When looking at data, the obvious first step is to ensure it is reliable. Once you determine it is reliable, there are certain conclusions that are made.
After making conclusions, where many managers go wrong is acting on those conclusions. They miss the critical third step - is the conclusion from causation or correlation? When you don't ask that question and come up with a solid answer, your next action becomes risky.
One of my favorite examples is a bit silly, but bear with me.
The data clearly shows that in Chicago, ice cream sales and gun crimes move in tandem. When ice cream sales go up, gun crime goes up. The data on this is reliable. However, managers with this information could make any number of wrong decisions.
This data is classic correlation. Which means one really has nothing to do with the other.
Suppose a hospital manager saw this data and decided that he would track ice cream sales and base his ER staffing based on that. It might be ok, but most likely it would lead to some bad staffing decisions.
FYI - ice cream sales and gun crime actually are correlated because they are both correlated to the weather. Warm weather increases ice cream sales and also increases the likelihood of gun crime.