Agile decision making capabilities
IBM® Business Monitor supports agile decision making by providing historical analysis, future prediction, and dynamic alerting capabilities. The KPI history functionality captures the values of key performance indicators (KPIs) over time. These values are analyzed using the KPI History and Prediction widget and this information becomes the basis for predicting the expected future values of the KPI. Alerts can be created directly from the dashboard to notify users of a current business situation or of potential problems or opportunities in the future.
KPI history
KPI history is an optional feature for a KPI. You enable KPI history from the Monitor Model editor or KPI Manager dashboard widget. The KPI history feature captures and stores KPI data on an hourly basis. Using the KPI History and Prediction widget, business users can perform extensive analysis on the collected data directly from a dashboard.
The KPI History and Prediction widget provides a graphical time-based chart where users can specify a default time range and data frequency (such as one year of daily data points) that will be displayed each time a user navigates to the widget. Users can also interact with the chart on an on demand basis, for example, by changing time ranges and data frequencies, specifying start and end dates for arbitrary time ranges, displaying a tabular view of the data, clicking data points, or by hovering over certain points on the chart to display data values, range names and values, and targets.
An administrator-level user can manage the historical KPI data that has been stored and perform operations such as import and export functions.
KPI prediction
Business users can define and view KPI predictions directly from a dashboard. KPI predictions are made using an exponential smoothing algorithm that is based on the KPI's past performance, current trends, acceleration and deceleration, and cyclical patterns when applicable. Prediction models are defined with a future time span and frequency of prediction points. Prediction data is displayed on a graphical chart alongside of the history in the KPI History and Prediction widget after it has been configured. Similar to the history data, a user can interact with the chart, for example, to toggle the display of the prediction patterns (on or off), and so on.
Business-user alerts
A business user can define alert situations, alert content, alert delivery mechanisms, and alert recipients directly from a dashboard. One or more conditions can be defined in an alert created for a KPI. For example, an alert can be generated for a call center when both response time and call volume increase above target levels at the same time. When defining an alert, a user can specify the frequency with which to be alerted; for example whether to receive a single notification or repeated ones. The subject and body text of the alert is generated by default, based on the conditions that were selected. This text can be manually edited or a user can create a custom message. Alerts can be shared with other users, giving them the ability to subscribe to the same alert.
Predictions can be used in alert conditions, enabling users to create notifications that are sent when a KPI prediction indicates a potential problem or opportunity. For example, the recent KPI history and historical pattern for a cumulative quarterly sales KPI shows that the quarter might end below expectations. The prediction model can be configured to show the expected daily values through the end of the quarter, and alert notifications can be sent when the end-of-quarter value is expected to be above or below target. This enables users to make changes based upon the data presented.
The Alert Manager widget lists all personal and shared alerts. From here users can filter, edit, copy, or delete alert definitions.