The Coke II Mistake: Why Data Beats Traditional Research