I had an interesting question the other week from a customer who was trying to get some particular analysis of data from their PCM model. Their project had already done a good job of analysing their cost base, just like many other project teams in lots of different industries, and they were getting some useful information about their (in this particular case) products and channels. Like all good PCM models, their PCM model was giving them interesting breakdowns of activity costs and unit rates, and they could do some comparisons across the business to determine best practice.
But, the customer wanted to know, how could they see which cost drivers were most significant to the business? Having done the analysis of the activities that their business carried out, and identified the drivers that each activity used, an interesting analysis, they thought, would be how the costs flowed through each driver. With that sort of analysis, the organisation would be able to take a view on which cost drivers were the most significant. Then, they would be able to pick the three or four most significant drivers. Is most of the company’s cost base allocated using the "sales volume" driver? Or does more go through "number of deliveries", or "enquiries"?
There isn’t a cost object cost table in PCM that provides this analysis, so we set about building a proof-of-concept of how we might develop this sort of information. First of all we pulled the SummaryActivityValue table out of the model and loaded it into an Access database, and then we captured the CostObjectDriver table into the same database. From this, we were able to join the two tables and fairly easily do a set of summary reports and charts.
A caveat or two: this is a prototype, and needs some work to make it ready for production. It should also be noted that the customer’s model is particularly straightforward, and doesn’t include overrides in the cost object assignments. Furthermore, all its activities were driven to products and channels on the same driver; there were no two-stage assignments. This made the analysis easier, and avoided any awkward questions about which driver gets the cost when an activity goes to products on a (one dimensional) driver and then to channels on another (two-dimensional) driver. My opinion is that this kind of assignment is usually included in a model when the driver that you really want to use is only one-dimensional, but you want it to go to two dimensions. In this case the second driver is actually a proxy for the first. So the first, one-dimensional driver is the one that is considered the important one in the cost object assignment.
Without much further effort, we managed to generated some nice pie charts that very clearly indicated the drivers through which the majority of the costs flowed. We also determined that there were a couple of cost drivers that took a very small proportion (less than 2% each) of the overall cost base. One of these cost drivers, it should be noted, was a particular pain to collect and verify; the project team now has put that particular cost driver “on notice”, and it is possible that it will be replaced in the model in a future version. This proposed change will not materially affect the result of the calculation, but it will enable the project team to complete its monthly reporting cycle more quickly - an outcome that we hadn't considered at the start of the exercise.