There is no single travel recovery number. Leisure, business, and international inbound diverged, so one blended read misses operating reality. Property-level foot traffic built on polygon POI cross-checks reported occupancy by separating guests from street traffic, and destination marketing joins CTV exposure to later visits using governed identity.
How to use this article
Read the checklist here, then use the linked hub and product pages for procurement citations.
Travel and hospitality in 2026 no longer runs on a single recovery narrative. Leisure spend overshot 2019 baselines in many markets; business travel has settled below 2019 levels by different amounts per subsegment (our reading, not a published figure); international inbound recovered unevenly; longer-stay domestic absorbed demand business travel used to carry. Reading this correctly needs foot-traffic, CTV-exposure, and card-spend signals, not aggregate ADR and RevPAR alone. GSDSI travel and hospitality frames the catalog; this piece covers what the data says in 2026 and how operators and investors cross-check branded prints against observed behavior.
Investors and brand operators who rely on company-reported metrics alone miss segment divergence until earnings. Alternative data provides independent triangulation: when foot traffic at flagged properties diverges from reported occupancy, the question is geometry, seasonality, or reporting: each with different implications for the thesis.
Segment Divergence: Leisure, Business, International
Leisure travel in drive and sun markets shows sustained overshoot versus 2019 in many DMAs; urban business corridors lag, a split tracked in the US Travel Association's travel research. Convention and group business recovered slower than transient leisure. International inbound depends on origin market and visa corridor, which the Commerce Department's National Travel and Tourism Office reports by origin country: aggregate "inbound recovery" hides country-level gaps. Segment dashboards should never blend leisure and business without labeling: media and revenue teams optimize against different cohorts.
Foot Traffic at the Property Level
Hotel and resort foot-traffic reads require polygon POI on property footprints, not radius fences around lobby pins that capture sidewalk and adjacent retail. Join global mobility to POI & Geofencing with dwell rules separating guest parking from through-traffic. Compare observed visits to reported occupancy by STR segment and seasonality, with segment definitions taken from STR's hotel performance benchmarking. Bureau of Transportation Statistics travel monitors provide external benchmarks for macro sanity checks.
Spend Panels and Card Data
Transaction panels illuminate average ticket, category mix, and geo allocation, especially for F&B, entertainment, and retail adjacent to hospitality assets. Panels skew toward carded spend and under-sample cash-heavy cohorts; document bias before binding to RevPAR models. Triangulate spend momentum with mobility leads: spend confirms wallet behavior mobility suggests.
Destination Marketing: CTV Exposure to Visit
Tourism boards and hotel groups measure destination campaigns by joining CTV/ACR exposure to devices observed in market polygons during booking windows. Requires household or device graph with documented match rates, aggregation floors, and sensitive-location exclusions. See cross-channel measurement for governed join patterns: destination marketing fails when exposure logs never meet place truth.
Property-Level Reads Beat Brand-Level Reads
Hospitality analysis goes wrong when it aggregates to the brand. A flag spans wildly different assets: an airport property serving delayed business travelers, a resort running on leisure seasonality, and a suburban box dependent on a single corporate account can all carry the same sign over the door. Averaging them produces a number that describes no real hotel and hides the divergence that matters commercially.
Work at the property level and roll up only at the end, with the segment mix stated alongside the aggregate. When a brand-level figure moves, the first question is whether the underlying properties moved or whether the mix changed, because a portfolio that opened resorts and closed airport properties will show a trend that has nothing to do with demand. Keep an explicit open-and-close log for the portfolio you are tracking, since same-store comparisons are the only ones that mean anything across a period with meaningful churn.
Choose a Baseline Year Deliberately, and Say Which One
Every recovery chart is an argument about its baseline, and the choice usually gets made by accident. Indexing to a pre-disruption year makes recovery the story. Indexing to the trough makes growth the story. Indexing to last year makes normal seasonal variation look like momentum. All three can be defensible, and presenting one without naming it is where hospitality analysis loses its audience, particularly when the audience includes people who lived through the period being indexed.
State the baseline in the chart title, not the footnote, and where a decision depends on the comparison, show two baselines side by side. Also account for calendar effects before declaring a trend: holiday placement, day-of-week composition, and event calendars move hospitality demand enough to manufacture a recovery narrative from an ordinary month. A same-weekday, same-season comparison is worth more than a sophisticated model applied to mismatched periods.
Triangulate Before You Publish a Recovery Claim
No single sensor sees the travel economy. Foot traffic sees presence but not spend, and it struggles to separate a guest from someone using the lobby restaurant. Card panels see spend but skew by issuer and payment method, missing both corporate billing arrangements and cash-heavy segments. Booking and rate data see intent and price but not attendance. Each has a bias, and the biases do not point the same way.
Treat agreement between two independent sensors as the threshold for publishing a claim, and treat disagreement as a finding worth investigating rather than a problem to average away. When foot traffic rises while spend per visit falls, that is a real and reportable pattern about visit composition. When one sensor moves alone, the most likely explanation is that sensor, and the fastest way to find out is to check whether the divergence starts on a specific date. For the sensor-by-sensor trade-offs, see foot traffic versus card panels.
Procurement for Hospitality Alt-Data Stacks
Scope pilots with property lists, date range aligned to seasonality, pre-registered segment definitions, and polygon samples for every flagged asset. Ask mobility vendors for false-positive methodology on hospitality sites: resorts, casinos, and airport hotels need custom footprint rules. Request spend panel composition disclosure before licensing. When ready, validate POI data on your property list with operational status and refresh dates before production foot-traffic joins.
Hospitality operators running competitive benchmarking should license polygon POI data for their comp set with brand hierarchy and a refresh schedule set per dataset: radius defaults at airport and urban hotel sites inflate visit indices and distort share calculations.
Frequently Asked Questions
Why does foot traffic diverge from reported hotel occupancy?
Geometry (radius vs polygon), dwell rules, seasonality, and property mix (street traffic vs guest parking) all affect observed visits. Polygon POI on hotel footprints reduces false positives versus centroid radius fences.
Is business travel recovered in 2026?
It has settled below 2019 levels by different amounts per subsegment (our reading, not a published figure): group and convention lag transient leisure. Single aggregate recovery metrics mislead operators and investors.
How should destination marketers measure CTV campaigns?
Join CTV exposure to devices observed in destination POI polygons during booking windows with documented match rates, aggregation floors, and holdout design, not DMA-level proxies alone.
What bias affects hospitality spend panels?
Carded spend skew, panel composition, and lag versus real-time mobility. Triangulate spend with foot traffic and document bias before binding to RevPAR or ADR models.
What belongs in a hospitality alt-data pilot charter?
Property list with polygon samples, segment definitions, seasonality window, pre-registered KPIs, false-positive methodology, and refresh cadence for POI and mobility feeds.
Why analyze hospitality at property level rather than brand level?
One flag can span airport, resort, and suburban assets with completely different demand drivers. Averaging them produces a figure describing no real hotel, and a brand-level move often reflects a change in portfolio mix rather than demand. Keep an open-and-close log so comparisons are same-store.
Which baseline year should a recovery chart use?
Whichever you can defend, named in the chart title rather than a footnote. Indexing to a pre-disruption year tells a recovery story, to the trough a growth story, and to last year a seasonality story. Where a decision hangs on it, show two baselines together.
How many data sources does a recovery claim need?
At least two independent ones that agree. Foot traffic sees presence but not spend, card panels skew by issuer and payment method, and booking data sees intent rather than attendance. When only one sensor moves, suspect the sensor and check whether the change starts on a specific date.
✓ Opt-Out Request Honored via Global Privacy Control
We use cookies and similar technologies to improve your experience. No marketing cookies are pre-selected. Learn more in our Privacy Policy. Your Privacy Choices.
