A useful hotel forecast helps you plan room pricing, staffing and purchasing. Improving accuracy starts with consistent data and a repeatable review process, rather than a promise that software can predict every change in demand.
For the wider pricing context, read our practical hotel revenue management guide. This article focuses on forecasting: what to measure, how to compare predictions with results, and when to revise your assumptions.
1. Define the forecast before choosing a method
Decide whether you are forecasting rooms sold, occupancy, room revenue or another measure. Record the stay date, forecast date, available room count and reporting basis. A forecast made 30 days before arrival should be compared with other 30-day forecasts, rather than a prediction made the night before.
Keep forecasts separate from budgets. A budget is a target; a forecast is your current estimate of the likely outcome. Record changes and the information that prompted them.
2. Check data and compare equivalent demand periods
Review cancellations, duplicate bookings, room blocks, out-of-order rooms and segment coding. Separate confirmed reservations from tentative group enquiries. Use consistent treatment of taxes and inclusions when forecasting room revenue.
Start with a simple baseline, such as comparable weekdays or a similar event period. Then consider what has changed: available inventory, local events, group commitments, renovations or marketing activity. Last year’s calendar date may be a weak comparison when holidays fall differently.
Understand booking pace and pickup
Booking pace compares reservations for a future stay date at equivalent lead times. Pickup measures bookings added between snapshots; specify whether the figure is gross or net of cancellations. A date with few bookings far in advance may still sell well if that property’s guests usually book late.
Illustrative example: a 30-room motel has 18 confirmed room nights for a future date. Between two snapshots, six room nights are added and two cancelled. Net pickup is four room nights, bringing the total to 22, assuming no other adjustments. This is an explanation of the calculation, not a demand forecast or customer result.
3. Measure forecast errors consistently
Save each forecast before the outcome is known. Compare it with actual results at fixed horizons, such as 7, 14 and 30 days before arrival. Do not overwrite old predictions and then measure only the latest version.
Mean absolute error (MAE) is the average of the absolute differences between forecasts and actual results. For rooms sold, it gives an understandable answer in room nights. If you forecast 24 rooms and sell 21, the absolute error is three rooms.
Mean absolute percentage error (MAPE) averages those absolute errors as percentages of actual results. For each period, calculate |actual − forecast| ÷ actual × 100, then average the percentages. It is undefined when actual results are zero and can be unstable when actual results are very small. These definitions and limitations are explained in Forecasting: Principles and Practice by Hyndman and Athanasopoulos.
There is no single accuracy percentage that this article promises for every hotel. Compare methods against your own baseline at the same horizon and investigate whether you consistently over-forecast or under-forecast.
4. Use software with clear responsibilities
Software can help organise property data, monitor booking patterns and support rate decisions. Confirm the exact reports, integrations and data sources available before relying on a supplier’s forecast.
Hotel RevBoost analyses property performance, booking pace, competitor rates and market information to support pricing recommendations. Operators can review recommendations or configure automated rate updates through supported PMS connections. Confirm suitability for your specific property and system during a demonstration.
Forecasting informs pricing, but a pricing change is a separate decision. Our hotel dynamic pricing article explains the relationship.
5. Review room types, segments and events
Where your data supports it, review room types and meaningful business segments separately. Group bookings, contracted business and transient demand can have different lead times and cancellation patterns. Validate those patterns using your own records.
Record event assumptions explicitly. A busy local event can change demand, but historical performance and competitor prices still provide context. Do not assume a sold-out night reveals all the demand you could have captured, or that tighter restrictions always improve the overall stay pattern.
6. Make the review a weekly routine
- Save an updated forecast for the next 30, 60 and 90 days.
- Review pickup, cancellations and remaining availability.
- Compare completed stay dates with their earlier forecasts.
- Record large errors and revise the relevant assumptions.
- Share useful estimates with the people planning staffing and purchasing.
Check staffing and selling costs alongside room revenue. A higher occupancy forecast does not by itself establish the best price or the most profitable outcome.
Choose a measurable next step
If you are evaluating software, use our small-hotel software break-even guide to consider the full cost. Book a Hotel RevBoost demonstration with your PMS details, room types and current forecasting process so the discussion addresses your actual requirements.
Content reviewed and updated by Intellisoftware on 9 October 2026. Forecast accuracy and revenue improvements are not guaranteed.
