Alex,

Thanks for the quick (and useful) reply. Although it doesn't exactly solve the issue I'm facing it gives me an idea for another problem I'm working through.

My original example was not completely clear so I've created a new one. Because I don't know how many products are in a set or how they will be indexed I won't be able to use the lookup() and first() logic you pointed out. I will know what my "anchor" products are because they will all be marked. (In my new example this is column C.) If I'm able to potentially propagate the anchor volume out into a calculated field (in my example: column E) then they final calc is really easy. And this allow the final % calculation to work dynamically at different levels of aggregation which is one of my goals.

Let me know what you think. Thanks.

Great stuff, thanks Alex. However, to answer your first question: no, there may be more than one anchor SKU. As such for each roll-up (in this case, geography) it would be the SUM of volume for the anchor SKUs. I created a new example Excel file which includes a Northeast geography which shows what I mean.

Sorry for the confusion. In my effort to make the example relatively easy to understand I've not included all the relevant details.

That ... is ... frickin' ... awesome. Thanks Alex!! Very much appreciated.