Gathering Requirements for BI Implementation Project

What it is not?

It is not a report specification document - All those who despise reading a report with a lot of information, please raise your hands. Most reports designed today are crammed with information. What happens when a BI implementation project starts with a bunch of these reports, and it is successfully delivered. Three weeks later, a user needs a new report for a customer request or a change in the business. Are you going to have to design a new report? Your solution was not a BI solution - so do it right the first time. Don't focus on creating complex reports as part of the BI implementation.

It is not just a process for using the new BI tool that has been purchased - Too often a brand new BI tool is purchased by management and they want everyone to start using it and derive the ROI for the investment. A project is kicked off for beginning to use the BI tool. Few selected users are given training on the product. When BI Implementation project is kicked off, the business and functional users must determine the true requirements which should not be driven by the features provided by the BI tool.

It is not for estimating the cost and duration of the project - The requirements of many IT projects are used to determine the scope and therefore the cost. In a BI implementation scenario, there are many unknowns that will be determined during the course of implementation. These factors such as unknown data relationships and decision criteria will become apparent during the implementation stage and therefore the scope and cost should be more flexible compared to implementing an application.

What it is?

It is an exercise for defining mapping and relationships in existing data - Most times, the users are not sure how they want to use business intelligence, so they create a list of all the data elements for all the major aspects of the business such as products, manufacturing locations, distribution points and organizational units.

It helps understanding context in data for decision-making - The business users require contextual information for decision support to improve the management, customer and operational processes. For example, if the business users are concerned about time-to-market, then that context determines what data elements are required. The business strategies will help determine all the major context elements required. The BI Systems should end up supporting the business processes and the decision contexts.

Identifying automated decision making possibilities - Selecting the decisions to automate, for ex: if we consider pricing as a decision variable, we can automate it for achieving various business objectives such as increasing market share, increasing revenues and profits, stimulating demand, etc. However, deciding to raise or lower prices would be useless if all your customers are on fixed price contracts. The business users can decide which decision variables to automate so that the BI system can be built to contain the context and automation required to either send alerts or perform the automatic change within the ERP systems.

Confirm that the data integrity supports BI Implementation requirements - As a trivial example, if the state field is a free-form text field where users are typing it out, there could be several variations. Business Intelligence requirements cannot be complete without recognizing these challenges related to data integrity.

Best Practices

Make speed a priority in the requirements - In the latest BI Usage Survey, it was discovered that the top reason for users to reject their BI system was speed. Requirements should list out the speed at which information requested will be fulfilled by the BI system. The functionality of the BI tool should be capable of delivering at the required speeds and the cubes should be optimized for expected queries to run fast.

Security - Document the security requirements such as roles, permissions and information. A matrix of this will be exceptionally useful for BI implementation to be successful at the time of deployment. Information that falls into the wrong hands is worse than not having the information at all.

Make the establishment of a BI Competency Center a priority - Creating and staffing a BI competency center for giving access to the right people, determining the changes and training the users is essential. The key stakeholders in the BI implementation project should extend their support and establish the BI Competency Center during the early stages of the project lifecycle.

Summary:

Gathering Requirements for BI Implementation Project should not focus only on reporting specs but also on documenting data elements, understanding context in data for decision-making, identifying automated decision making possibilities, security requirements, performance requirements and establishment of a BI Competency Center.