pdca-plan

Begin by defining the problem with the IS/IS NOT technique across what, where, when, and magnitude dimensions, using observable data rather than assumed causes. Then establish a numeric baseline, set a SMART objective, and formulate one testable hypothesis before moving to the Do phase.

The Plan phase supports 5 Whys for linear causal chains, Fishbone/Ishikawa diagrams for multi-category brainstorming, Pareto 80/20 for prioritizing vital causes, IS/IS NOT for problem delimitation, and 5W2H for full context. Step-by-step guides are included in the references.

Use DMAIC when the problem involves complex statistical variability requiring deeper analytical tools. PDCA Plan is better suited for iterative improvement cycles where a single hypothesis can be tested quickly with available data.

A valid SMART objective names one exact metric, includes numeric baseline and target values, is achievable by the team, connects to business impact, and has a deadline. Objectives combining two metrics should designate a primary goal and list others as constraints.

Testing multiple hypotheses simultaneously makes it impossible to determine which change caused the observed result during the Check phase. Each cycle should isolate one variable, with additional hypotheses queued for subsequent cycles.

The skill guides the planning process but does not replace domain expertise or judgment. If no historical data exists, the first Plan iteration may only define how to collect data, and statistically complex problems may require DMAIC instead.