"But maybe sales want to cross reference order arrival time with time to ship out, customer geographic region and even products ordered. The goal: see if there is a correlation between the length of time it takes certain products to make it through manufacturing and customer retention in specific geographic regions."
I used to generate this sort of thing out of the normal working RDBMS quite regularly. It wasn't a big deal if the right indexes were in place. If they weren't it might have take a little time to work out a plan with temp tables. Scouring stuff out of middleware wasn't a necessity (or, indeed a possibility) because I was working with systems that handled the entire selling process. What's described seems to be the consequence of buying a load of systems that each do part of the job and cobbling them together.
And I wonder if manglements take the slightest notice of the sort of thing in the example. They don't seem to notice even less subtle distribution issues such as the package allegedly sent from one carrier depot to another that's never seen again by the system and the absence never noticed until the customer complains. But I suppose it gives all the MBAs something to compose PowerPoints about.