Gartner came out with a study recently stating that only 50% of BI projects actually succeed. You can argue on what denotes failure and I’ve heard discussions of failure rates ranging from 35-80%. Whatever the actual rate, the failures of BI projects is way too high in comparison to other IT implementations. I feel the success rate of BI projects is directly related to how involved the business users are involved in the process and the level of business knowledge of the IT implementers.
BI Systems are characterized by transforming a company’s Online Transaction Processing (OLTP) database system into a format that is more easily understood by decision makers. The transformed data is known as a Data Warehouse. This process involves transforming normalized individual transactions into a set of data characterized by measures, such as sales amount and quantity and dimensions, such as customer and product. The data in this format can then be pivoted against each other by decision makers to analyze the enterprises status, such as “let me see my company’s sales by customer by product by fiscal quarter”.
This can be a hugely complex and hardware/software intensive process. The main goal of the process is to take very voluminous and complicated data and turn it into very precise information that decision makers can quickly look at to see the enterprises status. This effect is exemplified through the use of Key Performance Indicators (KPI), which takes the data about current status and compares it to a goal to return a value of 1 (above goal), 0 (on goal) or -1 (below goal). The KPI is rendered through the use of an icon, the most well known one being the stop light with green for above goal and so on.
The users of this BI system are decision makers and come from the business side of the enterprise while the generators of the BI system are IT personnel. As in most IT projects, the business users define and approve the project deliverable and then IT goes off and implements. However, due to the complexity and scale of these data warehousing projects this traditional hand-off between the business users of IT and implementers is even more disconnected than on the usual IT project.
I have trained and mentored with many BI clients and it appears that the successful BI implementations have one main thing in common. They have someone who is well versed in business process and analysis involved in the planning, training, mentoring and review of the projects progress. This cuts down on the variance of the product that business wished for and what IT actually delivered. Due to the complexity, many specialists in BI technologies are missing this key component to success, which is being to apply business process and analysis understanding to the complete BI solution.
To make your BI project successful be sure to include in your team planning, training and implementation functions personnel who understand the business side of Business Intelligence. In other words, make sure you have the right amount of “Business” knowledge in your Business Intelligence project team.