Travelers booked flights and hotel stays at the same pace in the third quarter of 2026 as a year earlier, but paid about 10% more each time, according to Criteo's Winter 2026 Travel Pulse. The 53-page report combines a September 2026 survey of 10,500 travelers in ten countries with transaction and browsing data that Criteo collects from what it calls "hundreds of travel providers." It carries no publication date. Criteo's web page for the report was last updated on October 29, 2026.

The findings arrive as travel sellers head into the period that Criteo's data identifies as the strongest for bookings. Last season, for most travel categories, that period began after Christmas rather than around Black Friday. Weekly bookings for packages and tours peaked 67% above their October level between late December and mid-January, and hotel and cruise bookings 43% above it. The report also tracks how quickly ChatGPT has entered the booking path: by September, 86% of Criteo's travel partners had recorded at least one booking referred by the chatbot, up from 71% in April.

In Short

Criteo, an advertising company that sees shopping data from hundreds of travel websites, published a report on how people plan and book winter trips. It found that over the summer quarter people booked about as many flights and hotel rooms as a year earlier but paid around 10% more for each one, and that almost nine in ten of its travel clients had received at least one booking from someone who came from ChatGPT. For travel sellers that means growth from pricier trips rather than more travelers, an AI channel that is spreading but whose size remains unknown, and a booking rush that, going by last year, lands mostly in January rather than around Black Friday.

Two data sets and several calendars

The report draws on two sources that are worth keeping apart, because they answer different questions and cover different periods.

The first is the Criteo Survey - Consumer Travel Preferences, fielded in September 2026 among travelers in the United States, the United Kingdom, France, Germany, Japan, South Korea, Canada, Australia, Brazil and Singapore. Most survey slides cite a base of 10,500. The opening page of the report describes "insights from over 10,000 consumers globally," and some charts use smaller bases: 9,938 for the generational questions, and 3,674 respondents in five markets (France, Germany, Japan, South Korea and Brazil) for a question on why travelers shift their dates. The report gives no margin of error, no weighting method and no fieldwork provider.

The second source is what Criteo labels "Criteo organic data," drawn from the travel websites where its technology runs. Some slides call it "Criteo commerce data" instead. The report does not define the "organic" label, does not say how many travel partners contribute, and publishes no absolute volumes. Every booking figure is either a year-over-year percentage change or an index against a baseline.

The baselines differ from chart to chart. The quarterly metrics compare July to September 2026 with the same months of 2025. The seasonal curves run from October 2025 to March 2026, indexed to the average for October 1 to 28, 2025 - meaning the "winter" demand pattern in the report is last winter's. Several traveler-type analyses use air bookings from just two weeks, December 1 to 14, 2025. The flight-distance analysis covers bookings completed between September 7 and 20 in both 2025 and 2026, and the booking-window analysis covers September 14 to 27, 2026. The seasonal charts carry a warning that applies throughout: "Verticals are indexed independently; the chart does not represent relative sales or booking volumes."

Q3 2026: flat bookings, bigger baskets

Across the three segments Criteo tracks - online travel agencies, air transport, and accommodation (hotels and resorts) - the third quarter produced a consistent pattern. Bookings for air and accommodation were unchanged year over year. Revenue in both rose 10%, because average order value rose 10%. "Growth is coming from more expensive trips rather than more bookings," the report states.

The online travel agency segment, usually shortened to OTA, behaved differently. Its traffic rose 5%, but its conversion rate fell 9%, so bookings dropped 4%. Average order value rose 5% and revenue just 1%. Air transport showed the reverse: traffic fell 5%, conversion improved 5%, and bookings held steady. Accommodation traffic rose 2% while conversion slipped 2%.

The global average hides wide regional differences.

The Americas: fewer visitors, much larger orders

In the Americas, air traffic fell 17% year over year in the third quarter, the sharpest decline of any segment in any region. Conversion on airline sites rose 16%, which limited the fall in air bookings to 3%. Order values did the rest. Average order value rose 26% for both air and accommodation in the region, lifting air revenue 22% and accommodation revenue 23% even though bookings in all three segments fell between 1% and 3%.

The Spring/Summer 2026 edition of the report, published in May, described a different Americas. In the first quarter of 2026, air traffic in the region was up 20% and air bookings up 17%, according to that edition. Six months later, traffic was down 17%. The two quarters are not seasonally comparable, but the direction of travel changed.

EMEA: OTAs draw traffic, hotels lose it

In Europe, the Middle East and Africa, OTA traffic rose 7% while hotel traffic fell 8% and air was flat. OTA conversion fell 10%, the steepest OTA decline of the three regions, leaving OTA bookings down 4%. Air was the only EMEA segment where bookings grew (up 3%), helped by a 3% gain in conversion. EMEA was also the only region where accommodation conversion improved, by 5%, although accommodation bookings still fell 3%. Order value increases were modest: 5% for OTAs, 2% for air and 4% for accommodation.

APAC: accommodation interest, air weakness

Asia-Pacific stood out for accommodation. Traffic to hotel and resort sites rose 22%, and although conversion fell 11%, accommodation bookings rose 8% - one of only two segments worldwide to grow bookings. Air went the other way. Air bookings fell 11%, with conversion down 12%, though average order value for flights rose 10%. APAC accommodation was the only segment anywhere in which average order value declined, by 1%. Air bookings in APAC had already fallen 12% in the first quarter, according to the spring edition, so the weakness has now lasted two reported quarters.

The Middle East after March

One slide tracks international flight bookings to Turkey, Egypt and the Middle East from February to September 2026, indexed to average daily bookings for February 1 to 7. According to Criteo, bookings fell "when the war involving Iran began in March," and all three destinations rebounded in early June.

The recovery since then has been uneven. Turkey "has sustained that recovery and mostly remains above its early-February level," the report says. The Middle East "has climbed steadily from its deep spring trough but still sits 20-40% below pre-war levels." Egypt's early-summer rebound faded, and by August its bookings were at around 40-60% of the February baseline, leaving Egypt and the Middle East at similar levels by September.

The spring edition, which covered the conflict's first weeks, put the decline in Middle East air bookings at as much as 78% against early-February levels. Neither edition says which countries make up the "Middle East" series.

Last winter as the template

Much of the report's seasonal guidance rests on a single season: weekly data from October 2025 to March 2026.

Globally, retail sales peaked around 65% above their October baseline in Black Friday week and stayed high through the Christmas shopping weeks. Travel bookings moved the other way, dipping below baseline and bottoming out just before Christmas. After the holidays, the lines crossed: travel bookings rebounded in January and "mostly" outpaced retail through March.

The regional versions of that chart differ in instructive ways. In the Americas, Black Friday lifted travel too, with travel bookings up 37% that week while retail sales doubled their October level; travel then peaked at 31% above baseline in January. In EMEA, retail also doubled on Black Friday, travel bookings fell to around 20% below baseline before Christmas, and January took them to 20% above. APAC followed its own retail calendar, with sales rising in waves around Singles' Day on November 11 and Double 12 on December 12 and peaking around 20-30% above baseline. Travel bookings in APAC stayed relatively steady before Christmas and reached around 18% above baseline in January.

Which categories surge

Category by category, bookings peaked between late December and mid-January. Packages and tours led at 67% above the October baseline, followed by hotels and cruises at 43% each. OTAs (24%), activities (23%) and air transport (18%) recovered more moderately. The appendix slide for packages and tours gives the January peak as 66% rather than 67%. The surge was short. By late March, most categories were back near their October level, and air transport sat 3% below it.

Comparing peak bookings with traffic in the same week shows where January demand converted. Packages and tours bookings rose 67% on a 50% rise in traffic, and hotels and cruises also booked faster than their traffic grew. OTAs and activities went the other way: traffic rose 37% and 31%, but bookings only 24% and 23%. For OTAs the gap persisted. By late March, OTA bookings were 2% above baseline while traffic was still about 13% above it, so "travelers keep exploring OTAs but fewer of those visits turn into bookings," according to the appendix.

The appendix also shows that Black Friday mattered for some travel categories and not for others. Hotel and resort bookings reached 70% above baseline in Black Friday week, well ahead of traffic at 23% above. Air transport showed "no Black Friday effect," with traffic up only 12%, and then fell hardest before Christmas, with air bookings 30% below baseline and air traffic 25% below. OTAs showed no Black Friday surge either and dipped about 15% below baseline in the weeks before Christmas.

Ski gateways and first-time visitors

A separate slide tracks flight bookings to "selected core European ski gateways," using partner-reported sales events indexed to the first four complete weeks of October 2025. Bookings stayed close to that level through autumn, peaked 53% above it in the first week of January, fell below baseline in February and reached around 40% below by early March. The report does not name the airports.

Criteo also measured the share of first-time visitors to travel sites, defined as visitors new to a given site among those it could identify, averaged across sites with each site weighted equally. Roughly two-thirds of visitors in any given week were new. The share peaked at 70-71% across the three regions between Christmas and early January, with a brief spike to 70% in the Americas around Black Friday. By late March it had settled near 67% in the Americas and APAC and 65% in EMEA.

In European air travel, 41% of bookings made between December 1 and 14, 2025, across seven European markets came from customers making their first recorded purchase on that airline's website. The qualifier matters: "first" means first in Criteo's purchase history, not necessarily the traveler's first purchase ever. "This highlights an opportunity for airlines to acquire new customers during the holiday travel season," the report states.

ChatGPT in the booking path

The report's newest figure concerns OpenAI's chatbot. Among Criteo's global travel clients with consistent ad spend from April to September 2026, the share that recorded at least one ChatGPT-referred transaction rose from 71% in April to 79% in May and June, 81% in July and 86% in August and September.

Criteo is careful about what that measures. "While this doesn't indicate the proportion of bookings generated by AI, it shows how widely AI-referred booking activity has spread across travel providers in a relatively short period," the report page states. A partner counts once it records a single referred booking. The report gives no share of total bookings, no revenue and no conversion rate for ChatGPT visits.

A second metric covers what happens before the booking. From April to September 2026, "ChatGPT's share of product page landings was 1.5 times that of search," according to the report. The slide does not define the calculation further. It appears to compare how often visits from each source land on a product page, rather than the volume of traffic from each source, so it says nothing about whether ChatGPT sends more visitors than search engines.

What the spring edition said

Comparing the two editions requires care, because Criteo changed the base. According to the Spring/Summer 2026 report, 72% of Criteo travel clients had recorded at least one ChatGPT-referred booking in March 2026, measured among travel clients with consistent ad spend since November 2025. The winter report's series starts at 71% in April, among clients with consistent spend from April onward. On paper, the share slipped by a point between March and April. In practice, the two figures come from different groups of clients and cannot be read as a decline. The rise from 71% to 86% happened within one consistent group.

The landing metric changed form too. In March, ChatGPT drove a share of product page landings 13 percentage points higher than search, according to the spring edition. The winter edition expresses the April-to-September comparison as a ratio of 1.5 times. The two numbers cannot be converted into each other without the underlying shares, which neither report publishes.

Criteo has used the same partner-count metric outside travel. In its CPG Pulse 2026 report, Criteo found that 86% of its consumer packaged goods clients recorded at least one ChatGPT-referred transaction in July. In travel, the July figure was 81%.

What travelers say about AI

The survey adds the consumer side. Globally, 47% of travelers said they would consider a destination they had not previously thought of if an AI assistant recommended it. The range across markets was wide, from 36% in Canada to 62% in Brazil, with the United States at 46%, the UK at 49% and South Korea at 54%.

As a source of inspiration, though, generative AI ranks well behind older channels. Asked where they typically get ideas for leisure travel, 50% of respondents named friends or family, 44% social media or short-form video, 39% search engines, and 31% each travel blogs and review sites, and travel booking websites or apps. Generative AI assistants, with ChatGPT and Gemini given as examples, came sixth at 29%. Brazil was the outlier at 48%. In Germany the figure was 22% and in Canada 20%. Japan relied on search engines more than any other market, at 56%. Online advertising was named by 18% globally, ranging from 9% in Japan to 27% in Australia, and podcasts by 8%.

On the question of where AI chat tools would be useful, respondents chose accommodation suggestions and destination ideas (46% each), activities and sightseeing (45%), dining (38%), air travel (33%) and full-trip planning (30%). Full-trip planning reached 48% in both Japan and South Korea, against 18% in the United States and 15% in Canada. The question measures perceived usefulness, not actual use.

The spring edition asked the same question in six of the same markets in February, which allows a market-by-market comparison. Destination ideas rose in Japan from 48% to 52%, in South Korea from 38% to 40% and in the UK from 44% to 46%, held at 45% in the United States and 27% in Germany, and fell in France from 41% to 39%. Full-trip planning fell in the United States from 22% to 18%, while it rose in Japan from 45% to 48% and in France from 29% to 31%. The global averages from the two editions are not comparable, because the September survey added four markets.

Where the destination shortlist stands

Only 15% of travelers said they have one destination in mind and do not consider others. A further 43% start with a preferred destination but compare it against others, 27% compare several with no clear favorite, and 15% start with no destination in mind. Japan had the highest share of single-destination travelers at 25%; South Korea and Singapore had the lowest at 10%. Half of travelers in Brazil (51%) and Singapore (50%) start with a preferred destination they are willing to swap.

Shorter hops, fewer borders

Across domestic and international flights booked between September 7 and 20, using the same air travel partners in both years, long-haul flights (more than 2,500 nautical miles) lost share in every region. The decline was 2.1 percentage points in APAC, 0.7 points in EMEA and 0.4 points in the Americas. Medium-haul flights (750 to 2,500 nautical miles) gained 0.4 points in the Americas, 0.5 in EMEA and 0.6 in APAC, and short-haul flights (under 750 nautical miles) gained 1.5 points in APAC.

The share of international bookings fell 1.5 percentage points in EMEA and 1.2 points in APAC and was unchanged in the Americas. Within international travel alone, medium-haul gained 1.4 points in EMEA and 1.0 point in the Americas. APAC shifted toward short-haul international flights (up 1.8 points) as long-haul fell 2.3 points.

The spring edition, using bookings from April 1 to 14, had found APAC long-haul down 2.8 points and APAC's international share down 4.7 points, so the Asia-Pacific pull toward closer destinations shows up in both editions. EMEA moved the other way between the two snapshots: its international share rose 1.4 points in the April window and fell 1.5 points in the September window. The Americas, which shifted toward long-haul international trips in April (up 2.2 points), lost long-haul share in September.

Longer planning for longer trips

For flight, hotel and tour package bookings completed between September 14 and 27, 2026, the average stay reached 4.4 days, up 13.6% year over year, and the average anticipation window reached 44.9 days, up 12.8%. On PPC Land's calculation, that implies an average stay of roughly 3.9 days and a window of roughly 39.8 days in the same weeks of 2025.

The window stretches with the length of the trip. Criteo's chart shows travelers searching 92 days in advance for stays of eight nights or more (average duration 16.9 days), 70 days for four to seven nights, 36 days for two to three nights and 18 days for a single night.

The spring edition, based on bookings from April 1 to 14, 2026, gave 86 days for eight-plus nights and 29 days for one night, and reported that average lead time had fallen 7% year over year to 52.7 days. The winter chart is labeled as an "advance search window," while the spring edition described planning lead time, so the two sets may not measure exactly the same interval. Either way, the year-over-year direction differs: windows shortened in the April snapshot and lengthened in the September one.

Sun-seekers, niche destinations and parents

The third part of the report turns to traveler types, mostly using air bookings from December 1 to 14, 2025.

Sun-seekers

Travelers flying from cold US cities, defined as airports averaging 10 degrees Celsius or below in December, were 35% more likely to choose mild destinations (averaging 15-24 degrees in December, Miami being Criteo's example) in December than in June. For tropical destinations averaging 24 degrees or above, such as Cancún, the increase was 17%. These December trips were around 4% shorter, and airfares on the same routes were 6% lower for mild destinations and 9% lower for tropical ones than in June. Climate data came from NASA POWER.

Domestic warm destinations gained more than international ones: 41% against 19%. Fort Myers led US destinations at 119%, followed by Miami (59%), Tampa (39%), Orlando (36%) and Fort Lauderdale (35%). Internationally, Puerto Vallarta (138%) and Dubai (135%) led, followed by Los Cabos (97%), Cancún (58%) and Mumbai (43%).

The Dubai figure needs context the report does not give. The comparison period is June 1 to 14, 2026, which falls after the start of the war that, by Criteo's own chart, cut bookings to the Middle East. A December figure compared with a depressed June would show a larger increase for reasons unrelated to winter sun.

Niche destinations

Travelers flying to niche destinations - those outside the top 20% of most-visited locations - travel 1.7 times farther and spend 1.6 times more than those choosing mainstream destinations, according to the slide that presents the analysis, which covers air bookings from December 1 to 14, 2025, in the US, Germany, France, the UK, Japan, South Korea, Brazil and Canada. The report's opening summary gives different figures: 1.6 times farther and 1.5 times more. Criteo's web page uses 1.7 and 1.6 in its text but describes the spending gap as "1.6x higher flight spend" in its headline statistics.

The same two ratios, 1.7 times farther and 1.6 times more spend, appeared in the Spring/Summer 2026 report, where they were based on air bookings from July 1 to 7, 2025. Identical ratios from two different booking windows are possible, but the report does not address the overlap. A year earlier, a Criteo-sponsored summary of the Winter 2025 Travel Pulse published by Skift said travelers to less-visited destinations spend up to three times as much as mainstream vacationers.

Parents

Criteo identifies parents as travelers who bought from parent-related retail categories in at least two different months and two different categories between September 1 and November 30, 2025. On that definition, US parents flew approximately the same distance as non-parents in early December but paid 13% less per booking and 25% less per mile. "This suggests that parents may be more price-conscious or deal-driven when booking flights, rather than simply choosing shorter trips," the report states.

European parents showed the opposite pattern: 8% shorter flights, 20% more per booking and 25% more per mile. The per-mile figures do not follow from the other two by simple division - 20% more for 8% less distance would work out at about 30% more per mile on PPC Land's calculation, and similar distance at 13% less would give about 13% less per mile in the US. The figures may be averages of individual bookings rather than ratios of averages; the report does not say.

Parents were also more likely than equally active non-parents to book with an airline for the first time. The lift in new-to-partner booking share was 59% in the US, 39% in Brazil, 32% in South Korea and 16% in both Europe and Japan. Among parent travelers who also bought from specialty retailers in December 2025, US penetration was highest for apparel and accessories (73%), home and garden (54%) and health and beauty (52%), with toys and games at 45%. European penetration was flatter and lower: home and garden 36%, apparel 35%, health and beauty 34%, and toys and games 9%. Criteo's European analyses cover seven markets on two slides and nine on this one.

Budgets under pressure

Travel remains a priority for many respondents. Asked how essential travel is compared with categories such as fashion and beauty, 53% of Millennials, 50% of Gen Z and 44% of Gen X called it very or absolutely essential. Rising costs are affecting travel plans for more than 80% of each generation, according to the report. The chart's legend is ambiguous - it shows "Yes" and "A lot & Somewhat" as separate bands - but the two together come to 81% for Gen X, 82% for Millennials and 83% for Gen Z, with 17% to 19% answering no.

The most common ways to manage costs were booking far in advance (42%), traveling off-peak (40%), choosing cheaper destinations (35%) and choosing budget accommodation (35%). Off-peak travel was most popular in Singapore (52%) and South Korea (49%). Shorter distances, shorter trips and comparing three or more providers each drew 26%; loyalty programs and credit card rewards, 22%; travel packages, 19%.

The top three figures match those the spring edition reported from its February survey exactly: 42%, 40% and 35%. That survey covered six markets with a base of 6,379, while the September survey covered ten markets with 10,500 respondents. Identical results from two different samples are possible, but the report does not comment on it.

Overall, 69% of travelers said they were willing to adjust their dates to avoid the busiest periods, from 58% in Germany to 82% in Brazil, and 63% said they monitor geopolitical events such as wars or tense elections when planning. Among the five-market subsample willing to move dates, 59% cited lower prices, 55% less crowded destinations and attractions, 45% more comfortable weather, and 43% more choice of flights and accommodation.

What Criteo recommends

The final section of the report promotes Criteo's own products. Criteo recommends using video to build awareness "before the holiday season ends," expanding presence into AI discovery channels "beyond traditional search and social," using CTV to showcase destinations, and reaching traveler types through Criteo travel audiences. Its web page adds that "display and app advertising help re-engage interested travelers as booking intent strengthens."

That framing reflects Criteo's position. The company sells travel audiences, retargeting and, since March, access to advertising in ChatGPT. Its partner count for ChatGPT-referred bookings also serves as evidence for a channel it is selling. The underlying data, though, is published in enough detail - and with enough internal inconsistencies flagged above - for readers to judge it on its own terms.

What the report does not show

Several things are missing that would change how the numbers read. There are no absolute booking or revenue figures, so a 26% rise in order value in the Americas cannot be translated into dollars, and the relative size of the Americas, EMEA and APAC is unknown. The number of travel partners behind each chart is not given, and the panel for the ChatGPT metric is not sized. The survey has no stated margin of error.

The report also does not separate price from mix. A 10% rise in average order value could reflect higher fares and room rates, longer stays, more travelers per booking or a shift toward pricier routes, and the report's own data points in more than one of those directions: stays lengthened 13.6% in the September window, while long-haul flights lost share. Criteo does not break the increase down.

And the winter forecasts are, in effect, last year's pattern. The January surge figures describe the 2025-26 season; the report does not project the 2026-27 season.

Why this matters for the marketing community

For travel advertisers, the clearest commercial signal in the report is the gap between traffic and conversion. OTAs gained visitors and lost conversion in all three regions in the third quarter, and last season's OTA traffic stayed ahead of bookings well into March. That pattern fits a market in which travelers browse more widely before buying - 85% of survey respondents do not fix on a single destination - and in which the source of a visit matters more than its volume.

The search and AI context has moved quickly this year. Google folded hotel, flight and other travel formats into standard Search campaigns when it announced Search Campaigns for Travel on April 30, 2026, and in August Google AI Mode began completing hotel bookings with 10 partners in the US, including Booking.com, Expedia, Marriott and Hilton. On the programmatic side, Expedia's 200 petabytes of traveler data became available through Magnite after an April 16 announcement. And on October 21, the European Commission opened proceedings that put Booking.com under EU investigation over scams and boosted listings that may not be labelled as ads.

Criteo's role in ChatGPT advertising gives its travel numbers particular weight in that discussion. It became the first ad tech partner in OpenAI's ChatGPT ad pilot in March 2026, passed 2,000 brands advertising on ChatGPT by June, and OpenAI extended ChatGPT ads to 89 countries on October 21. The travel report does not say whether any of the ChatGPT-referred bookings it counts came from ads or from organic answers.

Measurement remains the weak point. Referral counts capture only visits that arrive with a ChatGPT referrer. Research by Similarweb found that 55.9% of AI-influenced traffic arrives via search rather than directly from the assistant, which means AI referral traffic as recorded by a site's analytics can understate the assistant's role. Consumer surveys also point to limits on how far travelers will let AI go. In a HUMAN Security survey, 54% of Americans were comfortable using AI to plan a vacation, but only 12% with fully autonomous booking, and a Fullstory survey found 15% of US travelers using AI tools early in trip planning, with 61% citing hidden fees as their top frustration. The questions differ, so the figures cannot be lined up directly, but Criteo's 29% who draw inspiration from generative AI assistants falls between them.

Budgets, at least among UK travel marketers, were expected to grow: 64% expected to raise marketing budgets in a survey reported in March. Criteo's report suggests that, on its partners' sites, the money is being chased by travelers who book about as often as last year but spend more each time - and who, if last season repeats, will do much of that booking in January rather than around Black Friday. Criteo's own business has been under pressure this year, with the company having cut its full-year guidance as second-quarter revenue fell 11% to $428 million.

Timeline

Summary

Who: Criteo, the advertising technology company, which published the Winter 2026 Travel Pulse using data from travel websites where its technology runs and a survey of 10,500 travelers. The findings concern airlines, hotels, online travel agencies, package and tour operators, cruise lines and the marketers who buy advertising for them.

What: In the third quarter of 2026, air and accommodation bookings on Criteo's travel partners' sites were flat year over year while average order values and revenue rose 10%; OTA traffic grew but conversion fell 9%. By September, 86% of Criteo's travel partners had recorded at least one ChatGPT-referred booking, up from 71% in April. Last season's data shows travel bookings peaking in January, led by packages and tours at 67% above the October baseline. The report contains several internal inconsistencies, including two sets of figures for niche-destination travel, and some figures match its spring edition exactly despite different samples.

When: The report is undated; Criteo's web page for it was last updated on October 29, 2026. The survey was fielded in September 2026, the quarterly data covers July to September 2026, and the seasonal patterns cover October 2025 to March 2026.

Where: The survey covers the US, UK, France, Germany, Japan, South Korea, Canada, Australia, Brazil and Singapore. The commerce data is reported for the Americas, EMEA and APAC, with separate analyses for Europe, the US and the Middle East.

Why: Travel advertisers plan winter campaigns around when travelers book, how much they spend and where they find ideas. The report shows growth coming from higher spending per booking rather than more bookings, a booking peak that last season came in January for most categories, with hotels the main Black Friday exception, and a rapid spread of ChatGPT-referred bookings whose share of total bookings Criteo does not disclose.