Soda can make up nice real-world statistics. For example, do you suppose that taste tests for New Coke led them to make the change the formula (for those of us old enough to remember that event) but looking too close at that quantitative data caused them to overlook other qualitative data, like perhaps a negative reaction to an iconic brand that would tank sales? They were inferring something (future sales) from only the data they had. Is anyone perhaps familiar with the term GIGO?

According to the case study on the new coke I found, Coca-Cola spent $4 million (way back when) on market research and concluded from its research and blind taste tests that people preferred the new formula. Unfortunately, they did not do a study to understand the "emotional attachment" consumers had with the classic coke. After launching the new formula, people were outraged, and Coca-Cola responded by returning to the original formula.

In this example the company did follow the statistics illustrated from the marketing research and ultimately made a very serious error. We could measure taste on a quantitative scale (for example 1 = really dont like taste and 10 = really like taste) but the emotional attachment would be qualitative (not able to quantify).

Soda can make up nice real-world statistics. For example, do you suppose that taste tests for New Coke led them to make the change the formula (for those of us old enough to remember that event) but looking too close at that quantitative data caused them to overlook other qualitative data, like perhaps a negative reaction to an iconic brand that would tank sales? They were inferring something (future sales) from only the data they had. Is anyone perhaps familiar with the term GIGO?