According to Search Business Analytics, “agile BI” is one of the most discussed concepts. As projects using traditional BI development approaches take too long to finish with high cost and failure rates, people are discussing how to apply agile development principles to BI projects.
I agree that, on the one hand, using an agile approach will improve the success of BI projects, and will also enable organizations to start more projects. On the other hand, I also believe that traditional BI solutions lack capabilities needed to achieve agile BI.
Here are some of the reasons:
- Traditional BI solutions comprise multiple disparate tools. Traditional BI solutions do not provide integrated data design, modeling, and visualization solutions. Because of the nature and the number of tool sets needed, each project step must be done sequentially, from data design to modeling to visualization. This increases the number of handoffs between resources, resulting in communication and vision gaps.
- Users have limited input into creation of the tool they need. Only when all steps have been completed can the project results be viewed by the users to determine whether the project actually meets their needs. That is why in these BI environments, the users are called “end users.” They are the end of the process, passively consuming the software and the information provided to them.
- Data prepration takes significant time and effort. With traditional BI solutions, the majority of development time is spent preparing the data: cleaning and consolidating it from different data sources and formatting it for the reporting and analysis tool. These solutions do not provide checkpoints for gathering user feedback during this phase. This increases risk and likelihood of failure.
- Traditional BI solutions cannot quickly respond to change. Business dynamics change constantly, and business users need to adjust their analytics accordingly. With traditional BI solutions, business users highly rely on IT department to make changes. In these environments, the ratio of business users to IT staff determines the backlog and time it takes to make needed changes.
Agile BI development is not something new for QlikView customers. QlikView has long supported rapid application development by delivering all the needed capabilities for BI development in one comprehensive product. This is part of the reason why Gartner describes the benefits of QlikView to include short time to value and low cost of implementation*. In addition, QlikView does not require any specific data formatting or pre-staging of data before visualization and analytics development can begin. Developers can load the data from different data sources very fast and can have daily/hourly checkpoints with the business users to validate the data during the development.
In QlikView projects, users are actively engaged with the development, even during the data preparation stage. That is why with QlikView we don’t have the concept of “end users.” Business users engage with the analytics development at each stage. Once the data is prepared by the developers, with QlikView’s intuitive and interactive user interface capabilities business users can explore the data, turn data into knowledge, and discover new insights. This enables organizations and IT departments to adopt a self-service approach. With QlikView, business users own future enhancements and changes to the analytics.
The usability and productivity challenges of traditional BI solutions make it very hard to take an agile approach to BI development. Hence, this blog post is named “Traditional BI in Babushka Doll”. A babushka doll, sometimes called Matryoshka doll, is a set of wooden dolls of decreasing size placed one inside the other. To be able to play with the smallest doll, you need to take off the bigger dolls, just like traditional BI solutions require many tools and resources for development and maintenance. Watch this video and you will better understand what I mean.
*Gartner report, “Magic Quadrant for Business Intelligence Platforms.” January 27, 2011 http://www.qlikview.com/us/explore/resources/analyst-reports/gartner-report-magic-quadrant-for-business-intelligence-platforms-2011