Focused Improvement and Chronic Loss Elimination

Introduction

Total Productive Maintenance (TPM) treats equipment effectiveness as a shared organizational responsibility rather than a maintenance-department concern. Among the eight traditional TPM pillars, focused improvement—often called Kobetsu Kaizen—provides the disciplined method for removing losses that reduce productivity, quality, delivery reliability, safety, and cost performance. It converts operational problems into quantified improvement opportunities and mobilizes production, maintenance, quality, engineering, supply-chain, and management personnel in cross-functional teams.

Focused improvement is especially important for chronic losses. A chronic loss is a recurring, relatively stable problem that may appear individually small but accumulates over time. Examples include repeated minor stoppages, gradual speed reduction, recurring adjustment, small defects, waiting for materials, and frequent tool changes. Sporadic losses are sudden and significant events such as a major breakdown, a serious quality failure, or an extended setup error. The two categories require different responses: chronic losses demand systematic analysis of patterns and conditions, whereas sporadic losses demand rapid containment followed by permanent corrective action.

The objective of this chapter is to develop the capability to lead cross-functional kaizen using Pareto analysis, root-cause analysis, 5 Whys, Ishikawa diagrams, PM analysis, ECRS, SMED, mistake-proofing, and structured problem-solving. The emphasis is not merely on generating ideas, but on demonstrating causal validity, implementing controls, and sustaining gains.

Learning Objectives

After completing this chapter, the reader should be able to:

1. Distinguish chronic from sporadic losses and relate them to TPM performance measures.
2. Select and prioritize improvement themes using loss data and Pareto analysis.
3. Lead a cross-functional kaizen team through a structured problem-solving cycle.
4. Apply 5 Whys, Ishikawa diagrams, PM analysis, and verification methods appropriately.
5. Use ECRS, SMED, and poka-yoke to redesign work and equipment conditions.
6. Evaluate improvements using OEE, quality, cost, safety, and sustainability measures.
7. Recognize the limitations, governance requirements, and organizational conditions of focused improvement.

1. Losses as the Object of Improvement

TPM commonly examines six major equipment losses:

A broader production system also includes material shortages, information delays, labor imbalance, utility interruptions, excessive movement, and unsafe conditions. Therefore, an improvement team should not assume that every loss is caused by the machine itself.

Chronic and sporadic losses

A sporadic loss is usually visible because its magnitude is large and its occurrence is irregular. A chronic loss is often hidden by normalization: operators accept it as part of the process, maintenance repeatedly restores the same condition, and accounting systems classify its cost as ordinary production expense. Chronic loss elimination is consequently more difficult than emergency repair. It requires a change from event-based thinking—“What failed today?”—to mechanism-based thinking—“What operating conditions repeatedly create this outcome?”

A useful distinction is:

• Symptom: the observable deviation, such as a stoppage or defect.
• Immediate cause: the condition directly preceding the symptom, such as a blocked sensor.
• Physical root cause: the mechanism producing the condition, such as dust accumulation caused by inadequate sealing.
• Systemic cause: the management or design condition allowing recurrence, such as missing cleaning standards or weak design review.

Improvement is incomplete if it removes only the symptom. Replacing a fuse, resetting a sensor, or retraining an operator may restore production without eliminating the causal mechanism.

2. Measuring the Opportunity

Overall Equipment Effectiveness

OEE combines availability, performance, and quality:

where each factor is expressed as a decimal or percentage. Under a common loss-accounting convention:

Planned production time excludes scheduled periods in which production is not intended, such as planned shutdowns. Operating time is planned production time minus recorded downtime. Ideal cycle time is the theoretically shortest cycle time at which the process can produce without violating quality or equipment constraints. The definitions must be fixed locally because organizations sometimes classify planned stops, minor stops, or startup losses differently.

Worked example

A line is scheduled for 480 minutes. Planned breaks account for 30 minutes, so planned production time is 450 minutes. Downtime is 45 minutes, giving operating time of 405 minutes. The ideal cycle time is 0.5 minutes per unit, and the line produces 720 units, of which 684 are good.

Availability = 405 ÷ 450 = 0.900, or 90.0%.

Performance = (0.5 × 720) ÷ 405 = 360 ÷ 405 = 0.8889, or 88.89%.

Quality = 684 ÷ 720 = 0.950, or 95.0%.

Therefore:

OEE = 0.900 × 0.8889 × 0.950 = 0.760, approximately 76.0%.

The calculation is meaningful only when the data definitions are consistent. OEE is a diagnostic measure, not a universal productivity target. Raising OEE by producing excess inventory, increasing operator strain, or hiding planned losses is not genuine improvement.

Pareto analysis

Pareto analysis ranks loss categories from largest to smallest and displays their cumulative contribution. It is based on the practical observation that a limited number of causes often account for a large proportion of the measured effect; it is not a universal law that exactly 80% of results arise from 20% of causes.

The procedure is:

1. Define the loss measure: minutes, units, cost, incidents, or defects.
2. Establish the time period and population.
3. Group events using mutually understandable categories.
4. Rank categories in descending order.
5. Calculate each category’s share and cumulative share.
6. Select a theme using impact, feasibility, risk, and strategic importance.
7. Confirm the theme at the workplace before analysis.

For category i, its percentage contribution is:

The measure must not mix incompatible units. A Pareto based on minutes identifies time loss, whereas one based on financial cost may prioritize expensive but infrequent events. Multiple Pareto views are often useful.

3. The Cross-Functional Kaizen Architecture

A kaizen team should contain the knowledge needed to observe, explain, alter, and control the process. A typical team includes an operator, maintenance technician, process or manufacturing engineer, quality representative, production supervisor, and facilitator. Procurement, logistics, safety, automation, or suppliers join when the causal system requires them.

The sponsor defines the business significance and removes organizational barriers. The team leader controls the improvement method and evidence. The process owner approves operational changes. Front-line employees provide practical knowledge and must be involved in testing. A responsible owner is assigned for every permanent action.

A robust charter states:

• the process and boundary;
• the problem in measurable terms;
• the baseline period;
• the target and completion date;
• exclusions and assumptions;
• team members and decision rights;
• safety, quality, and regulatory constraints.

A target should be challenging but credible. “Improve the machine” is not a target. “Reduce minor-stop time on Press 4 from 38 to 15 minutes per shift without increasing defect rate” is testable.

4. Structured Problem-Solving Sequence

A practical TPM sequence integrates PDCA, the DMAIC logic of measurement and control, and shop-floor observation:

1. Select: choose a loss using evidence and strategic relevance.
2. Define: describe the gap, boundary, customer impact, and standard.
3. Observe: go to the actual workplace; collect time-stamped and condition-specific data.
4. Analyze: separate facts from assumptions and identify causal mechanisms.
5. Act: implement countermeasures in order of effectiveness and risk reduction.
6. Check: compare results with baseline and verify that unintended effects have not appeared.
7. Standardize: revise standards, training, maintenance requirements, and visual controls.
8. Horizontally deploy: examine whether the lesson applies to similar equipment or processes.

The sequence is iterative. If the result does not improve, the team should revisit the problem definition, measurement system, and causal hypothesis rather than simply adding more countermeasures.

5. Root-Cause Analysis Tools

5.1 The 5 Whys

The 5 Whys method repeatedly asks why an observed condition occurred. “Five” is a prompt, not a mandatory number. Some problems require three questions; others require seven or more branches.

Example:

• Why did the conveyor stop? The drive motor tripped.
• Why did the motor trip? Motor current exceeded its limit.
• Why did current rise? The conveyor resistance increased.
• Why did resistance increase? Product fragments accumulated at the return roller.
• Why did fragments accumulate? The scraper design allowed material to bypass it, and cleaning frequency was not defined.

The final statement identifies two distinct countermeasure paths: redesign the scraper and establish a condition-based cleaning standard. The method is weak when it becomes a blame exercise, when each answer is an unsupported guess, or when multiple causal branches are forced into one chain. Each “why” should be verified through observation, records, measurement, or controlled testing.

5.2 Ishikawa diagram

An Ishikawa, or fishbone, diagram organizes possible causes into categories. Manufacturing teams commonly use 6M categories:

• Manpower: skill, staffing, fatigue, communication;
• Machine: wear, alignment, design, controls;
• Method: sequence, standard, inspection, setup;
• Material: specification, variation, contamination;
• Measurement: instrument accuracy, sampling, definitions;
• Mother Nature/Environment: temperature, humidity, dust, vibration.

The diagram is a hypothesis-generation tool, not proof of causation. Teams should mark causes as observed, suspected, or verified. A cause becomes credible when its mechanism is physically plausible, its occurrence correlates with the effect, and changing the cause changes the effect while other conditions are controlled as far as practical.

5.3 PM analysis

PM analysis in TPM means analyzing physical phenomena and their mechanisms, not merely “preventive maintenance.” It asks four questions:

1. What physical phenomenon produces the defect or loss?
2. What conditions are necessary for that phenomenon?
3. What equipment or process factors create those conditions?
4. What standards, controls, or design changes prevent them?

For example, a filling machine may underfill containers. The phenomenon is reduced flow during the filling interval. Relevant conditions may include increased viscosity, restricted nozzle area, low supply pressure, or incorrect valve timing. The team then measures these conditions and determines which one explains the variation.

PM analysis is particularly valuable for chronic losses because it moves beyond labels such as “operator error” or “machine problem” toward a physical explanation. It is also useful for setting inspection points: a standard should monitor a condition that is causally connected to failure, not merely one that is convenient to record.

5.4 Verification methods

Root-cause verification may use:

• stratification by product, shift, machine, operator, material lot, or environmental condition;
• time-series plots and run charts;
• before-and-after comparisons;
• controlled trials or designed experiments;
• measurement-system analysis;
• fault reproduction under safe conditions;
• correlation supported by physical reasoning.

Correlation alone is insufficient. A night-shift association may reflect product mix, ambient temperature, staffing, or reporting practices. The team must investigate confounding factors.

6. Countermeasure Design

ECRS

ECRS is a practical method for redesigning work:

• Eliminate: remove an unnecessary step, motion, inspection, or transfer.
• Combine: perform two activities together or integrate functions.
• Rearrange: change sequence, location, or orientation.
• Simplify: reduce complexity, force, adjustment, variation, or decision burden.

ECRS should be applied to the actual process, preferably through direct observation and video or time study where permitted. It should not be used to eliminate safety checks or quality controls without risk assessment and authorization.

SMED

Single-Minute Exchange of Die (SMED) reduces changeover time, traditionally with the aspiration of achieving a changeover in fewer than ten minutes. Its central distinction is between:

• Internal setup: work that requires the machine to be stopped.
• External setup: work that can be performed while the machine is operating.

The method is:

1. Observe and document the complete changeover.
2. Separate internal and external tasks.
3. Convert internal tasks to external tasks where technically possible.
4. Streamline remaining internal work through parallel operations, quick clamping, standard settings, and visual confirmation.
5. Eliminate adjustment through presetting, locating devices, and poka-yoke.
6. Standardize and repeatedly audit the new sequence.

Changeover time should be defined consistently, for example from the last acceptable unit of Product A to the first acceptable unit of Product B at the required rate. If the endpoint excludes trial units or stabilization, the apparent improvement may be misleading.

Mistake-proofing (poka-yoke)

Poka-yoke prevents an error or makes it immediately detectable. Strong designs operate at the source:

• shape or dimension prevents incorrect insertion;
• interlock prevents operation in an unsafe or invalid state;
• sensor detects missing or incorrect components;
• sequence control prevents skipped steps;
• feedback immediately identifies deviation.

Detection after the product has advanced is weaker than prevention at the point of occurrence. However, automatic controls require validation, fail-safe design, cybersecurity consideration, and a clear response to sensor failure. A bypass key or override function must be governed, logged, and periodically reviewed.

7. Chronic Loss Elimination in Practice

A chronic-loss project should examine frequency, duration, conditions, and recurrence. A useful loss tree is:

Business impact → process loss → event type → physical phenomenon → causal condition → control weakness.

For example:

Lost output → minor stoppage → cap feeder jam → cap orientation instability → excessive vibration and guide clearance → no standard for guide condition.

This structure prevents teams from stopping at “cap feeder jam,” which is an event description rather than a causal explanation.

Chronic losses often require restoration before improvement. If basic conditions are abnormal—loose fasteners, contamination, poor lubrication, damaged guards, incorrect settings—advanced analysis may produce unstable conclusions. TPM’s autonomous and planned maintenance activities therefore support focused improvement by restoring equipment to an appropriate basic condition.

8. Case Study: Illustrative Packaging-Line Kaizen

Background

The following is a fictional illustrative case. A beverage plant operates a high-speed filling and capping line. OEE has declined from 78% to 70% over six months. The largest recorded loss is minor stoppage at the cap feeder.

Problem definition

A four-week Pareto of downtime shows 1,920 minutes of loss. Cap-feeder stoppages account for 640 minutes, or 33.3% of total recorded downtime. The team defines the project as reducing cap-feeder stoppage time by at least 50% within eight weeks without increasing defects or safety risk.

Analysis

The team stratifies the 640 minutes by product, shift, cap supplier lot, and feeder. Product Family B accounts for 61% of the cap-feeder loss although it represents only 40% of production. A workplace observation shows that the cap has a slightly wider dimensional tolerance and that vibration causes occasional cap bridging at the guide entrance.

An Ishikawa diagram identifies machine vibration, cap dimensional variation, guide clearance, cleaning condition, and operator adjustment. The team uses 5 Whys to examine bridging. PM analysis identifies the physical mechanism: vibration and excessive entry clearance permit two caps to overlap, creating a blockage. A measurement check confirms that the guide clearance varies between 4.5 and 6.0 mm after adjustment, while stable feeding requires 4.5–4.8 mm for the relevant cap family.

Solution

The team applies ECRS by eliminating repeated trial adjustments, rearranging the guide-setting sequence, and simplifying the adjustment through a calibrated spacer. A poka-yoke prevents the wrong spacer from being installed. The feeder is fitted with a vibration-damping mount, and the supplier is asked to improve cap dimensional control. Cleaning and guide-clearance checks are added to the autonomous-maintenance standard.

A short SMED exercise separates preparation of the correct guide and spacer from the stopped portion of changeover. The operator can now prepare the kit while the line is running.

Outcome

During the next four weeks, cap-feeder downtime falls from 640 to 250 minutes, a reduction of:

The result exceeds the 50% target. Quality data show no increase in cap defects, and the safety review confirms that the new guide and damping mount do not obstruct guarding or access. The team audits the standard for four additional weeks and deploys the guide-setting method to two similar lines.

Learning points

The improvement succeeded because the team quantified the opportunity, stratified the loss, observed the physical mechanism, verified a condition, combined equipment and method changes, and controlled the new condition. A simple instruction to “be careful with the feeder” would not have addressed the mechanism or created a reliable standard.

9. Governance, People, and Sustainability

Focused improvement can fail when technical changes are implemented without social or governance discipline. Operators may resist if kaizen is perceived as a disguised labor-reduction program. Participation, transparent objectives, and respect for local knowledge are therefore essential. Changes affecting work content, staffing, ergonomics, or skill requirements should be reviewed with relevant employee representatives.

Safety must take priority over production gains. Risk assessment should address guarding, lockout/tagout, access, stored energy, manual handling, ergonomics, and foreseeable misuse. Quality changes require validation, traceability, and approval under the organization’s quality-management system. Environmental considerations may include energy, water, emissions, waste, and hazardous substances.

Sustainability requires more than a favorable first-week result. The team should update:

• standard operating procedures;
• maintenance and inspection standards;
• training and qualification records;
• spare-parts specifications;
• control plans and reaction plans;
• visual management;
• audit schedules;
• escalation rules.

A control chart or time-series display can reveal whether improvement is sustained. The baseline and post-improvement periods should be comparable in product mix, volume, staffing, and operating conditions. Where they are not comparable, the team should explain the difference rather than claim a simple causal effect.

10. Technology and Contemporary Practice

Digital systems can strengthen focused improvement by automatically collecting downtime codes, cycle signals, alarm histories, energy data, and quality records. Computer vision may detect defects or unsafe conditions; predictive models may identify changing vibration or temperature patterns. Artificial intelligence can help classify events, search maintenance histories, and suggest candidate causes.

However, technology does not eliminate the need for process understanding. Poorly defined event codes produce precise-looking but unreliable analytics. AI-generated recommendations require human validation, especially where safety, quality release, or equipment control is involved. Governance should address data ownership, cybersecurity, model drift, explainability, access control, and the risk that automated monitoring encourages surveillance rather than improvement.

The most effective digital TPM systems connect three levels:

1. Signal: what changed in the process?
2. Mechanism: why could that change create a loss?
3. Action: what standardized response prevents recurrence?

Without the third level, dashboards merely display problems.

Challenges and Limitations

Several limitations deserve attention:

• Pareto analysis prioritizes recorded losses and may understate unreported or normalized losses.
• OEE can encourage local optimization and is sensitive to classification rules.
• 5 Whys may oversimplify multi-causal failures.
• Ishikawa diagrams can generate long lists without verification.
• Before-and-after studies may confuse improvement with changes in product mix or demand.
• SMED may be unsuitable for processes constrained by curing, heating, sterilization, or regulatory release unless the process is redesigned.
• Poka-yoke can create new failure modes if sensors, interlocks, or overrides are poorly governed.
• Cross-functional teams may lose momentum when sponsors do not provide time, resources, or decision authority.

Advanced teams address these limitations through sound measurement systems, risk analysis, controlled trials, causal modeling, and periodic review of whether the chosen metric still represents customer and business value.

Best-Practice Checklist

Before closing a focused-improvement project, confirm that:

• the loss is defined in measurable terms;
• the baseline and data definitions are documented;
• the team has observed the actual workplace;
• suspected causes are distinguished from verified causes;
• the physical mechanism is understood;
• countermeasures address causes rather than symptoms;
• safety, quality, and environmental risks are assessed;
• results are checked against comparable conditions;
• standards and training are revised;
• an owner and audit frequency are assigned;
• lessons are considered for horizontal deployment.

Chapter Summary

Focused improvement is the TPM discipline for systematically eliminating losses through cross-functional analysis and kaizen. Chronic losses are recurring and often normalized; sporadic losses are irregular but conspicuous. Both require a clear distinction between symptoms, immediate causes, physical mechanisms, and systemic causes.

Pareto analysis helps prioritize opportunities, while OEE provides a structured view of availability, performance, and quality. Root-cause methods such as 5 Whys, Ishikawa diagrams, and PM analysis generate and verify causal explanations. ECRS redesigns work, SMED reduces changeover through internal–external separation, and poka-yoke prevents or immediately detects error. Effective projects use a structured cycle of selection, definition, observation, analysis, action, verification, standardization, and horizontal deployment.

Technical success must be accompanied by safety, quality, employee participation, data governance, and sustainment. Digital and AI technologies can improve visibility and prediction, but they do not replace workplace observation, physical reasoning, or accountable decision-making. The central test of focused improvement is not whether an isolated result looks better, but whether the loss mechanism has been removed and the improved condition remains controlled.

Key Terms

Focused improvement: A TPM pillar that uses structured, team-based kaizen to eliminate targeted losses.

Chronic loss: A recurring loss that is often individually small but cumulatively significant.

Sporadic loss: An irregular, usually conspicuous loss such as a major breakdown or quality failure.

OEE: Overall Equipment Effectiveness, calculated from availability, performance, and quality.

Pareto analysis: Ranking loss categories by contribution to identify priority improvement themes.

Root cause: A causal condition whose removal prevents recurrence of the problem.

5 Whys: A questioning method that traces an effect toward underlying causes.

Ishikawa diagram: A cause-and-effect diagram used to organize potential causes.

PM analysis: TPM analysis of physical phenomena, mechanisms, and necessary conditions.

ECRS: Eliminate, Combine, Rearrange, and Simplify; a process-redesign framework.

SMED: Single-Minute Exchange of Die, a method for reducing changeover time.

Internal setup: Changeover work possible only while equipment is stopped.

External setup: Changeover work possible while equipment is operating.

Poka-yoke: Mistake-proofing that prevents or immediately detects error.

Basic condition: The required state of cleanliness, lubrication, tightening, alignment, and adjustment.

Horizontal deployment: Applying a proven improvement to similar processes or equipment.

Standardization: Formalizing the improved method, conditions, responsibilities, and controls.

Loss tree: A hierarchical representation linking business impact to events, mechanisms, causes, and controls.

Review Questions

Short Answer Questions

1. Distinguish between chronic and sporadic losses.
2. State the three components of OEE.
3. What is the purpose of Pareto analysis in focused improvement?
4. Explain the difference between internal and external setup in SMED.
5. Why must causes identified in an Ishikawa diagram be verified?

Long Answer Questions

1. Explain how a cross-functional TPM kaizen team should progress from loss selection to standardization.
2. Compare 5 Whys, Ishikawa analysis, and PM analysis in terms of purpose, strengths, and limitations.
3. Discuss how ECRS and poka-yoke can be combined to eliminate recurring operator adjustments.
4. Explain the assumptions and practical limitations of OEE as an improvement measure.
5. Describe the governance requirements for sustaining a focused-improvement project.

Analytical / Application Questions

1. A line operates for 420 minutes after planned breaks, experiences 36 minutes of downtime, produces 768 units at an ideal cycle time of 0.45 minutes, and makes 744 good units. Calculate availability, performance, quality, and OEE.
2. A Pareto shows that minor stoppages account for 42% of downtime, but the largest individual stoppage is a monthly breakdown accounting for 18%. Design a two-track improvement strategy for chronic and sporadic losses.
3. A changeover contains 25 minutes of internal work and 35 minutes of external work. Propose an SMED study plan and identify at least four countermeasures that could reduce internal time without compromising safety or quality.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes a chronic loss from a sporadic loss?

A. A chronic loss is recurring and often normalized, whereas a sporadic loss is irregular and usually conspicuous.

B. A chronic loss always has a smaller financial impact than a sporadic loss.

C. A chronic loss affects quality only, whereas a sporadic loss affects availability only.

D. A chronic loss is caused by operators, whereas a sporadic loss is caused by machines.

Correct Answer: A. A chronic loss is recurring and often normalized, whereas a sporadic loss is irregular and usually conspicuous.

Question 2

Using the chapter’s OEE definitions, a line has availability of 90%, performance of 80%, and quality of 95%. What is its OEE?

A. 68.4%

B. 72.0%

C. 76.0%

D. 85.5%

Correct Answer: A. 68.4%

Question 3

What is the primary purpose of Pareto analysis in focused improvement?

A. To prove that every cause has equal importance

B. To rank loss categories and select high-impact improvement themes

C. To replace workplace observation

D. To calculate the ideal cycle time

Correct Answer: B. To rank loss categories and select high-impact improvement themes

Question 4

Which method specifically analyzes physical phenomena and the conditions that produce them?

A. ECRS

B. SMED

C. PM analysis

D. OEE analysis

Correct Answer: C. PM analysis

Question 5

In SMED, which activity is an external setup activity?

A. Removing a die after the machine has stopped

B. Adjusting the machine while production is stopped

C. Preparing the correct die and tools while the machine is operating

D. Testing the first product after changeover

Correct Answer: C. Preparing the correct die and tools while the machine is operating

Question 6

Which ECRS action removes an unnecessary inspection or movement?

A. Eliminate

B. Combine

C. Rearrange

D. Simplify

Correct Answer: A. Eliminate

Question 7

Why should a cause in an Ishikawa diagram be verified?

A. Because every listed cause is automatically a root cause

B. Because the diagram generates hypotheses rather than proving causation

C. Because verification is required only for safety problems

D. Because verification replaces the need for data collection

Correct Answer: B. Because the diagram generates hypotheses rather than proving causation

Question 8

A process produces 500 units, of which 485 are good. What is the quality rate?

A. 95.0%

B. 96.0%

C. 97.0%

D. 98.5%

Correct Answer: C. 97.0%

Question 9

Which poka-yoke design is generally strongest?

A. A weekly report showing that errors occurred

B. An instruction asking the operator to be more careful

C. A control that prevents incorrect insertion at the source

D. A final inspection performed after shipment

Correct Answer: C. A control that prevents incorrect insertion at the source

Question 10

What is the correct sequence after implementing a focused-improvement countermeasure?

A. Close the project immediately to avoid changing the standard

B. Check results, standardize the improvement, and consider horizontal deployment

C. Remove all existing standards before measuring results

D. Repeat the Pareto analysis without checking safety or quality

Correct Answer: B. Check results, standardize the improvement, and consider horizontal deployment