Marketing Research and Market Intelligence

Introduction

Hospitality organisations operate in markets characterised by perishability, fluctuating demand, intense competition, online transparency and highly variable guest expectations. A hotel room that remains unsold tonight cannot be inventoried for tomorrow; a negative review can influence prospective guests across several distribution channels; and a competitor can alter its price, package or positioning within hours. In this environment, managerial intuition remains useful, but it is not a sufficient basis for strategic decision-making.

Marketing research provides a systematic process for defining information problems, collecting evidence and converting evidence into managerial insight. Market intelligence extends this process by continuously monitoring customers, competitors, intermediaries, technologies, regulations and broader environmental developments. Together, they support decisions about segmentation, positioning, pricing, product design, communication, distribution, service quality and customer experience.

The central challenge is not simply to collect more data. It is to obtain relevant, reliable and ethically acquired evidence, interpret it within its context and communicate its implications clearly. A large database may contain little useful knowledge, while a small, carefully designed qualitative study may reveal a critical service failure. Postgraduate hospitality managers therefore need methodological competence as well as analytical judgement.

Learning Objectives

By the end of this chapter, the learner should be able to:

1. Distinguish marketing research, marketing intelligence, data, information and insight.
2. Formulate hospitality research problems and measurable research objectives.
3. Select appropriate qualitative, quantitative and mixed-method designs.
4. Design ethical sampling and survey processes.
5. Evaluate reliability, validity, bias and data quality.
6. Use social listening, online reviews, competitive intelligence and secondary data responsibly.
7. Interpret basic descriptive and inferential statistics for managerial decisions.
8. Integrate evidence into a market-intelligence system and communicate actionable recommendations.

1. Research, Intelligence and Evidence-Based Hospitality Management

1.1 Distinguishing related concepts

Data are recorded observations, such as room nights, booking dates, review text or responses to a satisfaction item. Information is organised data placed in a meaningful context, such as monthly occupancy by segment. Insight is an evidence-based interpretation that explains what is happening, why it matters and what action may be appropriate. Marketing research is the systematic design, collection, analysis and communication of information relevant to a specific marketing problem. Market intelligence is a continuing organisational capability that gathers and interprets information about the market, competitors, customers and environment.

Research is often project-based: a resort may investigate why repeat visits have declined. Intelligence is more continuous: the same resort may maintain a dashboard monitoring competitor prices, destination demand, review themes, air connectivity and source-market trends. Research answers a defined question; intelligence helps the organisation notice emerging questions.

1.2 The research-to-decision cycle

A sound process normally includes:

1. Decision situation: Identify the managerial decision, not merely a request for data.
2. Problem definition: Translate the situation into a precise research problem.
3. Objectives and questions: State what must be learned.
4. Design: Select methods, population, sample, measures and procedures.
5. Data collection: Implement the protocol consistently.
6. Data preparation and analysis: Check, code, clean and analyse evidence.
7. Interpretation: Relate findings to theory, context and decision alternatives.
8. Communication and action: Present implications, limitations and recommended actions.
9. Evaluation: Assess whether the decision and research produced value.

A manager who asks, “Can you survey our guests?” may actually need to decide whether to redesign breakfast, alter a loyalty benefit or target a different market. The research method should follow the decision problem rather than the convenience of a familiar technique.

1.3 A decision-oriented research brief

A professional research brief should identify the business background, decision to be supported, research problem, objectives, target population, proposed methods, timetable, budget, ethical safeguards, deliverables and limitations. Objectives should be specific and observable. “Understand satisfaction” is weak; “identify the three service attributes most strongly associated with intention to recommend among leisure guests during the last six months” is more useful.

2. Research Paradigms and Design Choices

2.1 Qualitative research

Qualitative research explores meanings, experiences, motives, language and processes. Common methods include in-depth interviews, focus groups, observation, ethnography, diary studies and analysis of guest narratives. It is particularly valuable when a market is unfamiliar, a service experience is complex or the organisation needs to understand why a pattern exists.

A semi-structured interview guide usually contains opening questions, core topics, probes and closing questions. Questions should be open, neutral and comprehensible. “What influenced your choice of this hotel?” is preferable to “You chose this hotel because of its location, didn’t you?” The researcher should probe for concrete episodes rather than accept general claims alone.

Qualitative analysis commonly involves familiarisation, coding, category development, theme construction, interpretation and comparison with alternative explanations. A theme is not merely a frequently repeated word; it is a meaningful pattern relevant to the research question. For example, “slow service” may conceal different experiences involving queueing, unclear responsibility and poor communication.

Quality is assessed through credibility, dependability, confirmability and transferability. Techniques include triangulation, reflexive journaling, searching for disconfirming evidence, maintaining an audit trail and using participant validation where appropriate. Small qualitative samples are not intended to estimate population percentages; their value lies in depth and conceptual understanding.

2.2 Quantitative research

Quantitative research measures variables numerically and examines distributions, relationships or differences. Surveys, experiments, transaction analysis and structured observation are common hospitality applications. Quantitative designs may be exploratory, descriptive, correlational or causal.

A variable is a characteristic that can take different values. Examples include booking lead time, perceived value, stay length and intention to revisit. An operational definition specifies how a construct will be measured. “Loyalty” might be operationalised through repeat-purchase intention, actual repeat stays, advocacy or a composite scale; these are related but not identical.

Descriptive statistics summarise observed data. For observations x₁, x₂, …, xₙ, the arithmetic mean is:

Here, x̄ is the sample mean, xᵢ is observation i, and n is the number of observations. The mean is appropriate for quantitative measurements when extreme values do not distort interpretation. The median, the middle ordered value, is often preferable for skewed variables such as hotel rates or spending.

The sample variance and standard deviation are commonly estimated as:

The denominator n − 1 is used when estimating population variability from a sample under the usual independent-sampling framework. Standard deviation has the same unit as the original variable; variance has squared units.

For a proportion p̂ estimated from a sufficiently large, approximately random sample, an approximate 95% confidence interval is:

The value 1.96 is the standard-normal critical value for a two-sided 95% interval. This approximation assumes an appropriate probability sample and adequate expected counts; it should not be treated as a remedy for coverage bias or non-response bias.

Worked example. In a random sample of 400 guests, 260 report that they are likely to return. Thus p̂ = 260 ÷ 400 = 0.65. The approximate standard error is √[0.65 × 0.35 ÷ 400] ≈ 0.0239. The interval is 0.65 ± 1.96 × 0.0239, or approximately 0.603 to 0.697. The manager may report that the estimated proportion is 65%, with an approximate 95% confidence interval from 60.3% to 69.7%, subject to the sampling assumptions.

Association does not establish causation. A correlation between review ratings and occupancy may reflect seasonality, property type, destination popularity or pricing. Causal claims require stronger designs, such as experiments, natural experiments or carefully justified quasi-experimental analysis.

2.3 Mixed-method research

Mixed-method research intentionally integrates qualitative and quantitative evidence. In an exploratory sequential design, interviews identify constructs and language, followed by a survey that measures their prevalence. In an explanatory sequential design, survey findings are followed by interviews to explain surprising results. In a convergent design, qualitative and quantitative data are collected during a similar period and compared.

Integration must occur at the level of research questions and interpretation, not merely by placing two unrelated reports together. If survey results show high overall satisfaction but interviews reveal anxiety about check-in privacy, the apparent contradiction should be investigated through segmentation, question wording and service context.

3. Sampling and Survey Design

3.1 The target population and sampling frame

The target population is the complete group to which conclusions are intended to apply, such as guests who stayed at a city hotel during a specified quarter. A sampling frame is the operational list or mechanism from which participants are selected, such as eligible booking records. Coverage error occurs when the frame excludes relevant members or includes ineligible cases.

Probability sampling gives known, non-zero selection probabilities. Simple random sampling selects units randomly from a complete frame. Systematic sampling selects every kth unit after a random start, where k is approximately N ÷ n, N is frame size and n is desired sample size. Stratified sampling divides the population into relevant subgroups, such as business and leisure guests, then samples within each. Cluster sampling selects groups, such as hotels or tour groups, and studies units within them; it may reduce cost but often increases sampling error because members of a cluster resemble one another.

Non-probability methods include convenience, purposive, quota and snowball sampling. They may be appropriate for exploratory research or hard-to-reach groups, but sampling error cannot be calculated in the same conventional way and generalisation must be cautious.

3.2 Sample size and precision

For a simple random estimate of a proportion, a commonly used planning formula is:

Here, n₀ is the initial sample size, z is the critical value for the desired confidence level, p is the anticipated population proportion and e is the maximum tolerated margin of error expressed as a proportion. If p is unknown, p = 0.50 gives the largest required sample under this model. The formula assumes a probability sample, independent observations and the intended confidence approximation.

If the population N is small relative to the desired sample, a finite-population correction may be applied:

These calculations address sampling precision, not non-response, poor measurement or coverage bias. Researchers may inflate the invitation count to allow for an expected response rate, but this does not guarantee representative responses.

3.3 Questionnaire construction

A questionnaire should follow a logical sequence: eligibility, easy introductory questions, core measures, behavioural questions, sensitive questions and demographics, unless another order reduces bias. Each question should address one idea, use a defined reference period and provide balanced response options.

Avoid double-barrelled items such as “How satisfied were you with the room and breakfast?” Avoid leading questions, ambiguous frequency terms and unnecessarily technical language. Rating scales should have clearly labelled endpoints and, when appropriate, a neutral or “not applicable” option. A five- or seven-point agreement scale may be useful, but the scale format does not itself guarantee interval-level measurement.

Pretesting should assess comprehension, recall difficulty, response burden, routing, mobile display and interpretation. Cognitive interviewing asks participants to explain how they understood and answered each question. Pilot data can reveal missing categories, completion time and technical problems.

3.4 Reliability and validity

Reliability concerns consistency. Internal consistency may be examined for a multi-item scale, but a high coefficient does not prove that the scale measures the intended construct. Test–retest reliability concerns stability over time, while inter-rater reliability concerns agreement between observers or coders.

Validity concerns whether the evidence supports the intended interpretation. Content validity asks whether important aspects are represented. Construct validity considers convergent and discriminant evidence. Criterion-related validity examines association with a relevant outcome. External validity concerns transferability or generalisation to other settings.

Common threats include social desirability, acquiescence, recall error, interviewer effects, common-method bias, non-response and survivorship bias. Neutral wording, confidentiality assurances, multiple recruitment channels, short questionnaires and comparison of respondents with known population characteristics can reduce—but not eliminate—these threats.

4. Secondary Data and Hospitality Information Sources

Secondary data were originally collected for another purpose. Internal sources include property-management systems, central-reservation records, point-of-sale data, loyalty databases, customer relationship management systems, website analytics, cancellation records and complaint logs. External sources include national statistics, tourism boards, destination organisations, airline schedules, industry reports, trade publications, economic indicators, competitor websites and online reviews.

Before using secondary data, assess:

Data integration requires consistent identifiers, dates, currencies, time zones and definitions. “Revenue,” for example, may mean gross room revenue, net revenue after commissions or total property revenue. Analysts should document transformations and preserve an untouched source copy.

5. Social Listening, Online Reviews and Digital Ethnography

Social listening monitors publicly available conversations to identify topics, sentiment, questions, complaints and emerging expectations. It may include review platforms, social networks, travel communities, blogs and public video comments. It is not equivalent to a representative survey: users are self-selected, vocal experiences may be overrepresented, and platform algorithms influence visibility.

A responsible workflow is to define the decision and search terms, establish a time period, collect permitted data, remove duplicates and spam, code topics and sentiment, compare channels and validate interpretations with human review. Automated sentiment classification can misread sarcasm, multilingual expressions, cultural references and context. Sentiment should therefore be treated as an analytical aid rather than an unquestionable fact.

Review analysis should move beyond an average star rating. Managers should examine review volume, recency, response time, recurring themes, guest segment, stay context and the relationship between positive and negative attributes. A low rating about an unavailable facility is different from a low rating about staff discourtesy, even if both reduce the average.

Ethical practice includes respecting platform terms, avoiding unnecessary collection of personal information, anonymising quotations, not attempting to identify private individuals, and giving special care to vulnerable guests. Public availability does not automatically mean unrestricted ethical use.

6. Competitive Intelligence

Competitive intelligence is the systematic collection and analysis of lawful, publicly available or legitimately obtained information about competitors and market forces. It differs from industrial espionage, deception or misuse of confidential information.

A hospitality intelligence programme may monitor competitor room types, public rates, restrictions, packages, loyalty benefits, distribution presence, facilities, service promises, review themes, staffing signals and major partnerships. Comparison should account for location, property class, date, occupancy conditions, taxes, cancellation rules, inclusions and channel commissions. A rate observed on a single date is not a complete measure of competitive positioning.

A useful competitor matrix combines observable facts with cautious interpretation:

Intelligence should be triangulated. A competitor’s website reflects intended positioning; reviews reflect selected experiences; observed prices reflect a particular moment. None alone proves the competitor’s actual performance.

7. Market-Intelligence Architecture and Decision Use

An effective intelligence system has four layers: data sources, governance and integration, analytical methods, and decision routines. A dashboard may combine occupancy, average daily rate, channel contribution, review themes, search interest, competitor rates and market events. However, dashboards can create false confidence if definitions differ or if users confuse correlation with causation.

Useful analytical questions include:

• Which segments are growing, declining or becoming less profitable?
• Which attributes distinguish high-value guests from low-value guests?
• What service failures recur across complaints and reviews?
• Are competitor actions changing booking behaviour, or merely coinciding with it?
• Which recommendation is supported by evidence, and what uncertainty remains?

The net promoter score is sometimes calculated as the percentage of promoters minus the percentage of detractors, based on a specified recommendation question and classification rule. It is a descriptive management indicator, not a complete measure of loyalty or profitability. It should not replace behavioural outcomes, satisfaction diagnosis or financial analysis.

Research findings should be translated into decisions using an evidence-to-action chain: finding, interpretation, implication, recommended action, responsible owner, time frame and success metric. For example, “late check-in complaints increased” is a finding; “arrival congestion is concentrated among weekend leisure arrivals between 16:00 and 18:00” is a more useful interpretation; “introduce timed pre-arrival registration and add one peak-period agent” is an actionable response requiring a subsequent evaluation.

8. Ethics, Privacy and Governance

Hospitality research involves personal data, including contact details, booking histories, location information, preferences, photographs and potentially sensitive complaints. Ethical research requires respect for autonomy, informed consent where applicable, minimisation of harm, confidentiality, fair treatment and responsible data stewardship.

Researchers should specify the purpose of collection, collect only necessary data, restrict access, establish retention periods and protect data during transfer and storage. Consent should be understandable and distinguish participation in research from receiving marketing communications. Participants should not be pressured by their dependence on the hotel, and incentives should not be coercive.

Research involving employees requires particular care because organisational power differences may inhibit honest responses. Anonymous collection, independent administration and clear separation between research and performance appraisal can reduce fear. If recordings, biometric information, location tracking or automated profiling are used, the legal basis, proportionality and governance requirements must be examined under applicable data-protection and research-ethics rules.

Artificial intelligence can assist with transcription, coding, translation, anomaly detection and summarisation. It can also reproduce bias, expose confidential data, hallucinate themes or obscure the reasoning behind a recommendation. Organisations should establish human review, approved tools, access controls, documentation, bias testing and disclosure standards. Generated analysis must be traceable to source evidence.

9. Illustrative Case: Harbourview City Hotel

Background

Harbourview City Hotel, a fictional 220-room business-and-leisure property, recorded stable occupancy but a decline in direct bookings and repeat-stay intention. Its general manager believed that an OTA was undercutting the hotel, while the marketing manager suspected that the hotel’s mobile booking journey was confusing.

Problem and approach

The research team defined the decision as whether to prioritise price intervention, website redesign or service changes. It used a mixed-method design:

1. Analysis of twelve months of booking, cancellation and channel data.
2. A review audit of 1,200 public reviews, sampled across four major platforms.
3. Twenty-four semi-structured interviews with recent guests, selected across business, leisure, domestic and international segments.
4. A structured survey of recent guests, with 420 usable responses.
5. A lawful weekly comparison of publicly displayed competitor rates, including cancellation conditions and inclusions.

The team pretested the survey, separated research invitations from promotional messages, removed identifying information from quotations and documented the limitations of self-selection.

Analysis

Transaction data showed that direct bookings were not universally less expensive. However, mobile users abandoned the booking process at a much higher rate than desktop users. Review coding identified recurring positive themes about staff warmth and location, but negative themes about unclear breakfast inclusion, delayed responses to digital enquiries and inconsistent room-readiness communication. Interviews explained that guests did not necessarily perceive the problem as “price”; they perceived uncertainty about what the price included.

The survey found that perceived booking clarity was more strongly associated with direct-booking intention than the stated room rate, although the cross-sectional design could not establish causation. Segment analysis showed that international leisure guests were particularly concerned about cancellation language and transport information.

Solution and outcome

Harbourview redesigned the mobile booking path, displayed inclusions and cancellation terms in plain language, created a pre-arrival message sequence and introduced a response-time standard for digital enquiries. It did not immediately reduce public rates. After a controlled three-month evaluation, mobile completion, direct-booking share and enquiry response time improved. Because several changes occurred together and the evaluation period was short, management appropriately avoided claiming that the redesign alone caused every improvement.

Learning points

The case demonstrates the value of defining the decision before selecting methods, triangulating behavioural, attitudinal and textual data, distinguishing price from perceived value and communicating causal limitations. It also shows that intelligence is most useful when it changes a specific managerial action rather than merely producing a dashboard.

10. Challenges and Best Practices

Research quality can be weakened by unclear objectives, unrepresentative samples, leading questions, excessive survey length, poor coding, unexamined platform bias, confusing association with causation and presenting estimates without uncertainty. Organisational politics may encourage selective reporting, while confirmation bias can lead managers to search for evidence supporting a preferred decision.

Best practice includes:

• Begin with the decision and define the population precisely.
• Use qualitative work to understand language and mechanisms before measuring prevalence.
• Match sampling logic to the intended claim.
• Pretest instruments and document all changes.
• Triangulate primary, secondary, behavioural and textual evidence.
• Report missing data, response rates, exclusions and limitations.
• Separate observed facts from interpretation and recommendation.
• Protect privacy throughout the data life cycle.
• Treat automated analytics as assistive, not authoritative.
• Evaluate whether the decision produced the expected operational and customer outcomes.

Chapter Summary

Marketing research is a systematic process for answering a defined marketing problem, whereas market intelligence is a continuing capability for monitoring customers, competitors and the environment. Effective hospitality research begins with a decision-oriented brief and uses qualitative, quantitative or mixed methods according to the nature of the question. Qualitative methods reveal meanings and mechanisms; quantitative methods estimate prevalence, differences and relationships; mixed methods integrate both forms of evidence.

Sampling quality depends on the target population, sampling frame, selection method and response process. Survey quality requires clear operational definitions, neutral wording, pretesting, reliability assessment and validity evidence. Descriptive statistics and confidence intervals are useful only when their assumptions and limitations are understood. Secondary data, social listening, online reviews and competitive intelligence can enrich primary research, but each contains coverage, selection, measurement and ethical risks.

A mature intelligence system integrates lawful data sources, governance, analysis and decision routines. It reports uncertainty, avoids unsupported causal claims and converts findings into accountable actions. Privacy, consent, confidentiality, fairness and human oversight are essential, especially when personal data and artificial intelligence are involved.

Key Terms

Data: Recorded observations collected for analysis.

Information: Organised data interpreted within a context.

Insight: An evidence-based interpretation with decision relevance.

Marketing research: Systematic collection and analysis of information for a defined marketing problem.

Market intelligence: Continuous monitoring and interpretation of market and environmental information.

Target population: The complete group to which research conclusions are intended to apply.

Sampling frame: The operational list or mechanism from which a sample is selected.

Probability sampling: Sampling in which units have known, non-zero selection probabilities.

Non-probability sampling: Sampling without known selection probabilities for all population units.

Operational definition: The specification of how an abstract construct will be measured.

Reliability: Consistency of a measure or research procedure.

Validity: The degree to which evidence supports the intended interpretation.

Triangulation: Comparison of evidence from different methods, sources or investigators.

Social listening: Systematic monitoring and interpretation of public digital conversations.

Competitive intelligence: Lawful collection and analysis of information about competitors and market forces.

Coverage error: Bias caused by mismatch between the sampling frame and target population.

Non-response bias: Bias arising when respondents differ meaningfully from non-respondents.

Secondary data: Data originally collected for a purpose other than the current study.

Mixed-method research: Intentional integration of qualitative and quantitative approaches.

Data governance: Policies and practices controlling data quality, access, security, use and retention.

Review Questions

Short Answer Questions

1. Distinguish marketing research from market intelligence.
2. What is the difference between a target population and a sampling frame?
3. State two advantages and two limitations of online-review analysis.
4. Why is pretesting important in survey design?
5. Explain why correlation does not establish causation.

Long Answer Questions

1. Explain the stages of a decision-oriented hospitality marketing research process, using a hotel example.
2. Compare qualitative, quantitative and mixed-method research designs and identify suitable hospitality applications for each.
3. Discuss probability and non-probability sampling, including their implications for generalisation.
4. Explain how reliability, validity, coverage error and non-response bias affect survey conclusions.
5. Design an ethical market-intelligence system for a resort operating in a competitive destination.

Analytical / Application Questions

1. A hotel’s online rating has fallen from 4.5 to 4.2, while occupancy remains stable. Design a research programme that determines whether the decline reflects a serious service problem, review-platform selection effects or a change in guest mix.
2. A survey of 400 randomly selected guests finds that 260 intend to return. Calculate an approximate 95% confidence interval for the population proportion and explain the assumptions and managerial limitations.
3. A competitor introduces a lower advertised rate, but its offer excludes breakfast and has stricter cancellation terms. Explain how you would compare the offers fairly and recommend an appropriate managerial response.

Multiple Choice Questions with Answers

Question 1

Which statement best distinguishes market intelligence from a single marketing research project?

A. Market intelligence is continuous monitoring, whereas research is often designed for a specific problem.

B. Market intelligence uses only primary data, whereas research uses only secondary data.

C. Market intelligence is limited to financial information.

D. Market intelligence does not require interpretation.

Correct Answer: A. Market intelligence is continuous monitoring, whereas research is often designed for a specific problem.

Question 2

Which method is most appropriate for exploring why guests interpret a booking process as confusing?

A. A structured occupancy report

B. Semi-structured guest interviews

C. A competitor rate scrape alone

D. A frequency table of room nights

Correct Answer: B. Semi-structured guest interviews

Question 3

A hotel surveys 400 guests and 260 report an intention to return. What is the sample proportion?

A. 0.40

B. 0.60

C. 0.65

D. 1.54

Correct Answer: C. 0.65

Question 4

Which issue is an example of coverage error?

A. A respondent selects the wrong answer accidentally

B. The sampling frame excludes guests who booked through an unlisted channel

C. An analyst rounds a percentage to one decimal place

D. An interviewer records an answer twice

Correct Answer: B. The sampling frame excludes guests who booked through an unlisted channel

Question 5

What is the principal purpose of cognitive interviewing during questionnaire development?

A. To increase the size of the sample

B. To test how participants understand and answer questions

C. To calculate competitor market share

D. To replace all quantitative analysis

Correct Answer: B. To test how participants understand and answer questions

Question 6

Which statement about online reviews is most defensible?

A. They provide a perfectly representative sample of all guests.

B. Their average rating automatically proves service quality.

C. They can reveal recurring themes but are affected by self-selection and platform effects.

D. They should never be used in hospitality research.

Correct Answer: C. They can reveal recurring themes but are affected by self-selection and platform effects.

Question 7

In competitive intelligence, why should cancellation conditions and inclusions be compared with public rates?

A. Because headline prices may not represent equivalent value propositions.

B. Because competitor prices are always illegal to observe.

C. Because room rates never change over time.

D. Because reviews cannot contain useful information.

Correct Answer: A. Because headline prices may not represent equivalent value propositions.

Question 8

For a simple random sample, what does a confidence interval primarily express?

A. A guarantee that the estimate is correct

B. Uncertainty associated with sampling under stated assumptions

C. The percentage of non-respondents

D. The causal effect of a marketing campaign

Correct Answer: B. Uncertainty associated with sampling under stated assumptions

Question 9

Which practice best supports ethical hospitality research?

A. Collecting every available personal attribute regardless of purpose

B. Using public data without considering platform terms or privacy

C. Explaining the research purpose and collecting only necessary data

D. Linking research responses to employee evaluations without disclosure

Correct Answer: C. Explaining the research purpose and collecting only necessary data

Question 10

What is a major risk of using artificial intelligence to classify guest-review sentiment?

A. It can never process text at scale.

B. It may misinterpret sarcasm, context or multilingual expressions.

C. It eliminates the need for human judgement in every case.

D. It prevents all forms of data storage.

Correct Answer: B. It may misinterpret sarcasm, context or multilingual expressions.