A bi implementation is a long journey, and in many cases, will be constantly evolving and changing. Here are some key steps to consider along the way:
Vendor or product selection: Obviously, you will need to have some way to deliver the data to the end users, and a software product is usually the best bet. Besides the business intelligence tool, you’ll also have to consider the data warehouse end of things as well. The selection should be based on data sources, design, and end user needs.
Data: What are the sources for the information and data to be used. What kinds of steps need to be taken from a capture, ETL (extract, transform, and load), and reporting perspective to make sure that the information will be in a form that can be reviewed easily by end users.
Architecture: How will the data, software, and tools sit together and be administered to be functional, efficient, and cost effective.
Design: What kind of questions and information do the end users need to have? What are the future BI consumers issues and information requirements. The system needs to be designed to scale and obviously this goes hand and hand with the data that is available.
The implementation itself can have different methodologies. A common method would be to build out the data structures to support the planned current and future required reports. While the most comprehensive, this can be a time consuming process, as data warehouses are not generally quick to build. Along the way, several other more flexible methodologies have emerged.
The agile method favors a quick to production approach. This might be on a departmental or functional level, or based on a subset of key reports. It’s more iterative in nature and can often resemble a quick release software implementation. However, it can lead to decentralization and data silos.
Copyright 2011-2018 BusinessIntelligenceBase.com