13. Tourism Research, Analytics and Managerial Decision-Making

Introduction and Context

Tourism is an information-intensive industry. Hotels, airlines, attractions, destination management organisations, travel intermediaries and public authorities make decisions about markets, prices, capacity, investment and sustainability on the basis of evidence. Yet tourism evidence is rarely simple. Visitor numbers may rise while local economic benefits decline; hotel occupancy may increase because of a single event; social-media sentiment may appear positive while resident opposition intensifies; and an apparently successful destination campaign may merely have reached people who were already planning to travel.

Tourism research and analytics provide methods for addressing such complexity. Research generates systematic knowledge about tourists, enterprises, destinations and their environments. Analytics converts structured and unstructured data into patterns, forecasts and indicators. Managerial decision-making uses this evidence, together with professional judgement, institutional values and stakeholder perspectives, to select and evaluate courses of action.

The purpose of this chapter is not to treat data as an automatic substitute for judgement. Rather, it is to show how high-quality evidence can improve judgement. The chapter brings together research design, measurement, tourism indicators, dashboards, geographic information systems (GIS), big-data analysis, benchmarking, forecasting and evaluation. It also examines the ethical and governance issues that arise when tourism organisations collect and use personal, behavioural and location data.

Learning Objectives

After studying this chapter, students should be able to:

1. distinguish between exploratory, descriptive, explanatory and evaluative tourism research;

2. formulate research questions and select appropriate qualitative, quantitative and mixed-method designs;

3. assess the validity, reliability and practical usefulness of tourism indicators;

4. interpret dashboards, GIS outputs, big-data sources and benchmarking results;

5. apply basic forecasting and scenario techniques to tourism decisions;

6. design evidence-based evaluation for tourism policies, services and campaigns;

7. identify analytical bias, data-quality problems, privacy risks and governance requirements; and

8. translate research findings into defensible managerial and destination decisions.

From Tourism Questions to Researchable Problems

A tourism decision begins with a problem, not with a data set. A hotel may ask why weekend demand is weakening. A destination may ask whether a festival generates additional visitors or simply shifts the timing of existing trips. A government may ask whether tourism growth is compatible with water availability and resident well-being. These questions must be converted into researchable problems with defined concepts, populations, time periods and decision criteria.

Four broad purposes of research are particularly useful. Exploratory research investigates a poorly understood issue and is often based on interviews, observation, focus groups, document analysis or social-media analysis. Descriptive research measures characteristics such as visitor profiles, length of stay, expenditure, satisfaction or occupancy. Explanatory research examines relationships and possible causal mechanisms, for example whether perceived value influences revisit intention. Evaluative research assesses the effects, efficiency or equity of a programme, policy or investment.

A sound research question is specific enough to investigate and important enough to influence action. “How can the destination improve tourism?” is too broad. “Which visitor segments are most likely to extend their stay when offered an integrated cultural itinerary during the shoulder season?” is more useful because it identifies a market, an intervention and an outcome.

The unit of analysis must also be made explicit. It may be an individual visitor, booking, hotel, attraction, neighbourhood, event, trip or destination. Confusing these levels creates ecological and atomistic fallacies. For example, a relationship observed between destination-level promotion expenditure and visitor arrivals does not prove that every individual exposed to the promotion decided to travel because of it.

Research Design and Evidence Quality

Qualitative, quantitative and mixed-method approaches

Quantitative research uses numerical measurement and statistical analysis. Surveys, transaction records, occupancy data and visitor counts can establish patterns across relatively large populations. Its strength lies in comparability and estimation, but numerical precision does not guarantee conceptual accuracy. A large survey measuring an ambiguous concept may be less useful than a smaller study with a well-designed instrument.

Qualitative research explores meanings, experiences, practices and institutional processes. Interviews with residents may reveal concerns about noise, displacement or cultural representation that are not captured by a satisfaction scale. Participant observation can show how visitors actually move through a heritage site rather than how they report moving. Qualitative findings are not simply “anecdotal”; when systematically collected and analysed, they provide contextual and interpretive evidence.

Mixed-method research combines both approaches. A sequential exploratory design may begin with interviews to identify dimensions of destination image, followed by a survey to test their prevalence. A sequential explanatory design may first identify an unexpected statistical pattern and then use interviews to understand it. Integration is essential: the two components should answer connected questions rather than operate as unrelated studies.

Sampling and measurement

A probability sample gives units a known, non-zero chance of selection and supports statistical generalisation when appropriately designed. Convenience samples are faster but vulnerable to selection bias. Online visitor surveys may overrepresent digitally engaged travellers, while intercept surveys at airports may exclude visitors departing by land. Sampling frames should therefore be examined rather than assumed to represent the target population.

Measurement involves operationalising abstract concepts. “Satisfaction” may be measured through overall evaluation, attribute ratings, recommendation intention or behavioural indicators such as repeat visitation. A good measure is valid when it captures the intended construct and reliable when it produces consistent results under comparable conditions. Reliability alone is insufficient: a consistently inaccurate measure remains unsuitable.

Researchers should pre-test questionnaires, use clear response scales, avoid double-barrelled questions and distinguish “not applicable” from neutral responses. Non-response should be analysed because those who refuse or fail to respond may differ systematically from respondents. Weighting can correct some known differences, but it cannot repair all forms of bias.

Causality and evaluation design

Managers frequently want to know whether an action caused an outcome. Simple before-and-after comparisons are often inadequate because tourism outcomes are affected by seasonality, economic conditions, weather, competition and unforeseen events. Stronger designs include control groups, matched comparisons, difference-in-differences analysis, interrupted time series and, where feasible, randomised experiments.

For example, if one group of hotels receives a digital marketing intervention and comparable hotels do not, changes in direct bookings can be compared between the two groups. The design remains vulnerable if the groups differ in important ways, but it is more informative than comparing the intervention group only with its own previous performance. Evaluation should define the theory of change: inputs lead to activities, activities produce outputs, outputs influence outcomes, and outcomes may contribute to longer-term impacts.

Tourism Indicators: Measuring What Matters

An indicator is a selected measure used to describe a condition, change or performance. Common tourism indicators include arrivals, overnight stays, occupancy, average daily rate, revenue per available room, average length of stay, visitor expenditure, employment, carbon emissions, water use, resident attitudes and congestion levels.

Indicators should be interpreted as parts of a measurement system rather than isolated targets. Arrivals measure volume but not necessarily value. Occupancy reflects the use of accommodation capacity but may increase through discounting. Revenue per available room combines price and occupancy, but does not represent profit. Expenditure estimates may omit informal transactions or spending outside the destination. Sustainability indicators should therefore accompany economic indicators.

Useful indicator systems normally observe several dimensions:

• Demand: arrivals, bookings, search interest, length of stay and seasonality;
• Economic value: expenditure, employment, tax contribution, local procurement and profitability;
• Experience: satisfaction, complaints, recommendation and repeat intention;
• Community well-being: resident attitudes, housing pressure, access to public space and perceived cultural effects;
• Environmental pressure: emissions, waste, water consumption, energy use and site carrying conditions; and
• Resilience: source-market diversity, recovery time, dependence on particular segments and operational adaptability.

A target can distort behaviour when it becomes a narrow performance objective. If a destination rewards only increases in arrivals, organisations may pursue high-volume, low-yield markets and ignore congestion. A balanced scorecard approach links financial, customer, internal-process, learning and sustainability perspectives. Indicators should be disaggregated by season, market, geography and visitor type where privacy and statistical reliability permit.

Dashboards and the Management of Information

A dashboard is a visual interface that presents selected indicators for monitoring and decision support. An effective tourism dashboard answers three questions: What is happening? Why might it be happening? What decision requires attention? It should not merely display every available measure.

Dashboard design begins with the decision cycle. Strategic dashboards may show long-term market diversification, investment and sustainability trends. Tactical dashboards may monitor campaign performance, booking pace, event demand or staffing requirements. Operational dashboards may support daily room allocation, queue management and service recovery. Each audience requires different levels of aggregation, refresh frequency and access control.

Good dashboards display definitions, data periods, sources, update times and quality warnings. They distinguish actual values from forecasts and avoid misleading visual scales. A sudden increase in online searches, for instance, may indicate interest but not actual travel. A dashboard should permit users to examine underlying data and segment results, while preventing unauthorised access to personally identifiable information.

The dashboard is a decision-support tool, not a decision-maker. Managers should ask whether a movement is statistically meaningful, operationally important or merely normal variation. Alerts should be linked to pre-agreed actions, such as investigating a fall in conversion, adjusting staffing or initiating a resident consultation.

GIS and the Spatial Analysis of Tourism

Tourism activity is spatial. Attractions, accommodation, transport networks, protected areas, residential communities and commercial districts interact in particular places. GIS combines maps with attribute data to analyse these relationships.

Basic GIS functions include mapping the distribution of hotels, measuring distance to transport nodes, identifying visitor concentrations and comparing demand with infrastructure. More advanced applications include spatial clustering, network analysis, accessibility modelling, viewshed analysis and overlay analysis. A destination may overlay visitor heat maps with resident population density, public-transport capacity and sensitive environmental areas to identify pressure zones.

Spatial analysis improves decisions about wayfinding, shuttle routes, new attractions, emergency response and visitor dispersal. It can reveal that a destination is not uniformly congested: pressure may be concentrated around a small number of streets or time periods. However, location data must be interpreted carefully. A mobile-phone signal may indicate presence without revealing purpose, spending or demographic identity. Spatial aggregation can also hide inequalities within neighbourhoods.

GIS should therefore be combined with field observation, stakeholder knowledge and qualitative evidence. Maps have persuasive visual power, but visual clarity can create false confidence when boundaries, missing data or positional uncertainty are not disclosed.

Big Data, Artificial Intelligence and Analytical Risk

Tourism big data may originate from booking systems, point-of-sale transactions, loyalty programmes, online reviews, web searches, social platforms, mobile devices, transport systems, sensors and Internet-of-Things equipment. These sources offer high volume, velocity and variety, often enabling near-real-time analysis.

Text analytics and natural-language processing can classify review topics and sentiment. Recommendation systems can personalise offers. Machine-learning models can predict cancellation, demand or crowding. Computer vision may estimate queues or occupancy in public areas. These applications can improve responsiveness, but they do not eliminate the need for research design.

Big data is often comprehensive in volume but selective in coverage. Online reviews are produced by a self-selected group. Search data reflect interest, not necessarily booking. Social-media language differs across markets and cultures. Algorithmic classifications may misread sarcasm, multilingual expressions or local meanings. Historical data can reproduce past discrimination, while opaque models make it difficult to explain decisions to employees, customers or regulators.

A responsible analytical workflow includes data provenance, validation against independent sources, documentation of variables, bias testing, human review and monitoring after deployment. Models should be judged not only by predictive accuracy but also by interpretability, fairness, cost and consequences of error. A demand forecast that is slightly less accurate but transparent and stable may be more useful than a complex model that managers cannot interrogate.

Benchmarking and Comparative Performance

Benchmarking compares performance with a reference point. Internal benchmarking compares departments, properties or periods within the same organisation. Competitive benchmarking compares similar enterprises. Functional benchmarking compares a process with organisations outside the immediate sector. Destination benchmarking may compare indicators across places, although differences in geography, accounting, market structure and policy must be recognised.

A meaningful benchmark requires comparable definitions. Comparing hotel occupancy without considering room supply, market segment, season and event effects can be misleading. Benchmarking should combine absolute performance with relative position and trend. A property may outperform its competitors during a weak market yet still have declining profitability.

Managers should distinguish descriptive benchmarking from causal explanation. A competitor’s higher direct-booking share may result from brand strength, distribution contracts, loyalty investment or a different customer mix. Benchmarking identifies a performance gap; further research is needed to explain and address it.

Forecasting and Scenario-Based Decision-Making

Forecasting estimates future conditions using historical data, explanatory variables, expert judgement or combinations of these. Tourism forecasts may concern arrivals, room demand, prices, staffing, transport flows or expenditure. Time-series methods use patterns such as trend and seasonality. Causal models incorporate variables such as income, exchange rates, air capacity, events and marketing. Qualitative approaches, including expert panels and structured scenarios, are valuable where historical relationships are unstable.

Forecast accuracy should be evaluated using a holdout period rather than only the data used to build the model. Mean absolute error and related measures help compare forecasts, but the appropriate metric depends on the decision. Forecast intervals are as important as point estimates because uncertainty affects capacity and investment choices.

Scenarios are not predictions. They are coherent accounts of alternative futures, such as rapid recovery, prolonged inflation, climate disruption or transport restrictions. Scenario planning helps managers test strategies against uncertainty. A robust strategy may perform reasonably well across several futures rather than maximising returns under one preferred forecast.

In practice, forecasting should be connected to decisions. A hotel can translate demand scenarios into staffing bands, inventory controls and pricing rules. A destination can connect visitor-flow forecasts with public-transport capacity and environmental thresholds. Forecasts should be updated as new evidence arrives and should record assumptions so that errors can be learned from rather than concealed.

Evidence-Based Managerial Decision-Making

Evidence-based management integrates the best available research evidence, organisational data, professional expertise and the interests of affected stakeholders. It is not the same as data-driven management, in which numerical indicators may dominate other forms of knowledge. A decision can be evidence-informed even when evidence is incomplete, provided uncertainty is acknowledged.

A practical decision process contains five linked stages. First, define the decision and its constraints. Second, assemble and assess relevant evidence. Third, generate alternatives rather than jumping to a single solution. Fourth, compare alternatives using explicit criteria such as financial return, guest experience, feasibility, equity and environmental impact. Fifth, implement, monitor and revise.

Decision matrices, cost-benefit analysis, multicriteria analysis and real-options thinking can support this process. Cost-benefit analysis monetises consequences where possible, while multicriteria analysis is useful when social, cultural and environmental values cannot reasonably be reduced to money. Real-options thinking recognises the value of delaying, piloting or staging an investment when uncertainty is high.

The quality of a decision depends partly on the quality of the question. Managers should challenge confirmation bias, distinguish correlation from causation, seek disconfirming evidence and identify who benefits or bears the cost. A technically sophisticated analysis can still produce a poor decision if it ignores implementation capacity or stakeholder legitimacy.

Illustrative Case: Evidence-Led Visitor Management in Harbourview

Background

Harbourview is an illustrative coastal destination with a historic waterfront, a cruise terminal and a growing short-break market. Local businesses welcomed tourism income, but residents reported congestion, noise and pressure on public space during summer weekends. The destination organisation proposed a campaign to promote lesser-known inland attractions and considered introducing timed access to the historic waterfront.

Situation and problem

Initial evidence was contradictory. Annual visitor arrivals had increased, but average length of stay had fallen. Social-media posts showed strong interest in the waterfront, while a resident survey indicated declining support for tourism. The organisation needed to decide whether to spend its limited budget on promotion, visitor-management infrastructure or both.

Analysis

The research team defined three outcomes: economic value, spatial pressure and resident support. It integrated accommodation statistics, anonymised mobile-movement aggregates, transport data, attraction ticketing, visitor surveys, resident interviews and environmental observations. GIS showed that congestion was concentrated within a small waterfront zone between 11 a.m. and 3 p.m. on cruise-arrival days. Interviews revealed that residents were less concerned about total visitors than about simultaneous crowding and blocked pedestrian routes.

A simple comparison of cruise and non-cruise days showed lower average local spending per visitor on cruise days, but the team avoided claiming that cruise tourism was inherently harmful. Weather, ship size and visitor itinerary were potential confounding factors. A pilot intervention was therefore designed rather than a permanent policy introduced immediately.

Approach and solution

For one season, the destination tested three measures: coordinated arrival information encouraging visits to inland attractions, timed entry at the most sensitive heritage point, and a shuttle connecting the waterfront with inland sites. A matched set of comparable days was used for evaluation. Indicators included crowd density at peak periods, visitor expenditure distribution, resident attitudes, complaints, shuttle use, heritage-site revenue and visitor satisfaction.

The dashboard separated outputs from outcomes. The number of promotional impressions and shuttle passengers were outputs; reduced peak density, improved spending distribution and resident acceptance were outcomes. Qualitative interviews were repeated after the pilot to identify unintended effects.

Outcome

The pilot produced a more nuanced basis for decision-making than the original arrival statistics. Peak density declined in the monitored zone, some inland attractions experienced increased visitation, and resident attitudes improved modestly. However, the shuttle required subsidy and some visitors disliked timed access. The organisation decided to retain the shuttle on selected high-pressure days, improve advance communication, and use timed entry only when forecast pressure exceeded a defined threshold.

Learning points

The case illustrates five principles. First, the decision was framed around multiple outcomes rather than arrivals alone. Second, spatial and temporal disaggregation revealed the actual pressure point. Third, mixed methods explained why the problem occurred. Fourth, a pilot reduced the risk of a costly permanent intervention. Finally, evidence supported an adaptive decision rather than a simplistic judgement that tourism growth was either wholly beneficial or wholly harmful.

Challenges, Ethics and Governance

Tourism research faces non-response, inaccurate records, inconsistent definitions, small samples, missing data, seasonal distortion and rapidly changing behaviour. Analysts may engage in data dredging, selective reporting or “dashboard theatre,” in which attractive visualisations create an appearance of control without improving decisions. Forecasts can be treated as commitments, and benchmarks can encourage imitation without contextual understanding.

Ethically, researchers must obtain informed consent where required, minimise data collection, protect confidentiality and explain how information will be used. Personal data should be collected for legitimate purposes, retained only as long as necessary and protected through access controls and security procedures. Anonymisation and aggregation reduce risk but do not always eliminate the possibility of re-identification, especially when multiple data sets are combined.

Governance should assign responsibility for data ownership, quality, access, correction and deletion. Organisations should document model assumptions, conduct impact assessments for high-risk analytics and provide routes for human review or appeal. Particular care is needed when analytics affect prices, employment, access to services or the treatment of particular visitor groups. Resident data should not be extracted without reciprocal value, consultation and respect for local rights.

Current Developments and Best Practice

Tourism analytics is moving towards integrated data platforms, real-time mobility analysis, digital twins, low-carbon measurement and AI-assisted decision support. Digital twins may allow managers to simulate visitor flows or transport changes before implementation. Generative AI can assist with coding, text classification and reporting, but its outputs require verification, especially when used for research interpretation. Climate-related analytics is also becoming more important as destinations assess heat, water stress, extreme weather and adaptation capacity.

Best practice can be summarised as follows:

• begin with a decision and research question, not an attractive data source;
• define indicators, units, time periods and responsibilities clearly;
• combine quantitative patterns with qualitative explanation;
• test data quality, representativeness and measurement validity;
• distinguish monitoring, prediction and causal evaluation;
• show uncertainty, assumptions and data limitations;
• protect privacy and govern access throughout the data life cycle;
• pilot interventions where feasible and evaluate unintended effects;
• involve residents, employees and other affected stakeholders; and
• create feedback loops so that decisions improve the data system and the data system improves future decisions.

Chapter Summary

Tourism research and analytics support decisions about demand, value, experience, community well-being, environmental pressure and resilience. Effective work begins by defining a decision and converting it into a researchable question. Qualitative, quantitative and mixed-method designs each provide different forms of evidence, while sampling and measurement determine whether findings are credible.

Tourism indicators must be interpreted as a balanced system rather than isolated targets. Dashboards organise evidence for strategic, tactical and operational use; GIS reveals spatial and temporal patterns; big-data and AI techniques enable prediction and personalisation but introduce bias, privacy and explainability risks. Benchmarking identifies comparative performance gaps, while forecasting and scenarios help organisations prepare for uncertain futures.

Evidence-based management combines research, organisational data, professional judgement and stakeholder interests. Evaluation designs, pilots and explicit decision criteria help distinguish genuine effects from coincidence. Ethical governance, transparency, data security and respect for residents are essential to legitimate tourism analytics. The strongest tourism decisions are therefore not those supported by the largest data sets, but those that use relevant, valid and responsibly governed evidence to guide adaptive action.

Key Terms

Analytics: The systematic examination of data to identify patterns, generate insight and support decisions.

Benchmarking: Comparing performance with a defined internal, competitive or external reference point.

Big data: High-volume, high-velocity or high-variety data generated through digital and operational systems.

Causal evaluation: Assessment of whether an intervention produced an observed outcome.

Dashboard: A visual interface presenting selected indicators for monitoring and decision support.

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

Difference-in-differences: An evaluation method comparing changes over time between an intervention group and a comparison group.

GIS: Geographic information system used to store, analyse and display spatially referenced information.

Indicator: A selected measure describing a condition, change or performance.

Mixed-method research: Research integrating qualitative and quantitative evidence within one study.

Operationalisation: Translating an abstract concept into observable variables or measures.

Predictive model: A statistical or computational model estimating an unknown or future outcome.

Reliability: The consistency of a measurement under comparable conditions.

Scenario planning: Developing coherent alternative future conditions to test strategic choices.

Sampling bias: Systematic distortion caused by a sample that does not adequately represent the target population.

Theory of change: A logical explanation linking inputs and activities to outputs, outcomes and impacts.

Validity: The extent to which a measure or design accurately represents the intended concept or relationship.

Visitor yield: Economic value generated per visitor, trip or other defined tourism unit.

Review Questions

Short Answer Questions

1. Distinguish exploratory research from evaluative research in tourism.

2. Why is reliability alone insufficient when assessing a tourism measure?

3. State two uses of GIS in destination management.

4. What is the difference between a forecast and a scenario?

5. Why should tourism dashboards display data definitions and update times?

Long Answer Questions

1. Explain how a tourism organisation can move from a managerial problem to a well-designed mixed-method research project.

2. Discuss why arrivals, occupancy and expenditure should be interpreted alongside social and environmental indicators.

3. Examine the benefits and risks of using big data and artificial intelligence in tourism management.

4. Explain the role of benchmarking, forecasting and scenario planning in strategic tourism decisions.

5. Discuss the ethical and governance requirements for using location, transaction and visitor-behaviour data.

Analytical / Application Questions

1. A destination’s arrivals have increased by 20 percent, but resident support and average length of stay have declined. Design an evidence system to diagnose the situation and recommend management action.

2. A hotel group wants to introduce an algorithm that predicts cancellation risk and changes prices accordingly. Identify the data, evaluation method, governance controls and possible unintended consequences that should be considered.

3. Design a pilot evaluation for a destination campaign intended to disperse visitors from a congested heritage district to lesser-known attractions. Specify outcomes, indicators, comparison logic and qualitative evidence requirements.

22. Multiple Choice Questions with Answers

Question 1

A destination wants to understand why residents oppose tourism despite rising visitor expenditure. Which research design is most appropriate as an initial approach?

A. Exploratory qualitative research with resident interviews

B. A forecast based only on previous arrivals

C. Competitive benchmarking of hotel occupancy

D. A dashboard showing social-media impressions

Correct Answer: A. Exploratory qualitative research with resident interviews

Question 2

Which statement best explains why occupancy should not be used as the only measure of hotel performance?

A. Occupancy cannot be measured numerically

B. Occupancy does not show price, profitability or environmental effects

C. Occupancy is unrelated to room demand

D. Occupancy applies only to destination-level analysis

Correct Answer: B. Occupancy does not show price, profitability or environmental effects

Question 3

A dashboard shows a sharp increase in online searches for a destination. What is the most appropriate interpretation?

A. The searches prove that bookings have increased

B. The searches demonstrate improved resident support

C. The searches indicate interest but require validation against travel or booking data

D. The searches provide a complete measure of visitor expenditure

Correct Answer: C. The searches indicate interest but require validation against travel or booking data

Question 4

Which GIS application would best help a destination identify areas where visitor pressure and resident population overlap?

A. Overlay analysis of visitor concentration and population layers

B. A reliability test of a satisfaction scale

C. A seasonal hotel pricing model

D. A text analysis of online reviews only

Correct Answer: A. Overlay analysis of visitor concentration and population layers

Question 5

Why can big-data sources produce biased tourism conclusions?

A. They always contain too few observations

B. They are often generated by selective users or systems and may not represent the whole population

C. They cannot be analysed using statistical methods

D. They are never updated after collection

Correct Answer: B. They are often generated by selective users or systems and may not represent the whole population

Question 6

Which evaluation approach provides stronger evidence than a simple before-and-after comparison?

A. Comparing an intervention group with a suitable comparison group over the same period

B. Counting only promotional impressions

C. Asking managers whether the intervention seemed successful

D. Comparing unrelated destinations without adjusting definitions

Correct Answer: A. Comparing an intervention group with a suitable comparison group over the same period

Question 7

A benchmark shows that one hotel has lower direct bookings than a competitor. What should managers do before copying the competitor’s strategy?

A. Assume the competitor’s strategy caused the difference

B. Examine differences in customer mix, distribution, brand strength and operating context

C. Remove all intermediary distribution immediately

D. Use the benchmark as proof of poor employee performance

Correct Answer: B. Examine differences in customer mix, distribution, brand strength and operating context

Question 8

What is the principal value of scenario planning in tourism?

A. It guarantees an accurate single prediction

B. It replaces all quantitative forecasting

C. It tests strategies against several plausible futures

D. It eliminates uncertainty from investment decisions

Correct Answer: C. It tests strategies against several plausible futures

Question 9

Which practice best supports responsible use of an AI demand model?

A. Using historical data without checking for inherited bias

B. Evaluating accuracy, fairness, explainability and consequences of error

C. Allowing the model to make decisions without human review

D. Removing documentation so competitors cannot understand it

Correct Answer: B. Evaluating accuracy, fairness, explainability and consequences of error

Question 10

In the Harbourview case, why was a pilot preferred to an immediate permanent intervention?

A. The organisation had no indicators available

B. A pilot allowed effects and unintended consequences to be assessed before wider commitment

C. Residents were excluded from the decision

D. The intervention was expected to have no operational cost

Correct Answer: B. A pilot allowed effects and unintended consequences to be assessed before wider commitment