A tourism survey says travellers want a new experience, but only a small number of bookings follow. Did the survey fail, was the price unrealistic or was the sample drawn from people who could never actually visit?
The gold standard of tourism market research is a reproducible understanding of visitor needs and real choices, with honest sampling and practical testing. This 23-chapter guide connects interviews, surveys, segmentation, conversion and community outcomes.
Quick information: choose the evidence for the question
| Research method | Can reveal | Cannot automatically prove |
|---|---|---|
| Qualitative interview | Reasons, language and unmet needs | Population-wide percentage |
| Representative survey | Estimated specified responses | Guaranteed purchase behaviour |
| Booking analytics | Events and completed transactions | Unique people from session totals |
| Official tourism data | Dated arrivals, receipts and hotel trends | Preferences for one new attraction |
| Concept test | Reactions under stated proposal | Actual future profitability |
| Market pilot | Observed choice in trial conditions | Universal success after full rollout |
Worked booking funnel with defined denominators
| Fictional journey step | Count |
|---|---|
| Eligible web page visits | 1,000 |
| Booking starts | 200 |
| Confirmed bookings | 40 |
| Start rate per page visit | 20% |
| Confirmed booking rate per page visit | 4% |
| Confirmation per booking start | 20% |
The percentages describe events under an assumed tracking system, not unique tourists. Each user could interact several times. The result is not evidence that one fifth of all visitors purchased the product.
Survey response and nonresponse example
| Fictional survey | Value |
|---|---|
| Past travellers invited | 400 |
| Respondents | 200 |
| Positive responses | 150 |
| Response rate | 50% |
| Positive among respondents | 75% |
The other 200 invited travellers did not respond, and their preferences are unknown. The observed 75% cannot automatically be applied to the entire customer population.
Changed context: bookings and completed experiences
| Fictional market estimate | Count or proportion |
|---|---|
| Potential eligible opportunities | 10,000 |
| Reached at assumed 20% | 2,000 |
| Confirmed at assumed 5% of reached | 100 |
| Completed at assumed 80% of bookings | 80 |
| Completed out of original opportunities | 0.8% |
Every percentage is illustrative. The model does not represent a real market forecast, and ‘opportunities’, bookings and unique people are not necessarily identical measurement units.
All 23 tourism market-research controls
Define the market and research method
Chapters 1–6: Market research begins with an honest customer question · Distinguish a commercial question from a public-policy one · Define the population before recruiting a sample · Use the right quantitative and qualitative methods · Build a visitor journey map from actual tasks · Distinguish audience segments from stereotypes.
Surveys, interviews and customer evidence
Chapters 7–12: Use official market data to frame, not replace, research · Check what a booking conversion rate really counts · Worked conversion example with changing denominators · Survey design can change the apparent answer · Worked response-rate and satisfaction example · Qualitative coding should preserve context.
Market sizing, segmentation and ethics
Chapters 13–18: Concept testing should reflect the actual offer · Stated willingness to pay is not a completed purchase · Estimate market size with an honest eligible population · Changed-context market size: attendance may differ from bookings · Seasonality changes both demand and sampling · Competitor research must not become unverified imitation.
Verification and accountable product design
Chapters 19–23: Digital behaviour data can contain coverage bias · Protect consent, privacy and vulnerable participants · Consider host communities when developing visitor products · Test actual outcomes and keep the research honest · Publish a research report that readers can verify.
- Market research begins with an honest customer question
- Distinguish a commercial question from a public-policy one
- Define the population before recruiting a sample
- Use the right quantitative and qualitative methods
- Build a visitor journey map from actual tasks
- Distinguish audience segments from stereotypes
- Use official market data to frame, not replace, research
- Check what a booking conversion rate really counts
- Worked conversion example with changing denominators
- Survey design can change the apparent answer
- Worked response-rate and satisfaction example
- Qualitative coding should preserve context
- Concept testing should reflect the actual offer
- Stated willingness to pay is not a completed purchase
- Estimate market size with an honest eligible population
- Changed-context market size: attendance may differ from bookings
- Seasonality changes both demand and sampling
- Competitor research must not become unverified imitation
- Digital behaviour data can contain coverage bias
- Protect consent, privacy and vulnerable participants
- Consider host communities when developing visitor products
- Test actual outcomes and keep the research honest
- Publish a research report that readers can verify
CONTROL 1 OF 23 · All controls
1. Market research begins with an honest customer question
Tourism market research investigates visitor needs, perceptions, constraints and choices so destinations can design better experiences and services. It is different from simply collecting reviews, copying a competitor’s marketing or predicting a trend from one photograph. The gold standard is a well-defined question answered with methods and evidence suited to that decision.
A destination might ask why visitors avoid an attraction, what information they need to plan transport or whether a proposed experience creates value for a particular group. Tourism Data Analytics handles broader measurement and modelling; market research concentrates on understanding audiences and testing proposed offerings. A business decision must also respect residents and the places being visited.
CONTROL 2 OF 23 · All controls
2. Distinguish a commercial question from a public-policy one
A tour operator may investigate which package travellers would buy, while a destination organisation may need to understand how visitors use public spaces and how tourism affects residents. Both can use surveys, but their objectives and legitimate decision rights differ.
Name the organisation asking the question, the outcome sought and the people potentially affected. A tourism research programme should not treat higher sales as a complete substitute for community or environmental outcomes. Tourism Impact Assessment and Public Policy connect market evidence to the wider responsibilities of destinations.
CONTROL 3 OF 23 · All controls
3. Define the population before recruiting a sample
Inbound visitors, resident day-trippers, families travelling with children and overnight hotel guests are distinct populations. An airport survey may miss some domestic visitors, while a hotel survey may exclude same-day travellers. The sample must correspond to the decision’s target population.
UN Tourism’s International Recommendations for Tourism Statistics provide definitions of visitors and tourism trips. Write the eligibility rule, geographic boundary, date range and recruitment route before analysis. A beautifully formatted sample of the wrong population cannot establish the answer.
CONTROL 4 OF 23 · All controls
4. Use the right quantitative and qualitative methods
Interviews can reveal why people experience a problem and how they describe it, while structured surveys can estimate the prevalence of defined responses under suitable sampling. Observation can expose behaviour that people do not mention in interviews. Desk research can establish existing market context.
Choose methods based on the question. A dozen interviews can surface valuable themes but are not a representative percentage of all tourists. A large survey can quantify selected responses yet miss the reasons behind them. Sound research may combine methods, explaining what each contributes and what remains uncertain.
CONTROL 5 OF 23 · All controls
5. Build a visitor journey map from actual tasks
A traveller’s experience can begin before a trip with search and booking, then continue through arrival, transport, accommodation, attractions, payment and return. Problems at one step may stop the trip even when later attractions are excellent.
Map decisions, information required, costs and observed failures. Verify the actual journey using appropriate participants rather than assuming that every traveller shares one itinerary. The Tourism Geography root explains real paths and transport constraints that a marketing funnel may overlook.
CONTROL 6 OF 23 · All controls
6. Distinguish audience segments from stereotypes
Useful market segments may reflect trip purpose, needs, timing, budget constraints or desired experiences. Nationality and age can be descriptive variables but should not automatically determine a person’s preferences. A destination can mislead itself by treating one demographic group as internally identical.
Test segment definitions against observed needs and behaviour. Segment size, willingness or ability to travel and access requirements must be measured appropriately. A segment label is a tool for understanding and communication, not proof that all members want the same product.
CONTROL 7 OF 23 · All controls
7. Use official market data to frame, not replace, research
Published visitor arrivals, tourism receipts and hotel measures provide valuable context but cannot answer every question about preferences or purchase barriers. A national aggregate may remain stable while interest in a specific local experience changes substantially.
Singapore’s Department of Statistics reports 2025 international visitor arrivals at about 16.9 million and tourism receipts of approximately S$32.8 billion. These are dated official national outcomes, not figures for a fictional tour business or forecasts of a specific attraction’s market. Research questions need their own evidence at the relevant level.
CONTROL 8 OF 23 · All controls
8. Check what a booking conversion rate really counts
A website may report visitors, search sessions, completed bookings and paid arrivals. These are different events. A single traveller can create several sessions or book for multiple people, so the numerator and denominator must be chosen consistently.
If 500 fictional eligible booking sessions lead to 25 confirmed bookings under a defined tracking method, session-to-booking conversion is 25/500 = 5%. It does not establish that 5% of all travellers booked or that each booking represents one visitor. Attribution and tracking omissions also matter.
CONTROL 9 OF 23 · All controls
9. Worked conversion example with changing denominators
Consider a fictional campaign with 1,000 eligible page visits, 200 booking-start events and 40 confirmed bookings. Booking-start rate per page visit is 200/1,000 = 20%; confirmed bookings per page visit are 40/1,000 = 4%; confirmations per start are 40/200 = 20%.
All three percentages are valid under those defined event populations, but answer different questions. If one visitor revisits the page, the event data may count several visits. A manager who says 20% of all visitors bought a tour based on booking starts would be mistaken.
CONTROL 10 OF 23 · All controls
10. Survey design can change the apparent answer
Question wording, order, language and offered response options influence what people say. Asking whether travellers enjoy cultural experiences is not the same as asking which experience they would choose at an actual price and available time.
Pilot questionnaires with representative readers and remove ambiguities. Distinguish prompted responses from spontaneous descriptions and disclose whether options were rotated or shown in a fixed order. Do not build a promotional question whose only attractive answer is the preferred product.
CONTROL 11 OF 23 · All controls
11. Worked response-rate and satisfaction example
Suppose a fictional operator invites 400 past travellers to a survey. Two hundred respond and 150 of those rate the experience positively. Response rate is 200/400 = 50%, while positive response share among respondents is 150/200 = 75%.
A statement that 75% of all customers were satisfied would require assumptions about the nonrespondents and survey design. The observed proportion is among respondents. To make a population inference, examine recruitment, language, timing and possible nonresponse bias rather than merely displaying a larger percentage.
CONTROL 12 OF 23 · All controls
12. Qualitative coding should preserve context
Interview research often groups observations into themes such as unclear transport information, crowded venues or poor accessibility. The categories can guide product design, but coding decisions and interpretation must be transparent enough to review.
Keep original statements appropriately protected, record coding definitions and examine contrary examples. Do not turn a participant’s complaint into a quote from someone else or claim a theme was universal when only a few interviewees raised it. Research provides a wider methodological foundation.
CONTROL 13 OF 23 · All controls
13. Concept testing should reflect the actual offer
A destination can ask participants about a proposed trail, heritage workshop or shuttle before building it. Responses may reveal confusion, enthusiasm and practical constraints. Yet a concept shown without price, travel time or operating restrictions may seem more attractive than the real product.
Include realistic alternatives, access conditions and the effort required to participate. Explain which parts are hypothetical and how the test was conducted. A concept preference can inform a pilot but is not evidence that the finished attraction will succeed commercially or socially.
CONTROL 14 OF 23 · All controls
14. Stated willingness to pay is not a completed purchase
In surveys, people may claim they would pay for an experience but act differently when faced with dates, transport, competing options and an actual payment. Stated preference and revealed behaviour are useful but different evidence sources.
A fair test may combine appropriately designed questions with observed pilot bookings or real customer choices where authorised. Do not present a hypothetical price response as a guaranteed future revenue stream. State currency, inclusions, date and relevant group.
CONTROL 15 OF 23 · All controls
15. Estimate market size with an honest eligible population
A target market calculation may start with a population, estimate who can practically reach the product and then estimate likely participation. Each stage is a separate assumption. Multiplying several optimistic shares can create a large-looking number that lacks independent support.
For a fictional market of 10,000 eligible potential visits, suppose 20% are reached by a campaign and 5% of those reached convert into a booking. The resulting illustrative bookings are 10,000×0.20×0.05 = 100. This is not an actual forecast; each percentage would need valid measurement and an operational definition.
CONTROL 16 OF 23 · All controls
16. Changed-context market size: attendance may differ from bookings
In the preceding invented scenario, suppose only 80 of the 100 confirmed bookings are fulfilled in the measured period. The booking-to-completed-experience share is 80/100 = 80%, and completed experiences equal 0.8% of the original 10,000-unit hypothetical population.
The calculation highlights how one funnel stage should not be substituted for the final outcome. Some bookings may be cancelled or legitimately moved, and travel parties may contain multiple people. A rigorous report distinguishes sessions, bookings, individual visitors and completed services.
CONTROL 17 OF 23 · All controls
17. Seasonality changes both demand and sampling
People interviewed during a school holiday or major festival may have different purposes from visitors in an ordinary working week. A survey at one attraction on one busy day cannot automatically represent the destination’s entire year.
Recruit across relevant periods, routes and traveller groups where the decision requires it. Distinguish seasonally adjusted and raw official data. Tourism Data Analytics explains how time-series and visitor definitions affect statistical interpretation.
CONTROL 18 OF 23 · All controls
18. Competitor research must not become unverified imitation
A competitor’s public price, opening time or review score may change and may apply to a different service package. Copying its visible offer without accounting for access, quality and local conditions can produce a flawed comparison.
Define equivalent product features and data-collection dates. Verify official pricing and inclusions where feasible, and treat user reviews as selected experiences rather than a scientifically representative sample. Do not fabricate competitor quotes or rely on stale screenshots to make an investment decision.
CONTROL 19 OF 23 · All controls
19. Digital behaviour data can contain coverage bias
Online searches, platform bookings, map interactions and reviews are useful signals, but they represent people using the relevant system. Visitors who pay cash, do not use a smartphone or have limited connectivity may be underrepresented.
Document source coverage, user consent, processing rules and privacy safeguards. Treat inferred movement and actual surveyed intentions as different evidence types. A large number of digital records can be precisely counted while still failing to represent the intended visitor population.
CONTROL 20 OF 23 · All controls
20. Protect consent, privacy and vulnerable participants
Market research may collect personal travel plans, location, spending or household details. Such data must be handled lawfully and with proportionate safeguards. Research involving children or vulnerable individuals may need additional permissions and protections.
Collect only what the question requires, explain purpose and retention, and report aggregate findings responsibly. Do not publish identifiable itineraries or infer sensitive traits from a market segment without a legitimate reason. Ethical research builds trust while protecting participants from unnecessary risk.
CONTROL 21 OF 23 · All controls
21. Consider host communities when developing visitor products
A new attraction may appeal to travellers while causing unwanted pressure on residents, transport services or cultural spaces. Market appeal is therefore not sufficient evidence that the proposed product is appropriate for its location.
Involve affected residents and custodians using suitable methods and decision processes. The Tourism Impact Assessment and Cultural Heritage roots clarify why community consent, conservation and legitimate public objectives may constrain a commercially promising concept.
CONTROL 22 OF 23 · All controls
22. Test actual outcomes and keep the research honest
Once a product launches, compare observed completed visits, cancellations, guest experience and community conditions with the expectations recorded before launch. A retrospective narrative claiming success from any positive number is weak evidence.
Keep survey versions, sample rules, model assumptions and time windows traceable. If observed use differs from stated intention, update the segmentation and product decisions rather than rewrite old forecasts. Research succeeds when it improves decisions and can openly acknowledge surprise.
CONTROL 23 OF 23 · All controls
23. Publish a research report that readers can verify
A gold-standard market research report specifies decision, population, sample and recruitment, methods, dates, questionnaire wording, data protection, analysis, uncertainty, practical recommendations and monitoring plans. It must clearly separate respondent observations from projections.
The immutable Secondary 1 Mathematics Tutor Clementi Small Groups Tutorials page provides an educational analogy: identify denominators, explain calculations and retest changed conditions. The same numeracy prevents misleading booking and satisfaction claims in tourism research.
Tourism market-research audit
| Review question | Evidence | Common mistaken conclusion |
|---|---|---|
| Who was studied? | Visitor definition, sampling, dates and access | Survey sample represents every traveller |
| What was asked? | Question wording and response options | Promotional prompt treated as neutral |
| What was observed? | Unique visitors, sessions, bookings and actual visits | Event totals equal distinct buyers |
| Will they buy? | Realistic offer, price and tested behaviour | Stated preference equals certain revenue |
| Who else is affected? | Residents, cultural custodians and environment | Visitor demand justifies any development |
| Can it be reproduced? | Questionnaire, source data, bias, privacy and corrections | Favourable chart substitutes for method |
Frequently asked questions
What is tourism market research?
Studying visitor needs, constraints, preferences and actual choices to inform travel products and destination decisions.
Is a tourism market segment the same as a demographic stereotype?
No. Useful segments should be supported by observed needs or behaviour, not automatic assumptions about a nationality or age.
Does 75% positive feedback among survey respondents mean 75% of all visitors are satisfied?
Not without appropriate representative sampling and nonresponse analysis.
Are booking-starts the same as confirmed visits?
No. Events, reservations, people and completed experiences are separate populations.
Can stated willingness to pay predict guaranteed revenue?
No. Real purchases depend on prices, timing, availability, alternatives and other constraints.
Can official national arrivals data replace a survey for one specific attraction?
No. National totals give context but not necessarily the preferences or intentions of the target product’s audience.
What makes research independently useful?
Clear decisions, honest samples, dated sources, measured outcomes, uncertainty and ethical handling of participants.
Official statistical context and methodological sources
Official references: UN Tourism Statistics Database (updated 21 July 2026), International Recommendations for Tourism Statistics 2008, Singapore Department of Statistics Tourism Data and Singapore Tourism Board 2025 performance release, published February 2026. These provide dated population context, not live booking forecasts. All funnel and survey figures are fictional.
Connected eduKate Gold Standard subjects
Tourism · Tourism Management · Tourism Economics · Tourism Geography · Tourism Data Analytics · Tourism Impact Assessment · Travel Journalism · Research · Statistics · Data Science · Business · Public Policy · Cultural Heritage · Sustainability · Education
All 23 chapters · Tourism Data Analytics · Research · Education