Consider that you have
a sales data cube that contains a level for the customer name, and
you have numerous data records for that level where the customer name
starts with the letter S. Each time a cube update takes place, a new
customer name is added whose name starts with the letter S. There
is a finite amount of reserved space for new last names when a cube
is first made. If all new names added sort in the same location of
the alphabet, rather than being evenly distributed throughout the
alphabet, then the odds of running out of reserved space increase
with each new cube update event.
Eventually, there is
no longer any reserved space left to fit in yet another last name
that starts with S. However, there might be ample room to add a name
that starts with the letter Q, X, or I. Nonetheless, the customer
name level will still have to be reorganized before new names starting
with the letter S can be added to the cube.