Our production consists of about 20 different test stations and we use two different kind of yield.
Total yield:
How many units passing all test stations without any remark, (1-Fail/Total).
Only looking at units ready for stock and whatever period you want.
All failed units are fixed and NOT scraped.

Station yield:
To analyze poor Total yield we look at yield per station to see where to focus for best impact.
Don’t actually know what to call these type of yield calculations but it works for us.

“Works for us” is one of the top justifications that I have heard for ineffective practices that I have heard.

If you think of SS in terms of “opportunities to make a mistake” and its iterations, you could run 10 parts through 3 tollgates and measure units into the tollgate vs. bad units, and thus each tollgate has a FPY. You could then sum the “opportunities” and the failures to then get system FPY- the 10 parts would generate 30 opportunities etc. But that might be a number that only impresses the suits- it isn’t tactical enough to be practical, unless you use it to see if improvements for one tollgate are hurting the KQC that you measure in another, but you could see that from that tollgate

I don’t think the key is being a SS company, but in being a Continuous Improvement company. SS is a tool for continuous improvement. When functional managers are themselves always improving and encourage their people to work cross-functionally on their priorities and also on other functions priorities, then you can pull out the bag of Lean/SS/TOC tricks as needed.