Demand Calendar Blog by Anders Johansson

Sold Out Is Not Demand. Your Forecast Records It as Both.

Written by Anders Johansson | 25 August 2026

Constrained dates are the one part of a hotel forecast that cannot correct itself. Rooms sold stops at capacity, so demand above capacity never enters the history next year's numbers are built from. What follows covers how the gap forms, what revenue management research says it costs, and three ways to reopen your own baseline before budget season closes.

Your Forecast Data Stops at 100%

Rooms sold is a capped number. On a night you sell out, your property management system records 180 because you have 180 rooms, not because 180 people wanted one. Guests who search after that point leave no trace in the file your forecast learns from.

Run the measurement on your own data. Pull last year's actuals, flag every date that reached full occupancy, then add the dates where you closed a segment or set a minimum stay. The percentage you get is the share of your demand history that records capacity instead of demand.

Across the rest of the year, your forecast improves every time it misses. On those dates it cannot miss, because the number it is checked against was set by your room count.

Budget, Forecast, and Plan Share One Input

Three documents run a hotel year, and the distinction between them is worth defending. The budget is a commitment: what the hotel agrees to deliver. The forecast is the current truth: where the hotel is actually heading. The business plan is the action that closes the gap between them.

All three descend from the same actuals. Get the distinction right and the input wrong, and you have three well-governed documents resting on a demand history with a hole in it, on precisely the dates that matter most.

In June, nobody asks whether the baseline was censored. They ask why the forecast missed.

The Bias That Compounds Every Year

Censoring does not stay still. A capped baseline produces a conservative rate, the conservative rate sells out again, and the sellout confirms the cap. Revenue management research calls the result a spiral-down effect: expected revenue declines steadily over time when the censoring is ignored.

The measured figures come from airline revenue management, where the problem was studied first (Guo, Xiao and Li, Advances in Operations Research, 2012, a survey of more than 130 studies). A forecast carrying a negative bias can cost up to 3 percent of potential revenue. Underestimating demand by 12.5 to 25 percent costs 1 to 3 percent of revenue on high-demand dates. Studies on real booking data attribute revenue gains of 2 to 12 percent to unconstraining the history properly.

Airlines and hotels do not share a cost structure, and those percentages do not transfer directly. The mechanism does. Any system that stops recording at capacity teaches itself that capacity was the answer.

What the Gap Costs on Twelve Nights

Take the arithmetic on a 180-room hotel with twelve sold-out nights a year. Assume the market on those dates carries 12 euros more per room than the rate your baseline recommended. The room revenue difference is 180 rooms by 12 euros by 12 nights, or 25,920 euros.

Rate is where the number stops looking small. Revenue from an increase in rate carries almost no additional cost, so flow-through to profit on that line runs at roughly 90 to 100 percent. Nearly the whole 25,920 euros reaches GOP.

The figures above are illustrative, and your own room count and rate move them. The direction does not move. On a constrained date the occupancy variable is already spent, so rate is the only place the error can hide.

Three Ways to Reopen Your Baseline

  • Count the constrained dates before the budget is built. Flag every date last year that reached full occupancy, closed a segment, or carried a minimum stay. The share those dates represent is the share of your baseline that is not a demand number, and it is a figure you can put in front of your GM in one sentence.
  • Convert denied demand with your own ratio, never with raw search counts. Take the look-to-book ratio your booking engine already reports and apply it to the no-availability events on those dates before you add anything to rooms sold. One guest generates several searches across several devices, so raw counts inflate the estimate, and an inflated estimate is the fastest way to lose the argument you are trying to win.
  • Write the rate assumption down while the budget is still open. Record which dates you expect to constrain next year and what rate you believe they carry. When the variance question arrives in June, your reasoning is already on the record, with a date on it.

Your Best Call of the Year

The uncomfortable part of a constrained date is not the money. Your strongest judgment of the year leaves no evidence behind it. You argued for a higher rate, the market proved you right by selling out, and the file recorded a result identical to the one a lower rate would have produced.

Demand Calendar is a total-revenue forecasting and profit system for hotels. It sits alongside your RMS and PMS, not instead of them.

A baseline that stops recording at capacity is not conservative. It is wrong in one direction, on your strongest dates, every year, and no variance report will find it, because the number it measures against inherited the same blind spot.

See what your own constrained dates cost before this budget closes. Book a strategy call and we will run the count on your last twelve months.

BOOK A STRATEGY CALL → demandcalendar.com/book-a-call