Control Plans

Control Plans (CSSGB BoK VI.B.1 — Apply)

Quick Answer: A control plan is a living document that states how a process is monitored and controlled after improvements. It lists process steps, product/process characteristics, specifications, measurement methods, sample size and frequency, control methods, reaction plans, and owners. Green Belts apply control plans in DMAIC Control so gains survive handover to operations.

The Control phase fails when the team celebrates a pilot and walks away. Without a structured control system, operators revert to old habits, settings drift, and defect rates climb back toward baseline. VI.B.1 is at the Apply level: you must build and use a control plan, not merely define the term.

Purpose of a Control Plan

A control plan documents the system used to keep a process on target and within requirements. It answers:

• What must stay controlled (characteristics and process steps)?
• How do we know it is controlled (measurement method, sample size, frequency)?
• What method keeps it controlled (SPC, poka-yoke, checklist, visual standard)?
• Who acts when something goes wrong, and what do they do (reaction plan)?

Control plans bridge project work and day-to-day management. They are common in automotive and regulated industries (often aligned with AIAG-style process control documentation) and equally valuable in healthcare, services, and transactional processes when CTQs can drift.

When Green Belts Build the Control Plan

Typical timing:

1. Late Improve / early Control — draft after the pilot proves the solution works.
2. Control phase — finalize measurement methods, charts, reaction plans, and owners.
3. Project closure — hand the controlled process to the process owner with training and document control in place.
4. Ongoing — revise when process, equipment, specs, risk, or customer requirements change.

The control plan is not a one-page certificate of completion. It is the operating contract for sustaining the improved process.

Core Elements of a Control Plan

Exam questions often test whether you can name and apply the standard fields. A practical control-plan row or block includes:

Some formats also list machine/device, classification (critical/major/minor), and related FMEA references. For CSSGB, prioritize characteristics, measurement, control method, reaction plan, and ownership.

Product vs Process Characteristics

Product characteristics are features of the output (dimension, purity, on-time completion, error-free form). Process characteristics are settings or conditions that drive the output (temperature, cycle time, torque, hold time, queue age).

Strong control plans often monitor leading process parameters in addition to lagging product results. Waiting only for final inspection finds problems late. Controlling the parameter that creates the defect catches drift earlier and supports prevention.

Control Methods Green Belts Choose

Match method to risk and data type:

• Statistical Process Control (SPC): variable or attribute charts when data are available and special-cause signals matter.
• Poka-yoke (error-proofing): physical or system design that makes the error impossible or immediately obvious.
• Standard work / checklists: sequenced steps for consistent execution of critical tasks.
• Visual standards: photos, shadow boards, limit samples that make abnormal conditions obvious.
• Layered process audits or layered checks: short, frequent verification that standards are followed.
• First-piece / setup verification: confirm setup before a full run ships risk downstream.

The control plan names the method; related documents (chart worksheets, SOPs, work instructions) hold the how-to detail.

Reaction Plans

A reaction plan states what happens when a control limit is breached, a checklist fails, a poka-yoke trips, or a product is out of specification. Effective reaction plans include:

1. Containment — stop the process, quarantine suspect product, protect the customer.
2. Immediate correction — restore the process to a known good state if the fix is defined.
3. Escalation — notify the process owner, quality, or engineering when criteria require it.
4. Documentation — record the signal, action, and disposition for learning and audits.
5. Root-cause follow-up — when special causes recur, trigger investigation beyond the band-aid.

Vague reaction language such as “notify supervisor” without stop/quarantine rules is a common exam and real-world weakness.

Sampling and Frequency Decisions

Sampling and frequency should reflect risk (severity of failure), process stability, and cost of measurement. High-severity CTQs may need 100% inspection or continuous monitoring until stability and capability are proven. Stable, capable processes with strong poka-yoke may use reduced sampling. Frequency should catch drift soon enough to prevent large nonconforming lots—not only once per month for a high-volume critical characteristic.

Building and Implementing the Plan (Apply Steps)

1. Identify critical characteristics from CTQs, FMEA high RPNs, and capability risks.
2. Confirm measurement systems are adequate (MSA where relevant) so control data are trustworthy.
3. Select control methods appropriate to data type and failure mode.
4. Write reaction plans with containment, correction, and escalation.
5. Assign owners for monitoring and for reaction authority.
6. Train operators and supervisors on standards, charts, and reaction steps.
7. Pilot the control system alongside the improved process; adjust frequency and methods.
8. Place the plan under document control and schedule periodic review.
9. Transfer ownership to the process owner at project close with metrics still visible.

Control Plan vs Related Documents

• Control plan — what is controlled, how, how often, and what to do when out of control.
• Work instruction / SOP — detailed steps to perform the job correctly.
• FMEA — risk analysis that helps prioritize which characteristics need strong control.
• Process map / flowchart — visual of steps; the control plan attaches controls to those steps.
• Project charter / storyboard — project history; the control plan is the sustained operating system.

Do not confuse a control plan with a project plan. A Gantt chart schedules project work; a control plan sustains process performance after the project ends.

Exam Focus

Expect scenario questions: which fields belong on a control plan, what belongs in a reaction plan, whether a lagging-only inspection strategy is enough, and when to update the plan after a process change. Apply-level success means choosing a coherent monitoring and reaction system for a described CTQ—not memorizing a logo or form number alone.

Key takeaway: The control plan turns improved performance into a managed process with clear characteristics, measurement, control methods, reaction plans, and owners so DMAIC gains last.