Statistically speaking, the most valid procedure for collecting customer data is to randomly select a reasonably large representative group of customers and obtain complete and accurate data on each one. Well, think about that for a moment. a pharmaceutical company that is targeting a new drug for diabetic patients collect complete and accurate customer data of every diabetic patient in the United States. According to the 2014 statistics report, by Center for Disease Control and Prevention, there are 29.1 million diabetics in the US So, this procedure of collecting hundred percent customer data would not only be time consuming for the pharmaceutical company, but also would prove to be very costly. Since collection of feedback from hundred percent of our customers may not be possible, various other methods can be deployed. Each of the methods compromises a statistical approach in some way