A successful Kaizen event can produce an impressive result by Friday afternoon. A team removes unnecessary steps, redesigns a handoff, updates the workstation, and watches a key metric improve.
Then Monday arrives.
The workload changes. A trained team member is absent. A supervisor is pulled into another meeting. The visual board is not updated, the new standard is interpreted differently across shifts, and the process gradually returns to its previous pattern.
This is why improvement is not complete when the Kaizen event ends. The result must be converted into standardized work, visible measures, clear ownership, and a defined reaction plan. In the language of DMAIC, this is the purpose of the Control Phase.
A strong control plan turns a temporary improvement into a repeatable operating method. Over time, repetition and feedback create the operational equivalent of muscle memory.
Why Kaizen Gains Need a Control System
Kaizen is a practical approach to continuous improvement. It encourages teams to observe the work, identify waste, test changes, and improve the process through focused action.
However, a Kaizen event usually operates under unusually favourable conditions:
- The right people are gathered at the same time.
- The problem has leadership attention.
- Team members are actively observing the process.
- Decisions are made quickly.
- The improvement team is present to coach the new method.
Normal operations do not provide these conditions automatically. The control system must therefore carry the improvement forward after the event.
In the realm of Lean Six Sigma, the Control Phase protects the gains achieved during Improve. It confirms that:
- The new process is documented.
- The process owner accepts responsibility.
- Team members are trained and capable.
- Performance is measured at the appropriate frequency.
- Signals trigger a proportionate response.
- Benefits are reviewed after the project team withdraws.
This connects directly to standardized work: the agreed best-known method for completing a task safely, consistently, and efficiently at a defined point in time.
What a Control Plan Must Contain
A control plan should be detailed enough to guide action but simple enough for the people doing the work to use every day. At a minimum, include the following fields:
| Control plan element | Practical question |
|---|---|
| Critical-to-quality characteristic | What outcome must remain stable? |
| Measure | How will performance be calculated? |
| Target or specification | What result is expected? |
| Data source | Where will the information come from? |
| Frequency | How often will it be checked? |
| Owner | Who reviews and acts on the result? |
| Control method | What keeps the process within the desired condition? |
| Reaction plan | What happens when performance changes? |
| Escalation route | When and to whom is the issue raised? |
For example, a warehouse Kaizen may reduce picking errors by changing the layout, introducing barcode verification, and creating a new standardized work sequence. The control plan should not merely state “monitor picking accuracy.”
It should specify:
- Picking accuracy target: at least 99.5%
- Sample: all orders processed
- Review: daily by the shift leader
- Visual signal: green at 99.5% or higher; amber from 99.0% to 99.49%; red below 99.0%
- Reaction: verify scanner availability, review the last 20 errors, observe the standard work sequence, and escalate repeated causes to the process owner
That level of precision removes uncertainty when the process moves.
What Is a P Chart? Using the Voice of Process
One of the most useful questions in process monitoring is: what is a P chart?
A P chart is a statistical process control chart used to monitor the proportion of defective units in a sample over time. It is appropriate when the outcome is attribute data, such as:
- Pass or fail
- Correct or incorrect
- Complete or incomplete
- Complaint or no complaint
- Error or error-free
The chart normally displays:
- The proportion defective for each sample
- A centre line representing the average proportion
- An upper control limit
- A lower control limit, where applicable
The control limits describe the expected range of common-cause variation. A point outside the limits, or a non-random pattern within them, may indicate a special cause that requires investigation.
A P chart is not the same as a specification chart. A customer requirement might state that defects must remain below 1%. A control limit, by contrast, is calculated from process data and helps determine whether the process behaviour has changed.
This makes the P chart a valuable expression of the Voice of Process.
The Voice of Customer tells you what the customer needs. The Voice of Business identifies organisational priorities such as cost, capacity, or compliance. The Voice of Process shows what the process is actually doing over time.
When the Voice of Process is visible, teams can distinguish between:
- A normal fluctuation within the established process pattern
- A statistically unusual signal
- A sustained shift that requires a deeper improvement project
A P chart should only be used when the data definition is consistent. “Defective” must mean the same thing on Monday, Friday, and across every shift. That is why operational definitions and standardized work are essential companions to statistical process control.
Worked Example: When a Kaizen Gain Was Almost Lost
Consider a customer service team processing account-change requests.
Before the Kaizen event, the team reviewed 1,000 requests per week and found an average of 80 errors, producing an error rate of 8.0%.
The Kaizen team identified three major contributors:
- Re-keying customer information between systems
- Inconsistent document verification
- Unclear escalation criteria
The team introduced:
- A single-screen data-entry form
- A verification checklist at the point of work
- Standardized escalation rules
- A five-minute start-of-shift review
- A daily error board
During the four-week pilot, the results were:
| Period | Requests | Errors | Error rate |
|---|---|---|---|
| Baseline average | 1,000 per week | 80 | 8.0% |
| Pilot week 1 | 1,020 | 25 | 2.45% |
| Pilot week 2 | 1,015 | 18 | 1.77% |
| Pilot week 3 | 1,030 | 15 | 1.46% |
| Pilot week 4 | 1,025 | 14 | 1.37% |
The team had reduced the error rate from 8.0% to 1.37%, an improvement of approximately 82.9%.
The initial control plan required:
- Daily review of the P chart
- Weekly standard work observation
- A target error rate below 1.5%
- Process owner review of any sample above the upper control limit
- Monthly benefits reporting
For the first three weeks after handover, performance remained strong:
| Week after handover | Requests | Errors | Error rate |
|---|---|---|---|
| Week 5 | 1,010 | 13 | 1.29% |
| Week 6 | 1,040 | 14 | 1.35% |
| Week 7 | 1,025 | 16 | 1.56% |
In week 8, the daily P-chart review was missed because the team leader was reassigned temporarily. The checklist was still available, but it was not included in the new-starter induction. Two experienced employees also began using a previous escalation method because the old template remained accessible.
The result was 31 errors in 1,015 requests, or 3.05%.
The Kaizen gain had not disappeared because the improvement idea was invalid. It was weakened because the control system had an ownership gap, a training gap, and an obsolete document in circulation.
The corrective response was practical:
- Remove the old template from the shared drive
- Retrain all team members using a one-point lesson
- Restore daily P-chart review
- Add a backup owner for the metric
- Audit the standardized work sequence twice per week for one month
- Review the next 200 requests separately until performance stabilised
Within three weeks, the error rate returned to 1.41%, and the team continued monitoring it through the 30-, 60-, and 90-day sustainment reviews.
The lesson is important: a control plan must control the conditions that create the result, not only record the result after it changes.
How White and Yellow Belts Keep the Discipline Alive
Sustainment should not depend entirely on the original Black Belt. The broader Belt structure provides a practical way to distribute improvement capability.
White Belt contribution
A White Belt has foundational awareness of Lean Six Sigma principles and DMAIC. White Belts can help by:
- Recognising deviations from standardized work
- Understanding basic waste and variation concepts
- Raising issues through the agreed escalation route
- Participating constructively in daily improvement discussions
- Protecting the visibility of the control process
White Belts do not need to calculate advanced control limits to contribute. Their value is consistent awareness and timely escalation.
Yellow Belt contribution
A Yellow Belt has deeper knowledge of improvement tools and typically supports larger projects. Yellow Belts can:
- Maintain control charts and visual boards
- Conduct standard work observations
- Collect and validate data
- Lead small corrective actions
- Facilitate daily or weekly performance reviews
- Coach colleagues on the new process
- Identify whether a signal is local or systemic
The Black Belt may lead the original project, but Yellow and White Belts help embed the new method into daily operations. This creates shared ownership rather than dependence on one specialist.
Professionals seeking a structured pathway can explore Lean Six Sigma White Belt Training and Lean Six Sigma Yellow Belt Training. For project leaders, the Lean Six Sigma Black Belt course develops the advanced capability required to lead complex improvement work and mentor other Belts.
A Monday-Morning Sustainment Checklist
Before closing the Control Phase, confirm that the process owner can answer “yes” to these questions:
- Is the current standardized work available at the point of use?
- Has every affected team member been trained?
- Is the control plan visible and understandable?
- Is the data definition unambiguous?
- Is the P chart or other control chart reviewed at the right frequency?
- Does every trigger have a named response?
- Is there a backup owner for critical measures?
- Are obsolete forms, templates, and instructions removed?
- Are benefits tracked against the original business case?
- Are 30-, 60-, and 90-day reviews scheduled?
The Lean Six Sigma Practitioner Guide provides a broader framework for connecting DMAIC Control with Sustain, ownership transfer, leader standard work, audits, and management-system integration. You can also use the Lean Six Sigma Project Storyboard Toolkit to document the control plan, performance evidence, project closure, and sustainment reviews in one auditable record.
Make the Improvement the Default
Kaizen creates the change. Standardized work describes the change. The control plan protects the change. Daily management and Belt capability make the change repeatable.
That is how an improvement survives contact with Monday morning.
Build your capability through accredited, practical Lean Six Sigma certification and learn how to turn improvement activity into sustained business performance.
Kaizen. Kai-Care. Kai-Done. ( Lean Six Sigma)
