This is not all that simple. In well travelled areas, where data is complete and understood, data should be a priority.

But data is often incomplete, often wrong. Excessive reliance on data can also tend to keep one on the tried and true and missing the chance at true innovation. You can extrapolate on the past for only so long (RIM) but a really good excutive must be willing to take a risk and throw in a wild card. The famous Pontiac GTO (not the 200* version, you younger guys) was a classic example of breaking all the rules (including company directives) and it changed the automotive landscape. In fact marketing success has often been crippled by people who went by test group data ('new Coke' anyone?) not realizing that merely the fact of organizing a test group distorts the results.