4. Voice of the Customer and Customer Insight

Introduction and Context

Quality management begins with a deceptively difficult question: what does the customer actually value? Organizations often answer this question through internal assumptions, historical specifications, or the preferences of a small number of vocal customers. Such approaches are risky because customers do not usually express quality in the technical language used by designers, engineers, or process managers. They describe experiences, frustrations, desired outcomes, and trade-offs. A customer may say that a banking application should be “simple,” a hotel room should “feel clean,” or a delivery service should be “reliable.” These statements are meaningful, but they are not yet operational requirements.

Voice of the Customer (VoC) is the systematic process of collecting, interpreting, prioritizing, and deploying customer expectations. Customer insight is the deeper understanding created when customer statements are combined with behavioural evidence, operational data, market context, and analytical interpretation. Together, VoC and customer insight connect external expectations with internal quality design.

For Total Quality Management (TQM), this connection is central. Quality is not determined solely by conformance to internal specifications; it is judged through the customer’s experience of value. A well-designed VoC system therefore supports customer focus, evidence-based decision-making, process improvement, innovation, service recovery, and strategic alignment. It also reduces the danger of treating every customer statement as an isolated request. The purpose is not merely to listen, but to convert listening into disciplined organizational learning and critical-to-quality (CTQ) requirements.

Learning Objectives

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

1. Explain the meaning, scope, and strategic importance of Voice of the Customer and customer insight.

2. Distinguish stated needs, latent needs, expectations, requirements, and customer outcomes.

3. Select and design appropriate VoC methods, including interviews, observation, focus groups, surveys, complaints, reviews, and social listening.

4. Analyse qualitative and quantitative customer evidence using coding, affinity analysis, segmentation, and prioritization.

5. Translate customer language into measurable CTQ requirements using Quality Function Deployment (QFD) and the House of Quality.

6. Evaluate the ethical, technological, managerial, and operational challenges of customer insight systems.

7. Apply VoC principles to a quality-improvement or service-design decision.

From Customer Statements to Customer Insight

The term customer includes more than the person who pays. In many value chains there are users, purchasers, influencers, beneficiaries, intermediaries, regulators, employees serving the customer, and internal customers. A hospital, for example, must understand the needs of patients, family members, physicians, insurers, and regulatory bodies. These groups may have different expectations and different definitions of quality.

A customer statement is an expression of a desire, problem, perception, or preference. A need is the underlying outcome the customer seeks. A requirement is a precise condition that an organization can design, measure, and control. Consider the following progression:

• Customer statement: “I do not want to wait all afternoon.”
• Interpreted need: predictable and reasonable access to service.
• Requirement: 90 per cent of scheduled customers are attended within 15 minutes of the appointment time.
• CTQ measure: percentage of customers attended within the specified time.

This translation must be handled carefully. If the organization converts “I want a friendly service” into a superficial requirement such as “employees must smile,” it may measure an internal behaviour without addressing the real outcome, which could be confidence, respect, or ease of resolution.

Customer insight is not identical to customer data. Data may show that complaints about delivery delays increased by 12 per cent. Insight requires interpretation: which customers are affected, at which stage, under what conditions, with what consequences, and why has the pattern emerged? Insight combines evidence with context and leads to a decision or action.

Important concepts and terminology

Voice of the Customer may be understood through several distinctions:

• Stated needs: Needs explicitly expressed by customers.
• Latent needs: Needs customers may not articulate because they regard them as obvious or have not imagined a better solution.
• Unstated expectations: Assumptions customers hold about acceptable performance, such as privacy, accuracy, or basic safety.
• Delighters: Features or experiences that create disproportionate satisfaction when present, although customers may not demand them in advance.
• Customer outcome: The result the customer is trying to achieve, rather than the product or process supplied.
• Moment of truth: A significant interaction at which the customer forms or revises a judgement about the organization.
• Critical-to-quality requirement: A customer-derived requirement whose performance materially affects perceived quality, risk, or value.
• Customer journey: The sequence of stages and interactions through which a customer pursues an outcome.
• Customer effort: The time, cognitive work, repetition, or inconvenience required to obtain a service or resolve a problem.
• Unstructured data: Text, audio, images, video, or social content that does not initially exist in a numerical table.

A useful distinction is that satisfaction is an evaluation of an experience, whereas loyalty, retention, advocacy, and trust are possible consequences of repeated experiences. A single satisfaction score should therefore not be treated as a complete representation of customer value.

Designing a Voice of the Customer System

A credible VoC system is a management process rather than a one-time research exercise. It should answer five questions: whose voice is required, what decision will the evidence support, which method is suitable, how will findings be interpreted, and how will action be monitored?

Defining the decision and the customer population

Research should begin with a decision statement. “Understand customers better” is too vague. “Redesign the claims process to reduce abandonment while protecting fraud controls” provides a basis for selecting participants, questions, measures, and outputs.

The customer population should be segmented where needs, value, risk, or behaviour differ. Relevant bases may include use case, frequency, account value, geography, accessibility needs, channel, lifecycle stage, or problem severity. Sampling only active and satisfied customers can create survivorship bias. Complaining customers, customers who defected, infrequent users, vulnerable groups, and non-users may provide essential evidence.

Interviews: depth and meaning

Semi-structured interviews are appropriate when the organization needs to understand motivations, experiences, language, and causes. The interviewer uses a consistent topic guide while allowing participants to explain unexpected issues. Effective questions invite concrete accounts: “Tell me about the last time you used the service,” “What happened next?” and “What made that difficult?” Probing should explore consequences and workarounds rather than suggest preferred answers.

Interviews are especially useful at the exploratory stage, before a survey is designed. Their limitations include small samples, interviewer influence, recall error, and the tendency of participants to rationalize behaviour after the event. Findings should generally be treated as thematic evidence, not as estimates of population frequency.

Observation and contextual inquiry

Observation examines what customers do in the environment where use occurs. It can reveal workarounds, interruptions, accessibility barriers, informal practices, and differences between reported and actual behaviour. A customer may say that a form is easy to complete, while observation shows that the customer keeps a separate note because the form does not explain required information.

Contextual inquiry combines observation with questions asked at appropriate moments. Researchers must distinguish description from interpretation: “the user opened three browser tabs” is an observation; “the website is confusing” is an interpretation requiring further evidence. Ethical observation requires informed consent where individuals are identifiable and special care in private, workplace, or sensitive settings.

Focus groups and group interaction

Focus groups are moderated discussions involving participants selected for relevant experience. They are useful for exploring language, reactions to concepts, competing expectations, and social meanings. Group interaction can stimulate ideas that might not arise in an individual interview. It can also suppress minority views through conformity, status differences, or dominant participants. A skilled moderator establishes ground rules, invites quieter voices, tests disagreement, and avoids allowing consensus to be mistaken for truth.

Focus groups are weak instruments for estimating incidence. A comment made by six participants does not imply that a particular percentage of the market holds that view.

Surveys and measurement

Surveys are suitable when the organization needs structured comparison across a larger sample. A sound survey specifies the construct being measured, uses clear and neutral wording, avoids double-barrelled questions, provides balanced response options, and is tested before launch. “How satisfied are you with the speed and accuracy of the service?” combines two attributes and prevents precise interpretation.

A survey may measure perceived performance, importance, satisfaction, effort, trust, intention, or likelihood of recommendation. The instrument should distinguish importance from satisfaction. A low score on an attribute that customers do not value may not deserve immediate investment, while a moderate score on a highly important attribute may represent a serious quality gap.

Sampling error is only one source of survey error. Non-response, coverage error, poor wording, response scale effects, timing, and social desirability can be equally important. A high response rate does not automatically establish validity.

Complaints, reviews, and service records

Complaints are unsolicited signals of failure, unmet expectations, or perceived unfairness. They are valuable because they often contain a description of the customer’s effort and consequence. However, complaints are not a simple mirror of all customer experiences. Some customers complain repeatedly, while others silently leave. Complaint volume is affected by ease of access, organizational culture, customer awareness, and the seriousness of the problem.

Online reviews provide rich contextual information and can reveal recurring themes, but extreme experiences may be overrepresented. Organizations should analyse review content without manipulating, suppressing, or fabricating reviews. Complaint and review analysis should combine volume, severity, recurrence, customer segment, process location, and resolution quality.

Social listening

Social listening involves monitoring publicly available digital conversations, mentions, communities, and sentiment-related patterns. It can identify emerging concerns rapidly and reveal how customers describe an organization in their own language. Automated sentiment classification is useful for screening large volumes, but sarcasm, multilingual expression, context, bots, and coded language can reduce accuracy. Social listening should be treated as a signal-generation method requiring validation, not as an unquestionable measurement of public opinion.

Analysing Customer Evidence

VoC analysis requires disciplined movement from raw evidence to a manageable set of quality priorities. A common workflow is as follows:

1. Collect and preserve the evidence. Record source, date, customer segment, channel, context, and relevant consent conditions.

2. Clean and familiarise. Remove duplicates, correct obvious data errors, and read or review the material repeatedly.

3. Code the evidence. Assign labels to statements, such as waiting, clarity, reliability, empathy, price transparency, or recovery.

4. Separate need from solution. “Add a progress bar” may be a proposed solution; “I need to know what will happen next” is the underlying need.

5. Cluster related codes. Group similar ideas while preserving meaningful differences.

6. Prioritize. Consider importance, frequency, severity, strategic relevance, feasibility, and the consequences of failure.

7. Validate. Compare sources and test interpretations with customers, frontline employees, or additional data.

8. Translate into requirements. Express priorities in observable and measurable terms.

Affinity analysis

Affinity analysis is a collaborative method for organizing qualitative observations. Individual statements are written as separate items, then grouped according to natural similarity. Groups are named with an interpretive heading that captures the common need. For example, statements such as “I had to repeat my account number,” “the agent asked for information I had already submitted,” and “each channel seemed to start from the beginning” may form the affinity “customer information is not carried across channels.”

The process should not be reduced to mechanically counting similar words. Two statements may use different words but express the same need, while identical words may have different meanings in different contexts. An audit trail from raw quote to code, cluster, and requirement improves transparency.

Prioritization and triangulation

Prioritization can use a combination of customer importance, current performance, frequency, risk, and economic or strategic effect. A simple gap view compares perceived importance with perceived performance. More advanced approaches use weighted scores, driver analysis, conjoint methods, or linkage to behavioural outcomes such as retention and repeat use.

Triangulation means examining a question through different sources or methods. If interviews, complaints, and transaction data all indicate that customers experience uncertainty during delivery, confidence in the finding increases. Contradictions are not necessarily defects; they may reveal segment differences or a gap between intention and behaviour.

Translating Needs into CTQs through QFD

Quality Function Deployment provides a structured method for translating the Voice of the Customer into design and process priorities. Its best-known representation is the House of Quality (HoQ), although QFD can extend through several linked matrices from customer requirements to product characteristics, process plans, and control measures.

The HoQ normally contains the following elements:

• Customer requirements, or WHATs: What customers need or expect.
• Importance ratings: The relative significance of each customer requirement.
• Technical characteristics, or HOWs: Measurable design or process attributes that influence the WHATs.
• Relationship matrix: The strength of the relationship between each WHAT and HOW.
• Correlation roof: Positive or negative relationships among technical characteristics.
• Competitive assessment: Customer perception of the organization and alternatives.
• Technical priorities and targets: The characteristics requiring attention and the desired performance levels.

Suppose a university is redesigning online enrolment. Customer requirements might include “know whether my application is complete,” “receive help without repeating information,” and “obtain confirmation quickly.” Possible technical characteristics include real-time status visibility, integrated case history, response-time standards, validation rules, and confirmation-message design.

The relationship matrix asks how strongly each technical characteristic affects each customer requirement. A scoring convention such as 9 for strong, 3 for moderate, 1 for weak, and 0 for no relationship may be used. If a requirement has importance 5 and a technical characteristic has a strong relationship, its contribution is 45 before other adjustments. The calculation is less important than the reasoning behind the rating. A numerical matrix can create false precision if relationships are assigned casually.

The correlation roof identifies technical trade-offs. For example, increasing the number of security checks may improve fraud control but increase customer effort. Faster response may require additional staffing or automation. QFD makes such tensions visible so that managers can negotiate targets rather than optimize one attribute in isolation.

A CTQ should be specific enough to control. “Improve communication” is weak. “Provide an accurate status update after each processing stage, accessible through the customer portal and sent within five minutes of a status change” is more actionable. The requirement should also have an owner, a measurement method, a target, and a review frequency.

QFD does not replace managerial judgement or customer research. It structures judgement, exposes assumptions, and creates traceability from customer need to design decision. It is most effective when representatives from marketing, operations, design, quality, technology, finance, and frontline service work together.

Applied Case: Redesigning a Municipal Waste-Collection Service

Background

The fictional municipality of Riverdale received increasing complaints about missed household waste collections. The service department initially proposed purchasing additional collection vehicles. Senior management commissioned a VoC study before approving the investment.

Situation and problem

Complaint records showed that missed collections were concentrated on certain routes, but the records did not explain whether the core problem was vehicle reliability, collection timing, communication, or inconsistent household practices. An online survey had produced a generally acceptable satisfaction score, yet residents continued to post angry comments after missed collections.

Analysis

The project team segmented residents by housing type, collection frequency, route, and use of assisted services. It conducted interviews with residents, observed collection crews on several routes, reviewed complaints and social-media posts, and surveyed a representative sample of households.

Affinity analysis identified five themes: uncertainty about collection times, difficulty reporting a missed collection, inconsistent guidance about bulky items, lack of feedback after reporting, and concerns about cleanliness after collection. Observation showed that residents often placed waste outside early because collection timing was unpredictable. Crew interviews revealed that route changes were not always communicated to the public.

The team distinguished solutions from needs. “Send a text message” was treated as one possible solution to the need for reliable information. The underlying customer requirements were to know the expected collection window, report a failure easily, receive confirmation that the report was accepted, and understand when corrective action would occur.

QFD approach and solution

In the House of Quality, the highest-priority technical characteristics were route-status accuracy, exception detection, a simple mobile and telephone reporting process, and a defined response-time standard. The roof revealed a trade-off: narrowing collection windows could improve predictability but might increase routing cost. The team therefore chose a phased approach rather than promising a narrow window immediately.

The municipality introduced route-status updates, a single missed-collection case number across channels, automatic acknowledgement, and a service standard requiring action within one working day. Crew instructions were revised, and supervisors began reviewing exception data by route. A pilot was conducted before wider implementation.

Outcome

The pilot did not eliminate all missed collections, but it reduced uncertainty and repeated contacts. Residents could see whether a report was active, while managers could distinguish genuine route failures from incorrect placement or access problems. Importantly, the project avoided treating complaint volume alone as the target. It measured missed collections, reporting effort, response time, repeat contacts, and resident confidence.

Learning points

The case demonstrates that a visible complaint may be only the surface expression of a broader need. It also shows why VoC should combine expressed opinion with observation and operational evidence. QFD helped the municipality move from “buy more vehicles” to a balanced service design involving information, process control, technology, and frontline practice.

Managerial Applications and Professional Practice

A mature VoC system has implications beyond marketing. Operations managers use CTQs to define process standards. Quality professionals use customer evidence to select improvement projects and construct control plans. Product managers use insight to prioritize features. Service leaders use journey evidence to redesign handoffs and recovery. Executives use segmented customer insight to align investment with strategic value.

VoC should be integrated with complaint management, customer-relationship management, process-improvement methods, employee feedback, and performance review. When customer requirements are not linked to owners and measures, research becomes an attractive report rather than a management system. Governance should specify who can commission research, who approves instruments, how conflicting customer needs are resolved, and how actions are tracked.

Challenges, Risks, and Common Mistakes

Several errors repeatedly weaken VoC programmes. Organizations may listen only to current, cooperative, or highly satisfied customers; ask leading questions; confuse frequency with importance; equate social-media volume with market representation; or treat a single score as the customer experience. They may also collect more data than they can interpret, conduct research without a decision owner, or close the feedback loop only with internal stakeholders.

Another danger is solution bias. Customers are often asked to choose among organizationally familiar options before their underlying problem is understood. A related problem is metric substitution: an organization improves an easy-to-measure indicator while the customer outcome remains unchanged.

Conflicting needs require explicit trade-offs. One segment may value speed, another personal reassurance, and another low cost. The answer is not always to maximize every attribute. Segmentation, service tiers, informed prioritization, and transparent communication may be necessary.

Ethical, Governance, and Digital Considerations

Customer insight involves power and privacy. Participants should understand the purpose of data collection, the use of recordings, voluntary participation, retention periods, and any sharing with third parties. Sensitive information requires heightened protection and access control. Publicly available social content is not automatically ethically unrestricted; context, expectation of privacy, and potential harm remain relevant.

Surveys and analytics should avoid discriminatory sampling and inaccessible design. Automated text and sentiment analysis can reproduce language or demographic bias. Models should be validated across relevant segments, and high-impact decisions should not be based solely on opaque classifications. Organizations must distinguish anonymization from simple removal of names: combinations of location, time, and behaviour may re-identify individuals.

Digital systems can support transcription, topic modelling, speech analytics, dashboarding, journey analysis, and real-time alerting. Artificial intelligence can accelerate coding and identify emerging themes, but human review remains necessary for context, irony, ambiguity, minority signals, and ethical judgement. Automation should assist listening, not replace accountability.

Emerging Directions and Best Practices

Customer insight is increasingly continuous rather than periodic. Organizations combine transactional data, interaction records, product telemetry, reviews, research communities, and employee observations to detect changes in need. Journey analytics can connect an expressed complaint with preceding digital behaviour and subsequent service outcomes. Generative AI may help summarize large volumes of feedback, but governance must address confidentiality, hallucinated interpretations, model drift, and traceability.

Best practice is characterized by a few disciplines: begin with a decision; define customer segments and outcomes; combine methods; preserve the customer’s language; distinguish needs from solutions; document the route from evidence to CTQ; test interpretations; quantify uncertainty where appropriate; assign owners; close the feedback loop; and review whether improvements changed the customer outcome. Listening becomes valuable when it changes design, behaviour, and control.

Chapter Summary

Voice of the Customer is a structured approach to discovering and deploying customer expectations. Customer insight goes beyond collecting opinions by interpreting needs in context and linking them to decisions. Interviews, observation, focus groups, surveys, complaints, reviews, and social listening each reveal different aspects of experience and each has limitations. Effective systems use segmentation, coding, affinity analysis, triangulation, and prioritization to convert raw evidence into meaningful requirements.

Quality Function Deployment and the House of Quality provide a traceable method for translating customer requirements into measurable technical and process characteristics. CTQs must be observable, owned, measurable, and connected to customer outcomes. A reliable VoC system also requires ethical governance, privacy protection, careful use of automation, and continuous feedback into improvement. Its ultimate purpose is not to accumulate data but to align organizational capability with what customers value.

Key Terms

Voice of the Customer: A systematic process for collecting and deploying customer expectations.

Customer insight: Interpreted understanding of customer needs, behaviour, context, and value.

Stated need: A need explicitly expressed by a customer.

Latent need: An important need that customers may not articulate directly.

Customer outcome: The result a customer seeks to achieve.

Critical-to-quality requirement: A measurable requirement that materially affects perceived quality or risk.

Moment of truth: An interaction that significantly shapes customer judgement.

Customer journey: The sequence of stages and interactions involved in achieving an outcome.

Customer effort: The time, work, inconvenience, or repetition required from the customer.

Affinity analysis: A method of grouping related qualitative observations into meaningful themes.

Triangulation: Comparison of evidence from multiple sources or methods.

Quality Function Deployment: A structured approach for translating customer requirements into design and process priorities.

House of Quality: A matrix-based representation used within QFD to relate customer requirements to technical characteristics.

WHATs: Customer requirements represented in a House of Quality.

HOWs: Technical or process characteristics used to satisfy customer requirements.

Relationship matrix: The part of QFD that indicates the strength of links between WHATs and HOWs.

Correlation roof: The QFD section showing relationships and trade-offs among technical characteristics.

Social listening: Monitoring and analysing relevant public digital conversations.

Customer journey analytics: Analysis of behaviour and interactions across journey stages and channels.

Closed-loop feedback: A process in which feedback leads to action and the customer or organization receives follow-up.

Review Questions

Short Answer Questions

1. Distinguish between a customer statement, an underlying need, and a CTQ requirement.

2. Why should complaints not be treated as a complete representation of customer experience?

3. What is the purpose of affinity analysis?

4. Differentiate the WHATs and HOWs in a House of Quality.

5. State two ethical considerations in social listening.

Long Answer Questions

1. Explain how an organization can design an integrated Voice of the Customer system using qualitative and quantitative methods.

2. Discuss the strengths and limitations of interviews, observation, focus groups, and surveys.

3. Describe the process of converting raw customer evidence into prioritized CTQs.

4. Explain the structure and managerial value of the House of Quality, including the significance of technical correlations.

5. Analyse the role of governance, privacy, and human judgement in technology-enabled customer insight.

Analytical / Application Questions

1. A telecommunications provider receives complaints that customers find its billing system “confusing.” Design a VoC study and show how the resulting evidence could be translated into at least three CTQs.

2. A hospital survey shows high overall satisfaction, but online reviews frequently mention poor discharge communication. Explain how you would investigate the apparent contradiction and prioritize improvement.

3. Construct a brief QFD logic for an online examination-registration service. Identify four customer requirements, four technical characteristics, and one likely trade-off between technical characteristics.

22. Multiple Choice Questions with Answers

Question 1

Which statement best represents the progression from a customer voice to a CTQ?

A. “The app should be modern” becomes “use the newest software framework.”

B. “I do not want to wait” becomes “90% of customers are served within 15 minutes.”

C. “The service is expensive” becomes “increase the price.”

D. “I want help” becomes “hire more employees.”

Correct Answer: B. “I do not want to wait” becomes “90% of customers are served within 15 minutes.”

Question 2

What is the principal value of observation in a Voice of the Customer system?

A. It guarantees statistically representative findings.

B. It eliminates the need for customer interviews.

C. It reveals behaviour, workarounds, and contextual difficulties that customers may not report.

D. It measures satisfaction more accurately than every other method.

Correct Answer: C. It reveals behaviour, workarounds, and contextual difficulties that customers may not report.

Question 3

Why is a focus group generally unsuitable for estimating the percentage of customers holding a particular opinion?

A. Participants cannot discuss experiences in groups.

B. Group interaction and small, non-representative samples limit population estimation.

C. Moderators cannot ask follow-up questions.

D. Focus groups produce only quantitative data.

Correct Answer: B. Group interaction and small, non-representative samples limit population estimation.

Question 4

In affinity analysis, why should analysts distinguish customer needs from proposed solutions?

A. Solutions are always more reliable than customer statements.

B. Separating them prevents the organization from prematurely restricting the range of possible responses.

C. Needs cannot be grouped into themes.

D. Proposed solutions are required for all qualitative research.

Correct Answer: B. Separating them prevents the organization from prematurely restricting the range of possible responses.

Question 5

Which item is a HOW in a House of Quality for an online enrolment process?

A. Know whether the application is complete.

B. Receive help without repeating information.

C. Real-time application-status visibility.

D. Feel confident about enrolment.

Correct Answer: C. Real-time application-status visibility.

Question 6

What does the correlation roof of a House of Quality primarily reveal?

A. The demographic profile of respondents.

B. Relationships and trade-offs among technical characteristics.

C. The number of complaints received by channel.

D. The wording of interview questions.

Correct Answer: B. Relationships and trade-offs among technical characteristics.

Question 7

A survey finds that customers rate an attribute poorly, but interviews show that the attribute is unimportant to the relevant segment. What is the best interpretation?

A. The attribute must automatically become the highest improvement priority.

B. The rating should be ignored because interviews are always superior.

C. Performance scores should be considered together with importance and segment context.

D. The survey sample must be discarded without further analysis.

Correct Answer: C. Performance scores should be considered together with importance and segment context.

Question 8

Which practice best supports ethical use of automated social-listening analytics?

A. Treating automated sentiment scores as definitive decisions.

B. Using only comments with positive sentiment.

C. Validating automated classifications and reviewing context, bias, and privacy risks.

D. Collecting identifiable data indefinitely for future analysis.

Correct Answer: C. Validating automated classifications and reviewing context, bias, and privacy risks.

Question 9

A municipality assumes that buying more vehicles will solve missed waste collections. VoC analysis instead reveals uncertainty about collection windows and poor reporting follow-up. What does this illustrate?

A. Customer complaints always identify the correct technical solution.

B. The visible complaint may be a surface expression of broader customer needs.

C. Operational data should never be used with customer research.

D. QFD is useful only for manufactured products.

Correct Answer: B. The visible complaint may be a surface expression of broader customer needs.

Question 10

Which feature most clearly indicates a mature VoC system?

A. A large database of unanalysed comments.

B. A single annual satisfaction score used for all decisions.

C. Traceability from customer evidence to CTQs, owners, actions, and outcome measures.

D. Exclusive reliance on social-media monitoring.

Correct Answer: C. Traceability from customer evidence to CTQs, owners, actions, and outcome measures.