Process Management* is the practice of identifying, formalizing, evolving, and optimizing data processes, in the context of Business Intelligence.

Identifying Processes

The following list outlines the different types of processes that occur in a BI deployment.

1) Data Load Processes
2) Report Execution processes
3) QA and feedback processes
4) Development processes

Data Load Processes are required if data needs to be extracted in some way from the transactional systems for reporting purposes. The reasons for this are many. Multiple data sources, “dirty data”, transactional system is found to be too complex to report from directly, and to offload processing to another system or server are just a few examples.

Report Execution Processes are either pre-executions of operational reports, or cube builds. Examples are Transformer model builds to Powerplay cubes, or IWR (Impromptu Web Reports) scheduled, bursted, or public reports. These types of processes are required in order to present reports that take an unacceptable length of time to execute, or done to offload processing to outside of office hours.

QA and Feedback Processes are put in place to assure that the data presented is accurate, and to identify needed enhancements and/or corrections to the reporting deployment from the end user community. Two examples of feedback processes are Audit collection on report usage, and a Help Desk. Two examples of a QA process is a parsing of log files generated from a data load process, and Control Reports.

Development Processes detail a structured way that IT developers can deliver functionality to the end user community. This is successful only if there is a clear communications channel open between the user community (or a representative) and IT (management and staff). A well-structured development process minimizes development cost and maximizes user benefit.