Before we start, let me quickly summarise some of the problems with traditional BI that we are trying to resolve. Traditional BI is characterised by the need to process and store data in intermediary data stores (e.g. date warehouses or data marts) in a form that is suitable for analysis (e.g. on-line analytical processing (OLAP) cubes). The consolidation and transformation of all this source data into a single repository involves detailed requirements analysis, specialised skills, additional infrastructure costs, and long development lead times, which often results in a system that is slow, inflexible, expensive, and too IT-centric.
Self-service BI aims to empower business users to be able to access relevant corporate data, perform their own analysis, and collaborate with others, whenever they need results, and without being dependent on IT or requiring specialised skills.
To introduce some balance, I should also mention that I am a fan of data warehouses and that they serve a number of valuable functions that are not available in self-service BI. They can add metadata to conform different data sources in the absence of a master data management system; they protect the source operational databases from direct access that could compromise security or performance, they provide higher levels of confidence in the completeness and accuracy of the data; and they can provide better performance when analysing large volumes of data.
The fact is that self-service BI and traditional BI serve different needs and have their strengths and weaknesses. You can achieve the best of both worlds by recognising when one is more appropriate than the other and by building a self-service BI strategy on top of your existing BI platform rather than trying to create parallel systems.
In this series I will explore the purpose and benefits of self-service BI in more detail with the following topics: