Why is Business Intelligence so expensive?

When the resources required to implement and manage a software solution are several times the cost of the software itself, the solution is expensive. When the time it takes to implement the solution is measured in months and some times, in years, the solution should certainly be considered expensive.
Having worked in the business intelligence market for over 20 years, it has often surprised me at how much time and effort it takes to implement business intelligence solutions. Industry analysts have estimated that the resource and services costs are typically multiples of the software license cost. In his blog of July 10, 2013, Boris Evelson, analyst at Forrester Research submits that 20% of the costs for business intelligence implementations are related to software and hardware costs versus 80% for resources and services.
Open source and cloud based solutions have contributed to reducing the costs of licensing BI solutions, either on premise or as a service. However, many of these solutions still rely on traditional approaches to implementation involving “moving data” via ETL, defining, creating and managing complex star schemas and the deployment of multiple tools to handle a variety of business intelligence requirements. Basic reporting, multi-dimensional analysis, and dashboard creation often require separate tools, with little opportunity for reusing work from tool to tool. With software costs decreasing, one might believe that the service to software multiples will only grow larger.
Gartner estimates the overall business intelligence market (based on software revenues) will reach $13.8B in 2013 and grow to over $17B in 2016. This is a huge market representing a significant opportunity for innovators that can drive down the services to software cost multiples.