Customer Requirements and Quality Planning

Introduction

Quality is not defined solely by conformance to an internal drawing or process standard. In a competitive production environment, quality exists when a product or service consistently satisfies requirements that matter to customers and other stakeholders. The central task of quality planning is therefore translation: converting informal, incomplete and sometimes contradictory customer statements into measurable requirements, controlled processes and objective evidence of conformity.

This chapter presents an integrated approach based on stakeholder analysis, Voice of the Customer (VOC), Kano analysis, Critical-to-Quality (CTQ) characteristics, Quality Function Deployment (QFD), specifications, tolerances and quality planning. These methods are particularly important in Six Sigma because the Define and Measure phases depend on a clear understanding of what “quality” means to the customer.

Learning Objectives

After studying this chapter, the learner should be able to:

1. Identify customers and stakeholders and assess their requirements.
2. Collect, classify and prioritize the Voice of the Customer.
3. Use Kano analysis to distinguish basic, performance and excitement requirements.
4. Translate customer statements into measurable CTQ requirements.
5. Construct and interpret a basic House of Quality.
6. Develop specifications, tolerances and acceptance criteria.
7. Prepare a quality plan linking requirements, risks, controls and records.
8. Evaluate limitations, ethical issues and implementation risks in quality planning.

1. Quality as a Customer- and Stakeholder-Defined Concept

A customer is any person or organization that receives, uses, pays for or is affected by a product, service or process. The external customer may be an end user, distributor or industrial buyer. Internal customers include downstream departments that receive work from an upstream process. Stakeholders are broader: regulators, employees, owners, suppliers, communities and environmental interests may all influence requirements.

Customer requirements may be:

• Explicit: directly stated, such as “delivery within five working days.”
• Implicit: expected even when not stated, such as safe operation.
• Mandatory: imposed by law, regulation, contract or industry standard.
• Latent: not yet articulated because the customer has not imagined the possibility.
• Technical: expressed in engineering or process language, such as dimensional accuracy.
• Perceptual: concerned with appearance, convenience, trust or emotional response.

A useful distinction is between a need, a requirement and a specification. A need is a desired outcome, such as easy installation. A requirement is a precise statement of what must be achieved, such as installation by one trained person in less than 15 minutes. A specification defines measurable limits, such as an installation time of ≤ 15 minutes under stated test conditions.

Quality planning must also recognize that customers are not homogeneous. A purchasing manager may prioritize price and delivery, an operator may prioritize usability, a maintenance engineer may prioritize reliability, and a regulator may prioritize safety. A single product may therefore have several legitimate definitions of quality.

2. Stakeholder Analysis

Stakeholder analysis identifies who can affect or be affected by the product, process or project. It prevents the common error of treating the loudest customer as the only customer.

A practical procedure is:

1. Define the product or process boundary.
2. List direct, indirect, internal and external stakeholders.
3. Identify each stakeholder’s needs, influence, authority and exposure to risk.
4. Determine how reliable each source of information is.
5. Plan an appropriate engagement method.
6. Resolve conflicts using legal, safety, ethical and strategic priorities.

A stakeholder register can be structured as follows:

Influence does not automatically determine importance. A stakeholder with little commercial power may still represent a non-negotiable safety requirement. A matrix such as “influence versus interest” is useful for engagement planning, but legal and ethical obligations must not be ignored merely because a stakeholder has low influence.

3. Capturing the Voice of the Customer

VOC data should be collected through multiple methods because every method has bias. Interviews reveal motives and language; surveys provide broader coverage; observation reveals unspoken workarounds; complaints and warranty data expose failures; market research identifies competitors; and regulatory documents define mandatory boundaries.

3.1 Qualitative and quantitative VOC

Qualitative VOC consists of statements such as “the machine is difficult to clean.” These statements explain context but cannot yet be controlled. Quantitative VOC includes measurable evidence such as cleaning time, complaint frequency or a satisfaction rating. Both are needed: qualitative data explains what matters, while quantitative data helps establish magnitude and priority.

A useful interview practice is to ask open questions first, followed by probing questions:

• What are you trying to accomplish?
• What makes the current product difficult to use?
• What happens when the requirement is not met?
• How would you recognize an acceptable result?
• Which trade-offs would you accept?

The analyst should record the customer’s words before translating them into technical language. Premature translation can introduce engineering assumptions and conceal the original need.

3.2 Affinity analysis and translation

Similar statements can be grouped using an affinity diagram. For example, “does not jam,” “works continuously” and “few stoppages” may belong to a reliability or availability theme. The group is then translated into a neutral need statement, such as “the equipment remains available during the planned production period.”

The translation should preserve the customer outcome. “Use a stainless-steel shaft” is a technical solution, not necessarily a customer requirement. The underlying need may be corrosion resistance or hygienic cleanability. Keeping needs separate from solutions encourages better design alternatives.

4. Kano Analysis

The Kano model classifies requirements according to their effect on customer satisfaction. The most widely used categories are:

• Must-be (basic) requirements: their presence may not increase satisfaction, but their absence causes strong dissatisfaction. Safety interlocks are a typical example.
• One-dimensional (performance) requirements: satisfaction generally increases as performance improves. Delivery speed and battery life often behave this way.
• Attractive (excitement) requirements: unexpected features that increase satisfaction when present but do not usually cause dissatisfaction when absent.
• Indifferent requirements: features that have little effect on satisfaction.
• Reverse requirements: features that satisfy some customers but dissatisfy others.

Kano classification is not permanent. An attractive feature can become a performance requirement and eventually a basic expectation as competitors adopt it. Therefore, Kano analysis should be repeated when markets, technologies or regulations change.

A structured Kano questionnaire presents a functional question, asking how a customer feels if a feature is present, and a dysfunctional question, asking how the customer feels if it is absent. Response categories may include “I like it,” “I expect it,” “I am neutral,” “I can tolerate it” and “I dislike it.” The paired responses are classified using a Kano evaluation table. Results should be interpreted alongside customer segmentation and sample quality; a small or unrepresentative sample can produce misleading priorities.

Kano analysis is especially valuable when resources are limited. Mandatory requirements should be protected from failure, performance requirements should be optimized against competitors and attractive requirements should be assessed for strategic value and cost.

5. From Customer Statements to CTQs

A Critical-to-Quality characteristic is a measurable attribute whose performance is important to customer satisfaction, safety, compliance or business success. A CTQ tree provides a disciplined translation from a broad need to an operational measure.

Example:

• Need: “The product should be easy to install.”
• Driver: low installation effort and clear orientation.
• CTQ: installation time, number of tools and error-free first installation.
• Measure: median installation time ≤ 15 minutes, using no more than two specified tools, with a first-attempt success rate ≥ 98%.

A CTQ should have:

1. A clear characteristic.
2. An operational definition.
3. A measurement method.
4. A target or specification.
5. A sampling or test plan.
6. A link to customer or stakeholder value.

The operational definition must state what is measured, where, when, by whom and with which instrument or method. For example, “delivery reliability” is incomplete. “Percentage of confirmed orders delivered to the agreed destination by the promised date, measured monthly from the enterprise resource planning system” is operationally stronger.

A common prioritization method assigns each CTQ a customer importance score, business impact and risk. A simple priority index may be used for screening:

where P_i is the priority index for CTQ i, I_i is its importance score and R_i is its risk score. The scores must be defined, for example on a 1–5 ordinal scale. This multiplication is a management prioritization device, not a probability law; it should not be interpreted as a statistically precise risk estimate. Safety, legal and ethical requirements must override a low numerical score.

6. Quality Function Deployment

Quality Function Deployment translates customer requirements into technical characteristics and deployment actions. Its principal tool is the House of Quality (HOQ). The HOQ normally contains:

1. Customer requirements, or “WHATs.”
2. Importance ratings for the WHATs.
3. Technical characteristics, or “HOWs.”
4. A relationship matrix linking WHATs to HOWs.
5. A roof showing correlations among HOWs.
6. Competitive or benchmarking information.
7. Technical targets and priorities.

Relationship strengths are commonly represented as 9 = strong, 3 = medium, 1 = weak and 0 = no relationship. These values are conventions, not universal scientific constants. The team must define them consistently.

If w_i is the importance of customer requirement i and r_ij is the relationship score between requirement i and technical characteristic j, a basic technical priority is:

where T_j is the unnormalized priority of technical characteristic j, m is the number of customer requirements, w_i is the importance weight of requirement i and r_ij is the relationship score. The result has no physical unit because it is a weighted planning score. It is meaningful only for comparing characteristics within the same HOQ and scoring system.

Worked example

Suppose three customer requirements have importance ratings 5, 4 and 3. A technical characteristic, “seal compression,” has relationship scores 9, 3 and 1 respectively. Its priority is:

Another characteristic should not be declared less important solely because it has a lower score if it is safety-critical or legally required. The numerical result supports discussion; it does not replace engineering judgment.

The roof of the HOQ reveals technical conflicts. Increasing wall thickness may improve strength but increase mass. A relationship may be positive, negative or negligible. Such conflicts should be resolved through design experiments, simulation, risk analysis and customer-value assessment rather than hidden in the planning process.

Figure: The principal components of a simplified House of Quality.

QFD can be extended through several “houses”: product planning, part deployment, process planning and production planning. The method promotes cross-functional communication, but it is not a substitute for testing. A poorly understood customer requirement produces a detailed but invalid matrix.

7. Specifications, Tolerances and Acceptance Criteria

A specification states the required value, range or condition of a characteristic. A tolerance defines the permitted variation around a target. For a bilateral tolerance:

where x is the measured characteristic, x_L is the lower specification limit and x_U is the upper specification limit. If the target is T with symmetric tolerance t, then x_L = T − t and x_U = T + t.

For example, a shaft specified as 20.00 ± 0.05 mm has a lower limit of 19.95 mm and an upper limit of 20.05 mm. The tolerance width is 0.10 mm. Units and measurement conditions must always be stated.

Not all requirements are bilateral. A maximum delivery time is one-sided: x ≤ 5 working days. A minimum strength requirement is x ≥ 500 N. Attribute requirements may be binary, such as “no visible crack,” but the inspection conditions and definition of “visible” must be documented.

Tolerance allocation and capability

A tolerance must be achievable by the process and sufficient for function. If a process is approximately stable and normally distributed, the commonly used capability indices are:

Here USL and LSL are upper and lower specification limits, μ is the process mean and σ is the within-process standard deviation. C_p measures potential capability assuming centering; C_pk also reflects displacement from the target. These indices require a stable process, a meaningful estimate of within-process variation and appropriate measurement integrity. They should not be used to disguise instability or non-normality.

For the shaft example, suppose USL = 20.05 mm, LSL = 19.95 mm and a stable process has σ = 0.01 mm with mean μ = 20.00 mm. Then:

and, because the process is centered:

Capability indices are not acceptance criteria by themselves. The organization must define its required threshold, considering risk, industry practice and customer agreement.

Measurement-system integrity

A specification is only useful if the measurement system can distinguish conforming from nonconforming output. Gauge repeatability and reproducibility studies, calibration, resolution checks and standardized methods help establish confidence. Measurement uncertainty should be considered near specification limits, especially in regulated or safety-critical applications. A product should not be rejected or accepted solely because of an unreliable instrument.

8. Quality Planning Methods

Quality planning connects requirements to process controls. Common tools include:

• SIPOC: defines suppliers, inputs, process, outputs and customers at a high level.
• Process mapping: shows sequence, decision points, handoffs and controls.
• Design FMEA (DFMEA): examines potential product failure modes.
• Process FMEA (PFMEA): examines potential process failure modes and controls.
• Control plan: specifies characteristics, methods, frequency, responsibility and reaction plans.
• Inspection and test plan: defines verification stages and acceptance criteria.
• APQP-style planning: integrates customer requirements, design, process development, validation and launch controls.

A control plan should answer five questions: What characteristic is controlled? What is the specification? How is it measured? How often is it checked? What happens when the result is abnormal?

A useful control-plan structure is:

Prevention is normally stronger than detection. A keyed fixture that prevents incorrect assembly is preferable to relying only on final inspection. However, detection controls remain important where prevention is impractical or residual risk is high.

9. Illustrative Case: Packaging-Line Seal Failure

Background

A manufacturer of sterile fluid containers receives complaints that some packages leak during transport. The customer requirement is initially expressed as “the package must remain secure and hygienic.”

Problem

The statement is too vague for production control. The organization also discovers that purchasing, production and quality departments have different assumptions about the meaning of “secure.”

Analysis

Stakeholder analysis identifies hospital users, logistics providers, purchasing, production operators, the regulatory authority and the final patient as relevant stakeholders. Interviews and complaint analysis reveal three important outcomes: no leakage, easy opening by authorized users and evidence that the package has not been tampered with.

Kano analysis classifies sterile integrity and tamper evidence as must-be requirements. Opening force is a performance requirement: excessive force causes dissatisfaction, but insufficient force may also create risk. A new colour-changing indicator is classified as attractive, but only after regulatory acceptability is established.

The team develops CTQs:

• leakage rate under a defined pressure and time test: ≤ 2 mL/min;
• seal width: 3.0 ± 0.2 mm;
• opening force: 20–35 N under specified temperature and speed conditions;
• tamper indicator: correctly changes state after opening;
• label identity: 100% correct by barcode verification.

The HOQ shows that sealing temperature, pressure, dwell time and material compatibility strongly affect leakage. It also reveals a conflict: increasing seal strength improves integrity but may increase opening force. Engineering trials are therefore planned rather than assuming that one technical variable can satisfy all needs.

Approach and solution

The team conducts a measurement-system assessment, confirms that the leakage tester is calibrated, and performs a designed trial of sealing temperature and dwell time. PFMEA identifies incorrect material loading and inadequate seal temperature as high-risk failure modes. The control plan introduces automatic material verification, continuous temperature monitoring, first-off approval and periodic destructive seal testing. Nonconforming product is segregated, and a documented reaction plan defines when production must stop.

Outcome

After process validation, the organization demonstrates conformance to the CTQs under specified conditions. Complaints decrease, and the customer receives objective test evidence rather than a general statement of quality. The attractive tamper-indicator feature is introduced only after regulatory and usability review.

Learning points

The case demonstrates that VOC is not a list of slogans. It must be translated into measurable outcomes, linked to technical causes, verified by capable measurement systems and controlled throughout the process. It also shows why regulatory, user and operational requirements must be considered together.

10. Challenges, Ethics and Contemporary Practice

Quality planning can fail through biased samples, leading questions, excessive technical jargon, conflicting stakeholder objectives or false precision in scoring. Customer surveys may overrepresent frequent respondents, while complaint data may underrepresent silent dissatisfaction. Teams should document sources, assumptions, uncertainty and unresolved conflicts.

Ethically, organizations must not conceal safety or compliance requirements behind commercial priorities. Customer data must be collected and stored lawfully, with appropriate consent and confidentiality. Accessibility requirements should be considered so that quality is not defined only for the average user. Sustainability is increasingly a customer and stakeholder requirement, including energy use, recyclability, repairability and responsible sourcing.

Digital tools can support text mining of complaints, sentiment analysis, automated QFD updates and real-time traceability. Artificial intelligence can identify themes in large VOC datasets, but it may reproduce sampling bias, misinterpret ambiguous language or expose confidential information. Human review, data governance, explainability and cybersecurity are essential. AI-generated priorities must be treated as decision support, not as unquestionable customer truth.

Contemporary quality planning increasingly links QFD, FMEA, digital twins, product-lifecycle management and statistical process monitoring. The key principle remains unchanged: technology is valuable only when it improves the validity, timeliness and controllability of customer requirements.

Best-Practice Checklist

Before approving a quality plan, ask:

1. Have all relevant customers and stakeholders been identified?
2. Are customer statements supported by more than one evidence source?
3. Has each important need been translated into an operational CTQ?
4. Are legal, safety and ethical requirements explicitly identified?
5. Are specifications measurable, unambiguous and unit-consistent?
6. Are tolerance limits functionally justified and process-capable?
7. Is the measurement system adequate for the decision?
8. Does the QFD show technical conflicts and priorities?
9. Does the control plan include prevention and reaction actions?
10. Are assumptions, changes and approval responsibilities documented?

Chapter Summary

Customer requirements provide the foundation for quality planning and Six Sigma improvement. Stakeholder analysis identifies the different parties whose needs, authority and risks shape quality. VOC methods collect explicit, implicit and latent requirements, which are organized and translated into measurable CTQs. Kano analysis distinguishes basic, performance and attractive requirements and helps allocate improvement effort. QFD connects customer WHATs to technical HOWs, while specifications and tolerances define objective acceptance boundaries. Capability analysis can assess whether a stable process is able to meet those boundaries, provided its assumptions and measurement integrity are satisfied. Finally, control plans, FMEA, process maps and test plans convert requirements into preventive and detective controls. Effective quality planning is cross-functional, evidence-based, ethically governed and continuously updated as customer expectations and technology change.

Key Terms

Acceptance criterion: A defined condition used to judge whether an output is acceptable.

Attractive requirement: A Kano feature that increases satisfaction when present but may not cause dissatisfaction when absent.

Capability: The ability of a stable process to produce output within specification limits.

Control plan: A documented plan linking process characteristics to methods, frequency, responsibility and reaction actions.

Critical-to-Quality (CTQ): A measurable characteristic that is important to customer value, compliance, safety or performance.

Customer: A recipient, user, purchaser or beneficiary of a product or service.

FMEA: Failure Modes and Effects Analysis, a structured method for identifying and controlling potential failures.

House of Quality: The principal matrix used in Quality Function Deployment.

Kano analysis: A method for classifying requirements according to their effect on satisfaction.

Must-be requirement: A basic requirement whose absence causes strong dissatisfaction.

Operational definition: A precise statement of what, how, when and under what conditions a measure is obtained.

Quality Function Deployment (QFD): A method for translating customer requirements into technical and process requirements.

Specification limit: The boundary defining acceptable product or process performance.

Stakeholder: A person or organization that affects or is affected by a product, process or decision.

Tolerance: The permitted variation around a target or nominal value.

Voice of the Customer (VOC): Structured information about customer needs, expectations, experiences and priorities.

WHAT: In QFD, a customer requirement or desired outcome.

HOW: In QFD, a technical characteristic or method used to satisfy a WHAT.

Review Questions

Short Answer Questions

1. Distinguish between a customer need, a requirement and a specification.
2. What is the purpose of stakeholder analysis in quality planning?
3. Define a Critical-to-Quality characteristic and give one example.
4. What is the difference between a must-be and an attractive Kano requirement?
5. Why must a measurement system be evaluated before capability indices are interpreted?

Long Answer Questions

1. Explain a systematic method for collecting and translating the Voice of the Customer into operational requirements.
2. Describe the structure and purpose of a House of Quality, including the role of the relationship matrix and technical correlation roof.
3. Explain how specifications and tolerances should be developed for a manufacturing characteristic.
4. Compare DFMEA, PFMEA and the control plan in an integrated quality-planning system.
5. Discuss the ethical, regulatory and data-governance issues that may arise when customer requirements are collected and analyzed digitally.

Analytical / Application Questions

1. A customer says that a product is “reliable.” Develop a CTQ tree with at least two measurable CTQs, including operational definitions and suitable acceptance criteria.
2. A process has USL = 10.20 mm, LSL = 9.80 mm, μ = 10.10 mm and σ = 0.05 mm. Calculate C_p and C_pk, state the assumptions required and interpret the difference.
3. Construct a small QFD relationship analysis for three customer requirements and three technical characteristics. Explain how you would handle a technical conflict between a requirement for high strength and a requirement for low weight.

Multiple Choice Questions with Answers

Question 1

Which statement best describes a Critical-to-Quality characteristic?

A. A customer opinion that has not been analyzed

B. A measurable characteristic important to customer value or compliance

C. A production activity that has no acceptance criterion

D. A financial target used only by purchasing

Correct Answer: B. A measurable characteristic important to customer value or compliance

Question 2

In the Kano model, a must-be requirement is one that:

A. Always creates delight when improved

B. Has no effect on satisfaction

C. Causes strong dissatisfaction when absent

D. Is prohibited by regulation

Correct Answer: C. Causes strong dissatisfaction when absent

Question 3

In QFD, the customer requirements are commonly called:

A. HOWs

B. WHATs

C. ROOFs

D. LIMITs

Correct Answer: B. WHATs

Question 4

For a dimension specified as 20.00 ± 0.05 mm, the upper specification limit is:

A. 19.95 mm

B. 20.00 mm

C. 20.05 mm

D. 20.10 mm

Correct Answer: C. 20.05 mm

Question 5

Using T_j = ∑w_i r_ij, what is the priority when importance ratings are 5, 4 and 3 and relationship scores are 9, 3 and 1?

A. 30

B. 45

C. 60

D. 81

Correct Answer: C. 60

Question 6

Which tool primarily defines characteristics, measurement methods, frequency and reaction plans during production?

A. Control plan

B. Affinity diagram

C. Stakeholder register

D. Kano questionnaire

Correct Answer: A. Control plan

Question 7

The index C_p primarily indicates:

A. Customer satisfaction directly

B. Potential process capability assuming centering

C. The number of defects per customer

D. The average measurement value

Correct Answer: B. Potential process capability assuming centering

Question 8

Which is the strongest example of translating a vague need into an operational requirement?

A. The product should be attractive

B. The product should be modern

C. Installation shall take no more than 15 minutes under defined conditions

D. The product should satisfy everyone

Correct Answer: C. Installation shall take no more than 15 minutes under defined conditions

Question 9

Why should technical correlations be examined in the House of Quality?

A. To identify conflicts or synergies among technical characteristics

B. To replace customer research completely

C. To eliminate the need for measurement

D. To calculate employee wages

Correct Answer: A. To identify conflicts or synergies among technical characteristics

Question 10

A one-sided requirement is best represented by:

A. x = T

B. x_L ≤ x ≤ x_U

C. x ≤ USL

D. x ≠ T

Correct Answer: C. x ≤ USL