The day Spain and Portugal went dark

A one-day hole in the booking curve

We compared the reservations our 196 mainland properties received each day with the same weekday one year earlier. Across four weeks around the blackout, 2025 ran between 12% below and 26% above 2024. Then, on 28 April, bookings fell to 17% below the year before, the weakest day of the 28 we studied.

A drop on one day could be a quiet Monday anywhere. To rule that out, we compared it with properties that did not lose power. Against the two previous Mondays, bookings received fell 17% in mainland Spain and 11% in mainland Portugal. In the rest of Europe they fell 2%. That leaves a gap of about 14 percentage points that the blackout explains. With a baseline that avoids the Easter weeks, the gap widens to 19 points. On a normal day the two groups move within about 9 points of each other, so 28 April sits more than two standard deviations outside the usual range.

The bookings came back within two days

The next morning, the hole started to fill. On Tuesday 29 April our mainland properties received 20% more bookings than on the same Tuesday in 2024, and Portugal alone was 29% above its previous Tuesdays. Some of those were travellers who had tried to book the day before. Others were reservations taken by phone or on paper during the outage and entered once systems were back.

Adding up the days tells the story. After adjusting for the rest of Europe, the blackout cost our mainland properties about 369 bookings on 28 April. By the end of Wednesday the cumulative gap had shrunk to 127. By Sunday 4 May it had closed completely.

Revenue is the one place where the recovery looks incomplete. Booking revenue for the week ran about 6% below expected, even though the number of bookings caught up. It is possible that recovered bookings were shorter or cheaper, but the signal is weak and we would not lean on it.

Local guests pulled back hardest

Where we know a guest's nationality, Spanish guests booking Spanish properties fell 31% on 28 April. Everyone else fell 16%. In Portugal, domestic bookings almost halved, though on small numbers. Both groups rebounded the next day.

This may also explain a surprise. The Canary Islands kept their power, yet bookings there fell 19%, as much as on the mainland. Many guests of island properties book from mainland Spain, and on 28 April many of them had no power, no internet and other things on their minds. A shock can travel through a market's customers, not only through its buildings.

Every channel paused

We expected direct bookings to suffer most, since a hotel's own booking engine depends on its connection. That is not what happened. Every channel fell on 28 April, and every channel except the smaller online agencies bounced back the next day. The largest jump came from bookings entered at the front desk or by phone, which more than doubled on 29 April.

What it means for hoteliers

• Treat a shock like this as delayed demand, not lost demand. Cutting prices on 28 April would have given away rate on bookings that were coming back anyway.
• Judge the next 48 hours, not the day itself. The rebound on 29 April closed about half of the gap in one day.
• Keep an offline way to take reservations. A phone line, a paper log and a plan to enter them quickly protected bookings that would otherwise have gone elsewhere.
• Separate local and international demand in your forecasts. Domestic guests reacted twice as strongly, and the effect reached properties that never lost power.

A sudden drop in booking pace with a clear outside cause is a signal to read in context. Pricing that reacts to the drop alone would have discounted into a recovery.

Method and limits

We counted reservations by the date they were received, from 7 April to 25 May 2025, at 144 mainland Spanish and 52 mainland Portuguese properties with at least one booking in the window. Controls were 164 properties in the UK, Ireland, France, Germany, the Netherlands and Belgium, plus 22 in the Canary Islands and Azores. The year-on-year comparison aligns days by weekday (364 days earlier). Cancelled bookings are excluded unless stated.

• Our portfolio leans towards hostels and independent hotels, so these results describe that segment, not the whole Spanish and Portuguese market.
• Results depend on the baseline: −15% against the two previous Mondays, −21% against a baseline that avoids Easter. Both are well outside normal daily variation.
• The effect was uneven: some properties lost 30–40% of bookings that day, others about 10%.
• We saw extra cancellations for stays on 28–29 April, but they were concentrated in a small set of properties and look like cancelled group blocks. We do not draw conclusions from them.
• Easter (20–21 April), Labour Day (1 May) and the Madrid holiday on 2 May sit close to the event.