Continually adjusting a process that is already in a state of statistical control can be counterproductive. This is sometimes referred to as "tampering" with the process.
Deming's funnel experiment is one way to demonstrate the effects of unnecessary adjustments on the variation of a process. See the book "Out of the Crisis" by W. Edwards Deming.
An example from the University of Alberta.
Explanations from the Deming Institute.
Some theory (albeit hard to read).
The attached file [Funnel Experiment (supporting math).pdf] shows the mathematics behind this phenomenon. (Thanks to Dr. A)
Graphical Examples
The attached file "process_tampering.R.txt" has R code for a graphical/visual example of process tampering and how it inflates the process variation.
Here is a site that offers a package that expands on the funnel experiment with four rules and two dimensions.
There is a very nice demo version at this link.