IMO, this is an important consideration. Sure, the target level is SKU-store, but at what level are the purchase orders being made? The M5 Competition didn't say anything about this, but probably the SKU level is as important as the SKU-store, if not more.

For retail data in general, I think we need to see how well a method perfoms at different levels of the hierarchy. I've seen commercial and finance teams prefer a forecast that is more accurate at the top than another that is slightly more accurate at the bottom.