Take the case of your local instance of a national jeans store. It is probably staffed by 20-somethings or even teenagers, while far, far, away, a much older person makes decisions about the inventory in the store. Ms Store Clerk knows that the snap on those "boyfriend" style jeans has a little rough spot on the inside that is uncomfortable. She knows this she asks when the customer comes out of the dressing room, "How did those work out for you?" and the customer says, " What's the deal with the snap? It was totally scratching me!" And then Ms Store Clerk folds the jeans and puts them back on the shelf.
When Mr. Corporate sees the sales numbers for the store, he doesn't have access to that critical piece of information. But the human brain will fill in the blanks, if information is missing. Its the way we are hardwired. So Mr. Corporate will have a story for why the numbers are low on the "boyfriend" jeans. Perhaps he will think that the style is now out of fashion. So when he re-orders, he won't order more of those, but instread go with the capris. From the same manufacturer, who is still using the same snaps. And when the customers see that their favorites, the "boyfriend" style jeans are no longer stocked by that store, perhaps they will switch to the other national chain store in town. Just maybe.
In this case, the information costs are the costs of the lost sales, the lost future sales, and the purchases of the wrong stock. If the organization could somehow ask Ms Store Clerk about things, it would be able to make much better decisions.The take away for us is that decisions should be made as close to where the process is that they will affect as possible. It will be a negotiation to find the best place -- but I bet you will be surprised when you figure it out!