Journal · 12 April 2026
When a hotel occupancy forecast should ignore last August
A full house in a hot August will bully every cell that sits near it. The boulevard does not owe you a repeat.
Last August is the month directors remember. It is also the month that ruins a forecast if you let it sit in the average like a quiet neighbour.
On the Batumi boulevard a hot, dry August fills rooms that a wet Tuesday in July will not. If you fold that August into a single “summer factor,” the model will ask the kitchen to staff as if rain were a rumour. We have sat in rooms where last year’s record Saturday was still being used to justify an allotment that the night audit had already contradicted.
What to keep from a famous month
Keep the shape: which weekdays paid, which room types moved, whether walk-ins or buses did the work. Throw away the level if the weather, the festival calendar, or the border traffic is not the same animal this year.
A predictive occupancy curve for the next ninety days is safer when famous months are labeled as scenarios, not as the spine. One scenario can still be “August repeats.” The other two should be allowed to be ordinary.
A practical test
Print last August and last July on one sheet, by weekday. If July looks like a different trade, it is a different trade. Do not average them into a personality called Summer. Bring both sheets to a scoping visit. We would rather argue with two honest months than decorate one legend.