I ran across this article on how accounting can kill innovation, basically talking about the issues of applying our standard accounting rules and ROI calculations to innovative ideas. When an idea is too new, or too different, applying estimates of returns and value are often wildly incorrect.
A study showed that when strict financial controls were not applied at the beginning, but rather decisions were made along the way to continue investing in projects, there were more likely to be positive returns from the project.
This approach seems to be one that could be useful in many BI type projects, and to a large extent, seems to be a good use of Powerpivot as a way to prototype projects. Assemble some data by hand, making a limited investment, and then determine if further investment is warranted.
Companies often follow similar paths, with similar software and the differentiators between can turn out to be the innovation of the people inside the company. Making a case for innovation, with limited investment and small prototypes can be a great way to work on something interesting while potentially making yourself a very valuable employee to keep around.