Methods & evidence
Claims proportionate to evidence.
The initiative is designed to improve decisions, not to manufacture certainty. Its methods make visible the quality of the evidence, the assumptions within an analysis, the boundary of a finding, and the conditions under which that finding should change.
Five connected domains.
Trust and welcome
How international audiences perceive credibility, welcome, safety, predictability, mission, and institutional relevance.
Demand pipeline
Whether an institution or destination enters consideration, strengthens intent, and remains in future travel plans.
Conversion and anchoring
Whether intent becomes itinerary inclusion, whether an institution helps justify the trip, and whether prospective demand becomes an actual visit.
Onsite performance
What can be observed through attendance, visitor mix, experience, recommendation, earned revenue, and repeat behavior.
Place-based contribution
How institutional visitation may contribute to activity beyond the institution, including stays, spending, partner visits, and wider destination value.
These domains describe a sequence, but sequence alone does not prove causation. Evidence is required at each connection.
In use
Methods already at work.
The initiative's diagnosis of the current demand environment was not asserted. It was estimated — with named, examinable techniques.
Structural-break diagnostics
Standard gravity models explain 60–80% of bilateral tourism flows in normal times. From early 2025 their residuals for U.S. inbound travel turned persistently negative — 6% to 9% below prediction — a dated structural break (Bai-Perron). Adding trust and friction variables restores the fit: the operative factors are identified, not assumed.
Regime identification
Markov regime-switching distinguishes a price-elastic regime, where fluctuations mean-revert on their own, from the confidence-constrained regime now indicated — where damage persists until the underlying state variables change. The estimated regime probability is the initiative's most informative leading indicator.
Belief dynamics
Bayesian updating formalizes why trust falls faster than it recovers: negative signals carry higher effective precision than positive ones, and personal endorsement — the highest-precision channel a traveler has — has measurably weakened in continuous source-market tracking (to 61–64 on a 100-point scale).
Destination choice
Discrete-choice modeling (McFadden) treats visitation as utility-maximizing substitution across a global consideration set — quantifying competing regions' structural advantages in cultural yield per day, intra-regional connectivity, and friction premium.
Consideration-set dynamics
Markov-chain planning cycles: a destination's inclusion in the traveler's active set decays from roughly 65% to 31% over about four cycles; recovery takes six to eight. With reacquisition costing about three times one visit's earned revenue, prevention dominates cure.
Distributional risk
Loss functions here are convex — by Jensen's inequality, the average scenario understates expected loss. Monte Carlo simulation (200,000 draws per scenario) produces full loss distributions and tail measures for a normalized one-million-visit institution: mean annual earned-revenue losses of $219,000–$1.46 million; $2.23 million average across the worst 5% of outcomes.
The named tests, parameters, and estimates summarize the initiative's March 2026 executive intelligence report. Model outputs are not observed results.
The initiative publishes meaningful learning without disclosing private deliberations or unnecessary institutional detail. Participant discussion, institution-level data, personal information, and draft judgments remain controlled unless a separate public-use decision has been documented.