What do good candidates for datapool substitution do?

A.
simulate unique user credentials and data entry choices by individual users

B.
simulate random timing and workflow variations by individual users

C.
improve application performance by requesting pages that are already cached

D.
make performance more predictable by using the same data values for every user