There is a single quote that identifies one of the biggest mistakes you see using Statistical Process Control (SPC) in semiconductor industry.
"You can't lose weight just by wearing tighter pants" - Kathy Naraghi
Variability in a step in a process flow can often impact yield by driving product out of spec (bad CpK). When that variability is random, the temptation is to tighten the control limits and respond more aggressively and try to keep the equipment running close to target. Here’s the problem, adjusting for random noise always increases the noise. What’s worse is when this happens in the middle of the night when a process technician is the one who has to respond and is expected to fix a systematic process/equipment deficiency.
As long as you don’t have to shut down the line, here’s a way for the owner of the process to proceed
Set up a disposition for out of spec material that the 24/7 sustaining team can follow
Determine the systematic sources of variability (data analysis or DOE)
Run equipment or process projects aiming at reducing the highest sources of variation
Convert your baseline process over to the successful projects as your change control allows
Recalculate and set new tighter control limits
This saves your technicians the stress of responding to problems they can’t solve, gets your efforts focused on the fastest solutions, and as a result makes the biggest impact on your bottom line