierarchies or KPI definitions. In addition, all respective buttons in the user interface will be unavailable. For more information about enabling and disabling Static DB Schema mode for the Analytic Engine, see the PDF publication Configuring BAM.

Individual KPIs associate one measure with one dimension attribute, such as revenue by region or queue length by host. When you define an individual KPI, Optimize creates a KPI instance for each attribute of the dimension defined for that KPI. For example, if you define a KPI to measure revenue by region and you have three regions, Optimize creates three separate KPI instances to collect revenue data for the three regions.

Composite KPIs consist of the result of a mathematical operation performed against two other KPIs. For example, if you defined individual KPIs for a product's operating costs and revenue, you can define a composite KPI for the product's profit by subtracting the operating cost KPI from the revenue KPI. A composite KPI must be set up on individual KPIs that have the same collection interval, and dimensionality. Also, composite KPIs should be set up only on two “Sum” calculation or “Average” calculation KPIs. Although it is possible to set up composite KPIs involving a combination of “Sum” or “Average” calculation KPIs, the resulting calculations may be incorrect.

Note:
When you define an individual or composite KPI for which applicable data has already been collected, the KPI instances created for that KPI are automatically populated with that existing data. This enables you to determine a KPI's effectiveness without having to wait for new data to be collected. See KPI Back Population for more information.

There are two sections on the Add/Edit KPI page that enable you to define or update the characteristics of a KPI. These sections are KPI Information and KPI Definition. These sections and their related fields are explained below.

KPI Information: This section enables you to assign a name to the KPI and define a metric (including an appropriate unit of measure, a calculation type appropriate for data aggregation and a data collection interval). In addition, you can define a naming template which helps you to identify data on graphs and tables throughout Optimize. Finally, in this section you also set Empty Data Set Behavior (EDSB) for the KPI.

KPI Definition: The fields in this section enable you to define the KPI type (individual or composite) and to associate a fact with a dimension (a means for breaking out your data into slices), such as revenue by region or queue length by host. This creates a KPI Instance. In addition, you associate event maps and attributes with the KPI and you define a hierarchy.

Because of the inherent complexities involved in creating KPIs, you typically move back and forth between these two sections when defining a new KPI.

The fundamental factor to consider when creating a KPI is the event map selection. The event map you select defines the available dimensions, measures, naming templates, and other variables for the KPI. So, when creating a KPI it is important that you are familiar with the event maps configured on your system and know which one is appropriate for the KPI that you wish to create. For information about creating and editing event maps, refer to Mapping Events.

When creating a KPI, you must define Empty Data Set behavior (EDSB) for that KPI. EDSB behavior defines how the system handles intervals where no data is collected, and EDSB settings can have implications that ripple throughout the system. For an individual KPI with a calculation type of Sum or Count, the Empty Data Set Behavior can be defined as either “No Reading Provided” or “Value is Set To Zero.” For all other calculation types (Average, Last Value, and State), the EDSB is set to “No Reading Provided” and cannot be edited.

After the first KPI is defined on a given measure, all subsequent individual and composite KPIs on that measure must have the same EDSB setting. The EDSB setting for subsequently defined KPIs is defaulted to the setting specified for the first KPI defined on the measure and is not editable.

EDSB Setting	Description
No reading provided	Applies only to Sum KPI instances. If applicable, when calculations are performed, zero readings are counted and factored into calculations and are displayed on graphs and tables. Null readings are not factored into calculations and are not displayed on graphs and tables.
Value is set to Zero	Returns zero when zero or null readings are provided. The zero/null readings are factored into calculations, and they are displayed on graphs and tables; so they can potentially have a significant effect on values displayed throughout the system.

Also when you create an individual KPI, you can construct a naming template to use for that KPI using the Naming Template field. A naming template enables you to customize the name displayed for each KPI instance. By default, a KPI instance name is composed of the base KPI name plus the dimension hierarchy description. This naming structure is unique but can result in a lengthy display name that is difficult to interpret. Constructing an appropriate naming template enables you to customize the display name in a manner that suits your needs.

The event map and hierarchy you select for the KPI you are creating determine the variables available in the Naming Template field. Variables are constructed using the following format:

The naming template can contain any descriptive text in combination with one or more variables you select. When creating a naming template, you should select variables that make the derived name distinctive for each KPI instance. When working with composite KPIs, the variables available for use in the Naming Template field are those that are common across both KPIs specified.