PDSA is a variant of Shewhart’s PDCA (Plan Do Check Act) cycle where Check is replaced with Study. It is used to structure a series of trials or experiments and learn from the results.
Let’s look closer at each phase:
- describe what is to be achieved, the goal, the outcome
- frame the improvement efforts and
- align efforts onto the goal
- prepare a set of “experiments” supposed to improve the situation
- prepare the “how” the collected data will be analyzed
- carry out the planned actions
- collect data
- analyse collected data
- compare results to predicted outcome
- capture the lessons learned (influencing vs. neutral factors)
- according to latest results and accumulated experience, what is the next cycle about, i.e. what are the next assumptions?
- start a new PDSA cycle with a new set of assumptions
- limit the ambition for each cycle and prefer several small incremental steps rather than quantum leaps. “Learning will be improved if fed in small bits”
- keep the experiences simple by limiting the number of variables
Smaller steps allow to adjust better and faster to small (environmental or system) parameter changes. Smaller steps and limited number of parameter i.e. simpler experimentation is easier to carry out with non-specialists like shop floor personnel.
During the Study (S) phase:
- look for consistency of actions carried out vs. those planned
- check if conditions, support, etc were as planned, no biases
- is the result an improvement?
- is the system as a whole improved or is it just a local improvement?
- can we conclude the original assumption(s) is/are valid?
PDSA compared to PDCA
With PDCA, the Check phase suggest an existing benchmark, reference or standard against which to assess the result of the Do phase. It is a compliance check. In PDSA the importance is building knowledge and improve by trial-and-learning. In PDSA, the failure to improve is as valuable than the improvement, as there is a lesson to be learnt behind it.
Any comments?Follow @HOHMANN_Chris