The chance of it producing any result is incredibly low. Imagine trying to pin point the start of an outbreak when the US might have had a handful of cases each month, by randomly sampling people. You'd basically have to sample 100% of the population every month to catch a rare event like that.