We compared the 2025 Iberian blackout with the 2024 CrowdStrike outage and the 2023 UK rail strikes. Only one of them left a clear mark on hotel bookings.
When something big goes wrong, hotels brace for lost bookings. In our study of the Iberian blackout of 28 April 2025, they were right: bookings at our mainland Spanish and Portuguese properties fell by about a sixth that day, then came back within 48 hours.
We wanted to know whether that pattern holds for other shocks. We picked two more that made headlines across Europe. The first is the CrowdStrike outage of Friday 19 July 2024, when a faulty software update took down airline, bank and hotel systems around the world. The second is the series of national UK rail strikes in 2023. We ran the same kind of test on each: compare the affected day with normal days, and with places the event did not touch.
The short answer: only the blackout left a clear mark. The CrowdStrike outage did not measurably change bookings at our properties. The rail strikes may have trimmed London stays slightly, but not by more than an ordinary bad week. What matters is not how big the headline is, but whether the people making bookings lose the ability to do it.
The blackout is our reference point. Power failed across mainland Spain and Portugal at 12:33 and came back through the night. People could not use their phones, laptops or payment cards for hours, and bookings fell with them.
Bookings received by our 196 mainland properties fell 15% against the two previous Mondays, while the rest of Europe fell 2%. After removing that background, the blackout took out about 14 percentage points of the day's bookings. On a normal day the two groups move within about 9 points of each other, so the gap is well outside the usual range. By the end of the week, the missing bookings had been made up.
Early on Friday 19 July 2024, a faulty update from the security company CrowdStrike crashed millions of Windows computers. Airlines grounded flights, banks and broadcasters went offline, and some hotels lost their front-desk systems for part of the day.
We looked at bookings received by 171 European properties and 47 in Canada. The raw numbers suggest a dip: UK bookings were 16% below the four previous Fridays and Portugal's 19% below. But late-July Fridays were just as weak in 2023, by 11% and 15%. Once that seasonal pattern is removed, the UK is about 5 points below normal and Portugal about 6, well inside their normal day-to-day swing of about 13 points. Across Europe, bookings on the day were slightly above expectations. Canada was flat.
We checked channels too. Bookings via Booking.com, Hostelworld and properties' own websites were each 5% to 8% lower than on the two previous Fridays, which is within normal noise. Bookings keyed into one property-management system rose by about 30%, which fits staff entering reservations by hand, but on small numbers. Stays did not change either: the share of UK bookings for 19 and 20 July that were later cancelled or never arrived matched normal weeks.
The reason is simple. The outage hit back-office systems and airports. It did not stop travellers using their phones and the big booking sites, which kept working.
Rail strikes are a different kind of shock. They are announced in advance, they last a day or two, and they make it harder to reach a hotel rather than to book one. We looked at nine national strike days in 2023 between June and October and measured room nights stayed at our UK properties against the same weekday two weeks either side.
Regional properties in England and Wales, where we expected rail to matter most, showed no effect: stays averaged 1.6% below normal on strike days, against 0.4% above on other days. London properties were 10% below normal on strike days, or about 7 points below Scotland and Northern Ireland, which the strikes did not affect. That sounds meaningful, but the variation from one strike day to the next is large, and the biggest drops fell in weeks that were weak in London anyway, including non-strike days.
We also found no rise in cancellations for strike days. Guests seem to have planned around the strikes rather than abandoning their trips.
Put side by side, the three events show a clear pattern. The blackout hit the people making bookings: no power, no signal, no way to pay. That left a hole in the day, and the hole was filled within the week. The CrowdStrike outage hit companies' systems, not travellers' phones, and booking flows carried on. The rail strikes hit the journey, not the booking, and travellers worked around dates they knew about in advance.
Our properties are mostly hostels and small independent hotels, so these results describe them rather than the whole market. Booking dates are recorded in each property's local time, and bookings delayed by a system outage appear on the day they were finally entered. For the strikes we used nine dates we could confirm, and four properties in Scotland and Northern Ireland as the comparison, which is a small group. The London result is about 1.2 standard errors from zero, so it is weak. We did not include the 2023 French pension strikes: we have too few properties in Paris.