The answer is that it allows you to check viability of the stock primarily, and thereby also have a population that you know is growing. Bulk cultures from liquid will always contain a proportion of non-viable organisms as there are no easy ways to separate these from the viable.
It also allows you to check to see that the preparation hasn't been accidentally contaminated (or tubes swapped etc.) by another morphologically different colony former - having the right type of organism is very important experimentally. If you wished to check the morphology of the organism by either morphological (i.e. microscope + stains) or biochemical tests, both of these need to start with a pure culture. It is hard to get enough bacteria for most of these tests unless you pick them from colonies - liquid culture just isn't concentrated enough, and can be heterogenous if a mixed inoculum is (accidentally or otherwise) added.
Another reason is that there is some heterogeneity in the culture anyway, no matter what you do to characterize it after growth. Isolating from single colonies, which should be derived from single colony forming units, allows you some control over this population for your bulk preparation, whether this be growth under certain conditions (e.g. antibiotic resistance) or growth rate.
It for instance may allow you to select a large colony, which might indicate a slightly higher growth rate for that founder. If you have something that is grown on selective media, then it may be that a portion of a bulk culture has lost the characteristic for growth on this medium, which means that when you inoculate from the bulk, what you put in might not be what you expected. This can lead to deviations in expected generation time calculations and similar physiological or phenotypic measures.