In motion-control applications, the parameters of interest would generally be position and velocity.
Control Systems Science is an entire engineering discipline, so the ramifications of feedback control can go VERY deep. Stability (or the lack thereof) is the most vexing problem. To attempt to put it in a nutshell, by the time you observe an error (as between commanded position and actual position) it is already too late to correct it, as an error has already occured! But, you want to correct it without overshooting, and causing a possibly bigger error of the opposite sign. So, you have to moderate the reaction correcting the error so that the error is reduced moderately, rather than wildly rushing to (and then past) the zero-error point. This is called 'loop compensation', and generally involves reducing the gain and/or frequency response of the feedback loop. Reducing the frequency response can be thought of as reducing the gain at higher frequencies only, leaving the DC gain high, so as to minimize the error when stationary.
Some of the servo motor types are DC brush, DC Brushless, and AC Induction. Any of these designs can be either common rotary motors, torquer motors that are limited in how far they can rotate, and linear motors.
Here is a description of how to calculate the performance of a servo motor system using a ball screw to move a machine table. Motor Performance Calculation
A glossary of related terms -