Taking a "random" sample of your data for analysis that will drive your strategy is inherently risky. Why? Let's talk about it.