One of the things not mentioned in this thread but probably mentioned in the literature is the proper design of the procedure or action.
In the Demming method, the role of proper analysis of procedure and or "systems analysis" is critical. There are many simple procedures that do not require in depth analysis. There are however many that do.
The design and use of FMC's is a good example. A modern FMC is designed to ease the crew work load as far as Navigation is concerned. Limitations on the design of the FMC and the limited knowledge of engineers that built the system contribute to possible errors. Case in point is the Columbia crash of the AA 757 several years ago. There was confusion as to what fix they were navigating to.
In the US currently the FAA is attempting to gain additional airspace utilization by designing RNP RNAV departures. These departures are designed to allow simultaneous departures on close parallel runways with very accurate NAV separation. Depending on the FMC in use, there are built in hazards to this procedure. My company's old FMC's in our Boeings require extensive "work around" and programming to comply with these new RNAV departures.
Our Airbus fleet requires almost no modification. However despite this ease of use, last minute runway changes, a very common occurrence in LAX and ATL, can cause critical separation standards to be compromised.
Here was a problem that didn't exist until the new minimum separation departures came along. Now there are critical programming procedures that must be complied with or separation standards are not met. The earlier FMC's were not designed with this in mind and the programming procedure is
very error prone.
If a procedure is cumbersome or does not fit the flow of the work being done, or if the procedure actually hampers normal operations it is likely to be disregarded. The FAA needs to understand that the gain in utilization provided with RNP departure procedures needs to be tempered with a need by crews for no last minute changes requiring extensive reprogramming.
Demming learned early on that in the analysis of work systems those "work flow procedures" must support the job at hand and not detract from it.
The FMC RNP procedure is just one small example.