I'm doing some purchase price variance exercises as we speak. To be clear, you want to understand cost variation month over month, and break it down to understand how much of that variation was driven by product mix, channel and geography? My first guess would be to have your average cost on a line graph at the top, with more detailed breakdowns underneath (Maybe lines graphs, or simple tables) for top products, channels and locations. With pricing constant, you should be able to explain changes from the average cost above by looking at the stuff below. If the 80/20 rules applies, your top 5 or so products, channels and locations will do most of the driving here. You also want to probably capture any changes in price...more so for YOY comparisons.
Joe, You have it basically right... so one of the major dynamics that has been tough to capture is that price paid for a service is not constant. Meaning even in a month where mix by Product, Channel, and geography are constant, we can still have a price variance... ed
I'm thinking breaking down your service pricing might be the first step to figure out if possible. It'll be tough to pin it down on something if this pricing patterns are too vague. Is there anything in the data model that can you give more detail behind service pricing? Hours (Any unit of measure), service type, customer type, rep, etc.