This paper begins with the energy being put into a piece of rotating machinery and ends at the vibration graph displayed on the computer screen, taking into account each step this energy passes through from start to finish. We begin with inputs such as speed and load, the rotating shaft, the other sources of forcing frequencies, their journey from component to component, through the shaft, the bearings, the sensor mounting, the sensor, the electronics of the data collector, the filter and data collection setups etc. until they result in a piece of data that is ready for analysis. This paper will then posit that the data one analyzes at this last step may in fact be nearly meaningless if all the variables along the way are not controlled from test to test and the data trended. Ever wonder why sometimes the machine with really high vibration levels rarely fails but the one with low levels does? If you or someone you know has had bad luck with a vibration analysis program, this paper may explain why and describe an easier route to success.