The one that separates
Accommodation nights are the rare consumer measure that records what happened rather than what someone intended: nights actually spent in accommodation establishments, counted in the official frame. Cadence publishes a finding on them in all eleven markets this quarter, which makes Travel the only sector that classifies everywhere.
In ten markets nights and arrivals moved the same way. In Italy they did not: nights rose 1.7% year on year while arrivals fell 2.7%. Italy’s average stay, published separately in the same edition, lengthened by 0.18 nights.
Fewer stays began. More nights were spent. Both are true, and they are two measurements rather than one story.
The rest of the set
| Market | Nights | Arrivals |
|---|---|---|
| Netherlands | +7.4% | +8.3% |
| Austria | +6.4% | +10.4% |
| Sweden | +5.7% | +6.3% |
| Poland | +5.5% | +7.5% |
| Germany | +3.5% | +4.8% |
| France | +3.0% | +5.2% |
| Portugal | +1.8% | +0.4% |
| Italy | +1.7% | −2.7% |
| Belgium | +1.5% | +3.6% |
| Spain | +0.5% | +1.7% |
| Ireland | −2.4% | −1.0% |
Ireland is the only market where nights fell. Austria has the largest rise in both. Portugal and Italy are the two markets where nights grew faster than arrivals, and Italy is the only one where arrivals went backwards while nights grew.
Why the distinction is commercial
Growth in nights and growth in arrivals are different problems to sell against. More arrivals at a similar length of stay is a volume question — reach, routes, conversion. The same nights from fewer, longer stays is a composition question — mix, length, what is worth building a package around.
A single growth figure for European travel would have averaged a market gaining visitors against a market keeping them longer, and produced a number true of neither.
Same direction on the headline. Opposite problems underneath it.
What this does not establish
Accommodation activity is not travel demand, intent, interest or search, and nothing here reads any of those from it. Nights and arrivals are reported as two measures and are never combined into one; where average stay appears it is published in its own right rather than derived here. No relationship between these figures and any other signal is claimed, and none of it is a forecast.
Source: Eurostat tour_occ_nim and tour_occ_arm, July 2026 against the same month a year earlier. A read, not a forecast.