Most hotels grade every forecast with one ruler: how close it lands to actuals. The ruler fits the next 30 days and breaks beyond them, because a longer-range forecast exists to trigger action, and action changes the outcome it predicted. Here are the three jobs of a hotel forecast, the measure that fits each one, and the habit that turns a missed forecast into evidence of revenue you created.
Hotel Forecasting Has Three Jobs
A hotel forecast is one number doing three different jobs. The distance to the arrival date decides which job it is doing, and the job decides what a good forecast looks like. The full forecasting cycle is set out in the hotel forecasting guide; the three horizons are what matter here.
Inside 30 days, the forecast is operational. You execute against it: staff rotas, purchasing, restaurant covers and housekeeping hours. Between 30 and 120 days, it is tactical. You adjust with it and point sales and marketing efforts at the dates that need them. Beyond 120 days, it is strategic. You act on it with groups, contracts, campaigns, and next year's commercial plan.
Most hotels read all three horizons with the same ruler. The ruler is forecast accuracy, and it only fits the first horizon.
Inside 30 Days, Accuracy Is the Job
In the operational window, most of the business is already on the books, and little can still change. Nobody launches a campaign for next Tuesday. The forecast tells the chef, the executive housekeeper, and the front office manager how many guests will arrive, so they plan the right hours and stock.
Here, accuracy is the whole point. The measure is mean absolute percentage error (MAPE): the average daily gap between forecast and actual, as a share of actual. A target worth setting for the next 30 days is a MAPE of 2-3% for occupied rooms. A miss inside that window costs real labor hours and real stock, so tighten the model until it holds.
Every revenue manager accepts that ruler. The trouble starts when the same ruler travels out to day 90.
Beyond 30 Days, a Hit Proves Nothing
Take a 200-room hotel. Ninety days before a week in March, the forecast shows 812 room nights for the seven nights, or 58 percent occupancy. The same week last year closed at 71 percent. You flag the gap in the commercial meeting. Marketing moves budget into those dates, and sales calls two corporate accounts with projects in the area.
The week closes at 966 room nights, or 69 percent. Measured on accuracy, your day-90 forecast missed by 154 room nights, an error of 15.9 percent. On the variance report, it is the worst forecast of the quarter.
Now picture the forecast landing exactly on 812. The most reliable way to get there is for nobody to act on it. Beyond 30 days, a forecast that lands on actuals usually means nobody acted on it. Accuracy at that distance rewards the revenue manager who stays quiet and penalizes the one who raises the flag.
The forecast did its job. It moved sales and marketing into the right week while there was still time to fill it. Graded on accuracy, the success goes on record as a failure.
Three Horizons, Three Measures
Grade every horizon with the near-term ruler and the far horizons fail every time. Most teams respond the only way the ruler allows: they stop following up on the long-range forecast at all, and the forecast stops learning.
Each horizon needs its own measure, and each miss needs its own follow-up question. The question matters as much as the score, because it decides what the next forecast learns.
- Operational, 0 to 30 days: measure accuracy, with a MAPE target of 2 to 3 percent. Ask whether the miss came from volume or from price.
- Tactical, 30 to 120 days: measure lift, meaning actual revenue against the forecast as it stood before anyone acted. Ask which factor the forecast missed, and whether it was internal or external.
- Strategic, 120 days and beyond: measure outcome, meaning revenue against the path where the hotel changed nothing. Ask what changes in the next forecast, specifically.
A meeting that asks why the forecast was wrong gets a defense. A meeting that asks what changes in the next forecast gets an action. The tactical measure is the one most hotels never run, and it needs no new system. It needs one number that most revenue managers delete every week.
Save the Forecast Before You Act
Lift is the actual result minus the untouched forecast. In the March example, the lift is 154 room nights. At an average rate of €150, the lift is €23,100 in room revenue the forecast helped create.
The figure only exists if the day-90 forecast still exists. Most revenue managers overwrite the forecast at every update, which is the right habit for running the hotel and the wrong one for proving anything. Once the day-90 version is gone, the baseline is gone, and the €23,100 becomes an opinion.
Not all 154 room nights were attributable to the actions. Some would have arrived anyway, and a careful finance director will say so. The answer is not to claim every night. The untouched forecast is the only baseline written before anyone knew the outcome, making it the most honest available. Log each action next to it, with a date, and the attribution discussion runs on evidence instead of memory.
The saved forecast also answers the question that every commercial meeting circles around: which actions actually made money? Whether the hotel acted on your forecast at all is a separate measurement, set out in Your Hotel Forecast Was Right. The Month Still Missed. The saved forecast tells you what the action was worth.
Grade the Forecast on Its Job
Start with your own numbers. Pull last quarter's day-90 forecasts, if any version survives, and compare each one with the actual result. Then mark the weeks where someone acted in between. The weeks with the largest misses are often the weeks where the forecast did the most work.
From next Monday, save a copy of the forecast before each commercial meeting and log every action it triggers, with a date. In one quarter, you have a record of how much your forecasting is worth, measured on the job for each horizon it actually covers.
Demand Calendar is a total-revenue forecasting and profit system for hotels. It sits alongside your RMS and PMS, not instead of them.
A revenue manager graded on 90-day accuracy learns to stay quiet about the weeks that need help. A revenue manager graded on lift learns to raise the flag early, and raising the flag early is the reason a hotel forecasts at all.
BOOK A STRATEGY CALL → demandcalendar.com/book-a-call
