Control: Holding the Gain
Control makes the improvement permanent. The candidate must know the control plan - the document naming each metric, its specification, measurement method, sample size and frequency, the responsible person, and the reaction plan when it goes out of control. Without a reaction plan the control plan is incomplete.
Statistical process control (SPC) is central. Understand variable charts (X-bar and R, or X-bar and S) and attribute charts (p, np, c, u), and know the difference: p and np chart defective units, c and u chart defects per unit. Recognise special-cause signals using the Western Electric / Nelson rules - a point beyond three sigma, a run of points on one side of the centreline, trends, and repeated patterns. The key trap: reacting to common-cause variation as if it were special cause is tampering, which increases variation.
Also required: standard work and updated documentation, training and handover to the process owner, visual management and andon, mistake-proofing to prevent recurrence, and auditing to confirm compliance. Control charts versus run charts: run charts lack control limits and cannot distinguish common from special cause with the same rigour. Pre-control and reaction plans appear in questions about operator response.
Worked example / rule of thumb: after reducing order errors, the team plots weekly error rates with control limits and adds a rule: any point above the upper limit triggers a documented review within 24 hours, owned by the shift supervisor. That combination - chart, limit, owner, reaction - is what an examiner looks for as a complete control strategy.