A simple but inaccurate model as described in the previous post would suffice from a “behind the scenes” perspective. There’s no need to throw complex statistical manipulations at a PMO learning how to use the system. I’ll point out that from a training and education standpoint, I am all in favor of keeping the model simple. I must admit that working through these calculations did at times make me wish I’d spent less time translating classical Chinese poetry in college, and perhaps picked up a calculator once or twice.

The Excel workbook I used to develop this is available here, and continue reading for more information on how I did it. The good news is that I was successfully able to reproduce the server side driver calculations. This left me unsatisfied, and so I decided to play with the model a bit more to see if I could reasonably emulate the Project Server calculations. That model worked well with simple driver prioritization scenarios, but seemed to display a wide variance with Project Server calculations in more complex scenarios.

In my previous post, I gave an example of how Pairwise Analysis works, and presented a conceptual model in Excel of the driver and project prioritization engine within Microsoft Project Server 2010.

Matrix.xla excel 2016 download