Forecasting Algorithm

Summary

The forecasting algorithm uses historical data (from up to four years back) to come up with a trustworthy forecast.

Intended Users

General Managers, Hotel Managers, Revenue Managers and Booking Managers.

Instructions

Historical information from up to four years back is used, but the further away from the forecasted date, the less weight is used: e.g. last week will have a much higher impact than days four years ago.

The algorithm uses several parameters to come up with a forecast:

A. Season (special periods, etc. as per “Defining Seasons”) − Only historical days within the same season as the day being forecast are used.

B. Day of the week − Only historical days with the same weekday forecasted are used (i.e. Mondays to Mondays).

C. Days remaining till arrival.

D. Current OTBs − Historical days that fit the criteria A-C above. The more similar the OTB situation is to the same remaining lead days, the higher the weight.

E. If historical days have been constrained, i.e. more rooms could have been sold but there was no availability. A historical day that was sold out in advance will not be included in calculating future pickups.

After all the days that fit the criteria A-C have been weighted according to their relevance for the day that is being forecast (D-E), the algorithm will absorb all the values and calculate the pickups for both rooms and revenue for the day.

Was this article helpful?

Related Articles