Home/Evidence Base

Seventeen findings, each tied to the decision it informs.

The shelf we read from before we open your data. Every finding names what was measured, why it might matter for a cruise operator, what an operator would do with it, and the scope within which it holds. Where a line is our inference rather than the paper’s result, it says so.

17 findings · 9 decision areas · 9 with Rob Kwortnik as co-author · Evidence types: observed data, experimental, stated preference, framework, qualitative

Ethan Hawkes & Robert J. Kwortnik, Ph.D. · First published · Reviewed

In brief

What this is
The shelf we read from before we open your data: seventeen peer-reviewed findings, each tied to one of nine cruise commercial decision areas.
How to read it
Each finding states what was measured, why it might matter for an operator, what an operator would do with it and how long that takes, and the scope within which it holds. Our inferences are labeled as ours.
What to do
Filter to the decision on your desk, start with the findings marked “do now,” and bring the diagnostic question to the next revenue meeting.
Decision area
What to do with it
Showing 17 of 17
Do now · no data infrastructure requiredTest next · a pilot or controlled testRequires your data · an analysis on your booking, folio or loyalty data

The findings

Pricing & yieldTest nextObserved dataKwortnik co-author

Pricing architecture should reflect who is booking, not just what they are booking and when.

Cruise pricing simulations suggest third-degree price discrimination can improve revenue by segmenting who is booking, not only what and when they book.

Namin, Gauri & Kwortnik, 2020
“Improving Revenue Performance with Third-Degree Price Discrimination in the Cruise Industry” · International Journal of Hospitality Management, 89, 102597
Diagnostic question

Does your pricing reflect who is booking, or just what they’re booking and when?

What was measured
Modeled cruise demand with latent-class segmentation and simulated third-degree price discrimination across cabin categories and customer segments.
Why this might matter
Most cruise pricing is driven by cabin category and demand period, with minimal customer segmentation in fare-setting. Incorporating who is booking (not just what and when) may capture additional revenue; magnitude depends on demand mix, distribution, and segmentation maturity.
What an operator would do
Pull booking-level data for 24 months: fare paid, cabin category, booking lead time, party size, loyalty tier, channel. Cluster guests by behavioral attributes. Simulate: if you had offered different prices to different segments for the same sailing, what would total fare revenue have been? 4–5 weeks with your RM analytics team.
Scope
Cruise-specific simulation study using one operator dataset. Transferability depends on your segmentation maturity, distribution mix, and cabin structure.
Related
Joo, Gauri & Wilbur, 2019
Pricing & yieldRequires your dataObserved data

Guests become more price-sensitive as the sailing date approaches, so close-in pricing should reflect a different elasticity profile than early bookings.

Price responsiveness in cruise demand increases materially as the booking window shortens. Temporal distance from sailing moderates the effect of price on booking probability.

Joo, Gauri & Wilbur, 2019
“Temporal Distance and Price Responsiveness: Empirical Investigation of the Cruise Industry” · Management Science, 2019
Diagnostic question

Does your close-in pricing reflect the fact that the remaining demand pool has different price sensitivity than your wave-season bookers?

What was measured
Analyzed how price responsiveness changes as departure approaches using a large proprietary cruise data set of prices, bookings, and product attributes. Demand became more sensitive to price later in the advance-sales period even as overall demand was also strongest late in that window.
Why this might matter
Most cruise RM systems adjust price based on fill rate and pace, but few model the changing elasticity of the remaining demand pool. If close-in pricing assumes wave-season elasticity, you’re either leaving money on the table or failing to stimulate bookings for the segments that remain.
What an operator would do
Extract booking pace by sailing, cabin category, and lead-time bucket (90+, 60–90, 30–60, 0–30 days). Calculate effective price elasticity within each window. Compare current close-in rules against the elasticity pattern. If your system applies flat markdown logic, model time-varying pricing. 6–8 weeks with booking data including price changes and demand timestamps.
Scope
Cruise-specific with strong identification. Temporal elasticity pattern is likely generalizable; specific magnitudes depend on brand, itinerary, and distribution.
Related
Namin, Gauri & Kwortnik, 2020
Pricing & yieldDo nowExperimental

Demand-based pricing is more acceptable when fences feel fair and prices are framed as discounts rather than surcharges.

In restaurant tests across Singapore, Sweden, and the United States, perceived fairness of demand-based pricing depended on the fence used and on whether the price was framed as a discount or a surcharge.

Kimes & Wirtz, 2003
“Has Revenue Management Become Acceptable? Findings from an International Study on the Perceived Fairness of Rate Fences” · Journal of Service Research, 6(2), 125–135
Diagnostic question

For your three most common dynamic pricing touchpoints, can guests see why the price varies, and do they perceive it as a discount or a surcharge?

What was measured
Tested five restaurant rate fences across three countries. Coupons, time-of-day pricing, and lunch/dinner pricing were perceived as fair; weekday/weekend pricing was neutral to slightly unfair; table-location pricing was somewhat unfair. Framing demand-based pricing as discounts improved perceived fairness.
Why this might matter
Cruise application is an inference, but the fairness lesson is useful: if variable pricing is hard to explain or feels like a surcharge, acceptance risk goes up.
What an operator would do
Audit your pricing communication this week. For each dynamic pricing touchpoint (surge excursion pricing, dynamic bar pricing, upgrade offers, loyalty fences): does the guest understand why the price varies? Is it framed as discount from reference or surcharge? Pick the highest-volume touchpoint and A/B the framing language. Audit: one week. A/B: one sailing cycle. No data infrastructure required. This is a copy test.
Scope
Restaurant context across three countries. Cruise needs adaptation for inclusive pricing, ancillaries, and multi-day relationship dynamics.
Related
Namin, Gauri & Kwortnik, 2020, Noone & Mattila, 2026
Pricing & yieldTest nextExperimental

Conditional and bid-based upgrade models outperformed guaranteed upgrades on attractiveness and upgrade intent; bid behavior suggests higher revenue potential.

Despite greater uncertainty, consumers found Conditional and Bid upgrade models more attractive than Guaranteed upgrades, and the spread of bid amounts suggests stronger revenue potential.

Noone & Mattila, 2026
“Fee-Based Upgrading for Hotel Revenue Management: Conditional and Bid Models Outperform Guaranteed Upgrades” · Cornell Hospitality Quarterly, doi:10.1177/19389655261416854
Diagnostic question

Are you monetizing unsold premium cabin inventory through fee-based upgrades, and if so, are you using a model that captures heterogeneous willingness to pay, or a one-size-fits-all fixed price?

What was measured
Mixed-methods: Study 1 (n=248) compared Guaranteed, Conditional, Bid, and no-upgrade options; 63.3% chose to upgrade, with Conditional and Bid selected more often than Guaranteed. Study 2 (n=140) tested the mechanisms: perceived savings and attractiveness supported Conditional, while autonomy over price, savings, and attractiveness supported Bid. Outcome uncertainty did not materially deter either model.
Why this might matter
Cruise lines are increasingly monetizing unsold premium inventory. This paper suggests uncertainty does not automatically kill upgrade demand when guests already accept the base purchase, and that bid formats may capture heterogeneous willingness to pay more effectively than a single fixed offer.
What an operator would do
If you’re currently using Guaranteed upgrades only (fixed-price upsell at booking or check-in), pilot a Conditional or Bid model on a subset of sailings. For Conditional, emphasize savings with low risk (“Save $X on a suite. Only charged if available”). For Bid, emphasize autonomy (“Name your price. You keep your current cabin if not accepted”). Track conversion rate, incremental upgrade revenue, average upgrade revenue per transaction, and whether upgraded guests book the higher category next time. During high demand, prioritize displaying premium product at full price; during low demand, deploy Conditional/Bid to monetize unsold inventory. Scoped as a 3-month pilot.
Scope
Hotel setting. Cruise applications require adaptation for longer booking windows, inclusive dynamics, and multi-day context. The experimental evidence is hotel-based, even if the managerial logic is relevant to cruise.
Related
Kimes & Wirtz, 2003, Anderson & Xie, 2010
Loyalty & retentionRequires your dataObserved dataKwortnik co-author

New and repeat cruisers behave differently, and those differences should shape how you design retention strategy.

Using more than one million reservation records from one cruise brand, the paper compares new, first-time repeat, and multi-time repeat cruisers.

Sun, Kwortnik & Gauri, 2018
“Exploring Behavioral Differences Between New and Repeat Cruisers to a Cruise Brand” · International Journal of Hospitality Management, 71, 132–140
Diagnostic question

Do you know the rebooking rate from first sailing to second, and is your loyalty program designed to improve it?

What was measured
Used more than one million reservation records from one cruise brand. Compared with new cruisers, repeat cruisers were less price sensitive, lived closer to embarkation ports, were more likely to choose longer cruises and better cabin types, and booked further in advance; first-time and multi-time repeaters also differed.
Why this might matter
H&K inference: If your loyalty program concentrates on heavy repeaters whose behavior is already stable, you may be transferring margin rather than shifting trajectory. H&K inference: The first-to-second booking window can be a high-leverage place to test loyalty investment, not a published finding from this paper, but a practical hypothesis to validate with your guest data.
What an operator would do
H&K inference: Segment your guest file: new, first-time repeat, multi-repeat (3–5), heavy loyal (5+). For each: average fare, lead time, cabin mix, onboard spend/day, rebooking probability within 18 months, loyalty cost. If a large share of loyalty spend lands on heavy loyalists with very high baseline rebooking rates, model redirecting some of that investment toward first-time repeaters as a test. Pilot: targeted benefits for guests who sailed once but haven’t booked again. Measure at 6 and 12 months. Analysis: 4–5 weeks. Pilot: 6–12 months.
Scope
One cruise brand. Specific magnitudes depend on brand, itinerary, and program structure.
Related
Rust, Lemon & Zeithaml, 2004
Itinerary & destination economicsTest nextObserved dataKwortnik co-author

Shore excursions vary systematically by category and region, so destination product design should be managed as a portfolio rather than a generic add-on.

Using a large cruise-excursion dataset, the paper identifies excursion categories, a core-periphery structure, and regional differentiation in type, positioning, duration, and price.

Sun, Kwortnik, Xu, Lau & Ni, 2021
“Shore Excursions of Cruise Destinations: Product Categories, Resource Allocation, and Regional Differentiation” · Journal of Destination Marketing & Management, 21, 100660
Diagnostic question

Is your shore excursion strategy managed as procurement or as a product portfolio, and do you know which categories drive attach and contribution?

What was measured
Analyzed 3,259 shore-excursion products across 458 ports in Royal Caribbean’s network. Identified seven excursion categories plus a core-periphery structure, with regional differentiation in excursion type, image/positioning, duration, and price.
Why this might matter
For destination-intensive brands, excursions are core product and a meaningful revenue stream. The paper is descriptive rather than causal, but it supports managing excursions as a portfolio rather than a procurement afterthought.
What an operator would do
Pull your excursion catalog: categorize by type (cultural, adventure, culinary, nature, luxury private). Map revenue per excursion-guest-day, attach rate, and satisfaction by category and region. Identify: highest-contribution categories, lowest attach relative to satisfaction, and regional misalignment. Test category pricing and bundling on one region for one season. 4–6 weeks for analysis, one season for test.
Scope
Royal Caribbean excursion dataset. Most applicable to destination-intensive and premium brands.
Related
Lee & Ramdeen; Espinet-Rius et al.; Niavis & Tsiotas, 2013
Itinerary & destination economicsRequires your dataObserved data

Itinerary structure affects both demand and price. If you price all 7-day sailings as interchangeable, you’re averaging away willingness to pay.

Across multiple cruise studies, itinerary structure helps explain both occupancy and price alongside cabin type, booking timing, and ship attributes.

Lee & Ramdeen; Espinet-Rius et al.; Niavis & Tsiotas, 2013
“Itinerary Attributes, Occupancy, and Cruise Pricing (multiple studies)” · Tourism Management; European Journal of Management and Business Economics; Research in Transportation Business & Management (synthesis)
Diagnostic question

Does your RM price itineraries based on measurable willingness-to-pay differences, or treat all 7-day sailings as equivalent demand pools?

What was measured
This synthesis combines occupancy and pricing studies. Lee & Ramdeen (2013) found itinerary type explained 23% of occupancy-rate variance across nearly 30,000 voyages on 15 itineraries. Hedonic pricing studies then showed fares vary with nights, departure timing, booking timing, homeport, itinerary structure, cabin type, and ship attributes using 36,000+ actual prices and 176,000 listed prices.
Why this might matter
If your RM prices on cabin and pace without weighting itinerary attributes, you’re treating different demand propositions as equivalent. These studies show itinerary structure belongs in the pricing model, even if operator-specific values still need internal validation.
What an operator would do
Run a hedonic pricing decomposition on your fare data. Regress effective fare against itinerary attributes controlling for cabin and lead time. Compare implied attribute values to how your RM differentiates. 5–6 weeks with 2–3 years of booking data.
Scope
Synthesis of occupancy and hedonic-pricing studies using published or observed fare data rather than one operator’s effective net fares. Attribute values still need internal validation.
Related
Sun, Kwortnik, Xu, Lau & Ni, 2021, Namin, Gauri & Kwortnik, 2020
Onboard & ancillary revenueRequires your dataFramework

Revenue management should account for total revenue contributions, not just the primary ticket or room stream.

This Cornell Hospitality Quarterly review argues that hospitality revenue management should broaden from inventory pricing alone toward total revenue contributions, including ancillary streams.

Anderson & Xie, 2010
“Improving Hospitality Industry Sales” · Cornell Hospitality Quarterly, 51(1), 53–67, doi:10.1177/1938965509354697
Diagnostic question

Do you manage fare and onboard revenue as separate targets, or as one total-revenue system?

Operator field noteThis is the dynamic Ethan managed at MSC: every deep discount shifted the guest mix toward lower-spending passengers, compressing onboard revenue per passenger day. Coordinating ticket pricing and onboard revenue planning was the most consequential operational change during his VP Onboard Revenue tenure.
What was measured
Reviews more than twenty years of Cornell Quarterly revenue-management research across hotels, restaurants, golf, function space, distribution, and pricing, and points to total revenue contributions as an important direction for practice and research. It does not estimate a cruise-specific ticket-to-onboard coefficient.
Why this might matter
Cruise application is the next analytic step: fare strategy changes guest mix, guest mix changes onboard propensity, and those streams should be managed together rather than as isolated KPIs.
What an operator would do
Build a ticket-to-onboard bridge model. Pull ticket pricing and onboard spend at guest level for 12–18 months. Regress onboard spend/day against fare paid while controlling for itinerary, cabin, loyalty tier, and channel. Treat the result as an operator-specific estimate, not as a coefficient supplied by this paper. 3–4 weeks with booking and folio data.
Scope
Conceptual review article. It argues for broader revenue management but does not directly measure ancillary spillovers in cruise.
Related
Ng & Kwortnik, 2008
Brand positioningTest nextStated preferenceKwortnik co-author

Your competitive set may not be who you think. Guests often categorize by experience type, not cabin class. Two lines competing on fare may not compete for the same guest.

Studies how passengers categorize cruise lines by perceived experience.

Li & Kwortnik, 2017
“Categorizing Cruise Lines by Passenger Perceived Experience” · Journal of Travel Research, 56(7), 941–956
Diagnostic question

Is your competitive positioning built on how guests form consideration sets, or on how your revenue system categorizes inventory?

What was measured
Using J.D. Power survey data, found passengers form consideration sets around experience categories differing from marketer types. Measured choice determinants and stated loyalty intentions, not observed booking behavior.
Why this might matter
If pricing is anchored to cabin-class comparisons and guests decide along different dimensions, you may be pricing against non-substitutes. Stated-preference data: directional, requires transactional validation.
What an operator would do
Validate against switching: pull guests who sailed you and a competitor in 36 months. Map actual switching patterns. Compare to comp set your RM uses for fare matching. Also pull quote-shop or win-loss data. 2–3 weeks with cross-brand data.
Scope
Stated preference and loyalty intentions (survey), not observed transactions. Requires transactional validation.
Related
Namin, Gauri & Kwortnik, 2020
Experience designDo nowExperimentalKwortnik co-author

Where you place surprise, anticipation, and peak moments in the guest journey is a design choice you can test, not a solved revenue formula.

Studies surprise, anticipation, and sequence effects in how service experiences are evaluated.

Dixon, Victorino, Kwortnik & Verma, 2017
“Surprise, Anticipation, and Sequence Effects in Service and Experience Design” · Production & Operations Management, 26(5), 945–960 · Most Influential Paper Award
Diagnostic question

Does your experience arc build toward a peak, or away from one?

What was measured
Experimentally tested how placement of surprise, anticipation, and peak moments within a service sequence affects remembered experience evaluation.
Why this might matter
Your highest-margin experiences (specialty dining, premium excursions, spa) may anchor memory; if so, scheduling matters. Treat sequencing choices as testable, not as proven drivers of rebooking or revenue in your fleet until you run field tests.
What an operator would do
Map your experience arc for one itinerary. Plot by day: signature dining, best excursion port, flagship show, final-night event, next-cruise desk. If peaks cluster early, run a controlled test shifting key moments to mid-cruise on matched sailings. Measure satisfaction and any rebooking metrics you already track; treat revenue and rebooking links as hypotheses to validate, not as effects proven by this study for your operation.
Scope
Experimental mechanism in controlled settings. Field tests needed to link sequence choices to rebooking or revenue in your context.
Related
Kwortnik & Thompson, 2009
Experience designRequires your dataFrameworkKwortnik co-author

Managing service experiences requires a cross-functional system, not isolated touchpoint ownership.

Using a leisure-cruise case and a service-systems lens, the paper proposes service experience management as a way to unify service operations and marketing.

Kwortnik & Thompson, 2009
“Unifying Service Marketing and Operations with Service Experience Management” · Journal of Service Research, 11(4), 389–406
Diagnostic question

Does your organization have a cross-functional process for designing the experience arc, or does each department manage its own touchpoints independently?

What was measured
Using a multimethod empirical case in the cruise industry and literature from marketing and operations, the authors derive a service-systems model and propose a service experience management function to unify planning and delivery.
Why this might matter
Most cruise operators manage individual touchpoints (embarkation, dining, entertainment, excursions) through separate departments with separate metrics. What they do not manage is the arc: how those touchpoints interact, sequence, and compound to form the remembered experience.
What an operator would do
Audit whether your organization has an explicit experience design process that spans departments, or whether experience quality is managed implicitly. If the latter, pilot an experience management system on one itinerary: map journey stages, define the intended arc, assign measurement at each stage, create a cross-functional review cadence. 3–4 weeks to map; one deployment cycle to test.
Scope
Framework paper grounded in a leisure-cruise case. Operational gains still need validation within each operator.
Related
Dixon, Victorino, Kwortnik & Verma, 2017
Experience designTest nextQualitativeKwortnik co-author

The physical environment of the ship (layout, décor, ambient conditions, spatial design) shapes how guests perceive the brand and attach meaning to the cruise experience.

Ambient conditions, layout, décor, size, facilities, and social factors within the ship environment influence pleasure and the meanings cruisers attach to cruise brands.

Kwortnik, 2008
“Shipscape Influence on the Leisure Cruise Experience” · International Journal of Culture, Tourism, and Hospitality Research
Diagnostic question

Can you quantify which aspects of your ship’s physical environment most drive guest satisfaction and brand differentiation?

What was measured
Interpretive qualitative research exploring how cruise passengers experience and make sense of the physical ship environment. Identified multiple dimensions of ‘shipscape’ and their influence on guest experience and brand perception.
Why this might matter
Ship design decisions are typically driven by cost, capacity, and operational efficiency. This research shows the physical environment is also a brand attribute, one that shapes how guests differentiate between brands. For premium and luxury lines, shipscape is a competitive lever. For mass-market lines, it is where the gap between brand promise and delivered experience often shows.
What an operator would do
Conduct a shipscape audit: map key physical environment dimensions (ambient, spatial, social, décor) against guest survey data, identifying which attributes most predict satisfaction and brand differentiation. Compare across ships in your fleet. Use findings to prioritize refurbishment and newbuild design. 4–6 weeks with survey data and environment documentation.
Scope
Qualitative interpretive research establishing mechanisms by which shipscape influences brand perception. Does not quantify the effect on rebooking or revenue. That requires separate field measurement.
Related
Dixon, Victorino, Kwortnik & Verma, 2017, Kwortnik & Thompson, 2009
Service policy designRequires your dataObserved dataKwortnik co-author

Tipping-policy design can affect customer satisfaction, so gratuity changes should be treated as testable service-policy decisions rather than assumed neutral.

Cruise evidence suggests customer satisfaction differs across tipping-policy designs.

Lynn & Kwortnik, 2015
“The Effects of Tipping Policies on Customer Satisfaction: A Test from the Cruise Industry” · International Journal of Hospitality Management, 51, 15–18
Diagnostic question

When your gratuity policy last changed, did you measure the satisfaction effect, or was it assumed to be guest-neutral?

What was measured
Tests the relationship between cruise tipping-policy design and customer satisfaction using cruise-industry data.
Why this might matter
If policy design affects satisfaction, gratuity changes are commercial decisions, not just operating-policy changes. Revenue, labor, and repeat-booking effects still need to be measured separately.
What an operator would do
If you’ve changed your gratuity policy in the last 3 years, isolate the satisfaction effect: compare post-voyage scores (controlling for itinerary, ship, segment) before and after. If considering a change, design a controlled test on a subset of sailings. 3–4 weeks for historical analysis.
Scope
Cruise-specific. The paper focuses on satisfaction effects; revenue and rebooking effects must be measured separately.
Related
Kimes & Wirtz, 2003
Total revenue managementDo nowQualitativeKwortnik co-author

Total-revenue trade-offs get hidden when segments generate ticket and onboard value differently.

Teaching case in which a Southeast Asian cruise operator must allocate cabin inventory between casino players and leisure holidaymakers who differ in channels, pricing, and desired cruise experience.

Ng & Kwortnik, 2008
“Balancing Cruise Revenue Sources: The Case of Empress Cruise Lines” · Case Research Journal, 27(Spring), 105–127
Diagnostic question

Ask three VPs what your most profitable guest looks like. Would they agree?

Operator field noteAt MSC, Ethan saw this every wave season: pricing discounted to hit load targets, the mix shifted toward lower-spend passengers, onboard revenue per passenger day compressed. Every function hit its KPI. The enterprise left money on the table. The fix was a shared ‘total guest revenue’ metric.
What was measured
Teaching case set in the southeast Asian cruise market. The CEO of Empress Cruise Lines must decide how to allocate cabin inventory between casino players and leisure holidaymakers under different pricing methods, distribution channels, and experience expectations.
Why this might matter
The case illustrates how segment-specific objectives can pull revenue sources apart. It is useful for surfacing coordination problems, but it is not statistical evidence about the average size of those gaps in cruise.
What an operator would do
Start this week: 90-minute session where pricing, marketing, onboard, and loyalty each bring their ‘high-value guest’ definition and supporting data. The gap is the coordination cost. Then build a shared ‘total guest revenue’ metric (ticket + onboard + ancillary, net of distribution and loyalty costs) alongside existing KPIs for one quarter. Session: no prep. Metric: 2–3 weeks.
Scope
Single teaching case; not a statistical generalization to all cruise operators.
Related
Anderson & Xie, 2010
Guest portfolio valueRequires your dataFramework

Your guest base is a financial asset. Managing it like one changes which decisions get funded and what the commercial team optimizes for.

Provides a customer-equity framework for evaluating marketing strategy options by projected financial return, defined as change in customer equity relative to spend.

Rust, Lemon & Zeithaml, 2004
“Return on Marketing: Using Customer Equity to Focus Marketing Strategy” · Journal of Marketing, 68(1), 109–127
Diagnostic question

If you asked leadership whether last quarter’s decisions built or eroded the guest base value, could they answer with evidence?

What was measured
Built a framework linking marketing actions to customer equity through value equity, brand equity, and retention equity, and illustrated the method with an airline-industry application.
Why this might matter
Most hospitality teams optimize for period metrics that don’t connect to enterprise value. Customer equity makes that connection possible, requiring operator-specific assumptions. If your guest-level data isn’t readily segmentable, the first step is a data readiness assessment: can you link booking, onboard spend, and loyalty data at the guest level?
What an operator would do
Build a minimum viable guest equity model. Inputs: five contribution-based segments, 36-month repeat probability, contribution per sailing (net distribution and loyalty costs), onboard contribution, discount rate the CFO accepts. Output: portfolio value and sensitivity table (what is a 1-point retention lift worth per segment?). 3–4 weeks with 24+ months of guest booking history.
Scope
Framework paper with an airline example. Cruise requires adaptation for distribution, seasonality, and the ticket-to-onboard split.
Related
Gupta, Lehmann & Stuart, 2004, Sun, Kwortnik & Gauri, 2018
Guest portfolio valueRequires your dataObserved data

Customer lifetime value can help estimate what your company is worth. A 1% retention improvement compounds more powerfully than margin or acquisition gains.

Shows how customer valuation can be used to value firms, including high-growth firms with negative earnings.

Gupta, Lehmann & Stuart, 2004
“Valuing Customers” · Journal of Marketing Research, 41(1), 7–18
Diagnostic question

Can your finance team model what a 1% retention change is worth in enterprise value terms?

What was measured
Using publicly available data for five firms, estimated customer value and showed that a 1% improvement in retention increased firm value more than equivalent percentage improvements in margin or acquisition cost.
Why this might matter
Not hospitality. But customer value drives firm value, and retention compounds. The foundation for treating the guest portfolio as a valuation input.
What an operator would do
Model retention sensitivity: take rebooking rate by segment, shift 1 point, flow incremental margin through a 3-year DCF. Inputs: rebooking rate, contribution/sailing, discount rate. Output: a one-page sensitivity table your CFO can use in budget conversations. Two weeks.
Scope
Five public-company examples outside cruise and hospitality. Adaptation required.
Related
Rust, Lemon & Zeithaml, 2004, Keiningham, Cooil, Andreassen & Aksoy, 2007
Guest portfolio valueDo nowObserved data

NPS does not reliably predict revenue growth better than other satisfaction measures. If your reporting is built around it, test the assumption against your own rebooking data.

Net Promoter Score was not a superior predictor of firm revenue growth compared to other satisfaction metrics in a multi-industry longitudinal study.

Keiningham, Cooil, Andreassen & Aksoy, 2007
“A Longitudinal Examination of Net Promoter and Firm Revenue Growth” · Journal of Marketing, 71(3), 39–51
Diagnostic question

In your data, does NPS predict rebooking better than simpler measures, or is that an assumption you’ve never tested?

What was measured
Used longitudinal data from 21 firms and more than 15,500 interviews from the Norwegian Customer Satisfaction Barometer and compared the results with the American Customer Satisfaction Index. Net Promoter did not show clear superiority as a revenue-growth predictor.
Why this might matter
Most cruise lines have significant NPS infrastructure. If NPS is not predictively superior, is that infrastructure driving the right decisions? This is pro-testing, not anti-measurement.
What an operator would do
Two-week analysis: pull guest-level NPS, match to rebooking over 12–24 months. Same with overall satisfaction. Compare correlation coefficients with rebooking, spend, referral. If NPS isn’t meaningfully better, simplify your measurement system.
Scope
Multi-industry. Predictive power varies by operator and survey design.
Related
Rust, Lemon & Zeithaml, 2004, Gupta, Lehmann & Stuart, 2004
Filters ↑

Book a review · 30 minutes · Both partners

Bring one live decision.

In 30 minutes with both partners, we will stress-test the question and tell you plainly whether deeper work is warranted, including “do not hire us” if that is the right answer.

The leader who owns the decision usually joins: the CEO, CCO, or head of Revenue, Finance or Loyalty. We reply within one business day. What happens next →

Your information is confidential. No mailing list, no drip sequence. If it is not a fit, we will tell you plainly.