You would think that as much as I embrace the principles of direct marketing, that I would remember to create control groups in order to test my efforts. My time is so much more valuable than my money. So every effort I make must be carefully scrutinized for its value. If I don't split the sample I don't actually learn anything.
In marketing you ALWAYS have less than perfect correlation. So you KNOW that any change you see, however positive, is influenced by factors beyond the marketer's control. If you aren't splitting the sample you aren't really doing direct marketing at all.