[MUSIC] Welcome back. Predicting the future is part of our everyday life. We predict the future many, many times every day. In fact, nearly every decision we make inherently includes some kind of expectation for the future. This is probably most apparent when we talk about weather forecasts. In this video, we will define forecasting, discuss two different types of forecasting, and show you how to use a very simple forecast model to help you improve your forecast. Put simply, forecasting is in some ways predicting what will happen in the future. A lot of misunderstanding in the area forecast stems from the fact that we have two very different types of forecasts. But you almost never hear people clarifying what forecast they actually mean. In general, owners, asset managers, investors, and banks, as well as management and operations, mean a performance forecast when they say the word forecast. This is, quite simply, an estimate of what my future performance will be, either in terms of revenue, rooms sold, average daily rate, or some other metric. This is often a monthly level, but may be broken down by day to support operations in terms of staff planning or stock ordering. I was a cook, so I know how valuable it is to know that approximately 85 breakfasts are going to be served tomorrow morning. Revenue managers, however, are not really interested in the performance forecast, but instead, in the demand forecast. The demand forecast tells you how many people want to come to your hotel, including those who will be turned away because there is not enough space left. Remember back to the revenue management bingo game? If I had of told you that there were a total of 15 reservations coming, that would of been very useful to know, right? If I had of told you how many other requests we were going to get per day, per rate, or marketing segment, you would have been able to figure out in advance who to say no to to maximize your revenue. You would have been able to win that game. That is a demand forecast, and to be valuable, it has to be broken down by market segments. Think about a specific day exactly two weeks from today. If I was to tell you that you already have 169 rooms on the books or reserved for that day, as of today, how many requests do you think you will get in the next two weeks? Difficult question, right? Would it help if I told you that one year ago, two weeks prior to the same week day arrival, we had a 141 reservations on the books, and we ended with 198 out of 212 rooms sold. In this picture, we are looking at a graph of how reservations increased for a single arrival day as we get closer and closer to it. On the bottom right, you can see how many rooms were reserved a 190 days prior to arrival, that is, zero. As we move towards the left, we get closer and closer to the arrival day. One could argue that if we currently have 28, that is 169- 141 reservations more than the same time last year, then we might end up with 28 reservations more than the final last year. So,198 + 28 = 226 in demand, even though we only have 212 rooms. This is no different than saying we expect to pick up the same number of reservations that we picked up last year in the last two weeks of prior to arrival. However, to add a little bit of finesse, we might use a model where we assume that pick up pace in percent is the same this year, two weeks prior to arrival, as it was for the same day last year, two weeks prior to arrival. This would mean that if last year, two weeks prior to arrival, we had already picked up 70% of the total bookings we were going to get, then we could easily assume that this year, two weeks prior to arrival, we also have 70% of the total bookings we are going to make for that first Wednesday in December. To calculate this, we simply need to divide the number of rooms we have on the books this year two weeks prior to arrival, and divide it by last year's pace also two weeks prior to arrival, in percent. Last years pace would be 141 / 198, which is 71.2. In other words, last year, two weeks prior to arrival, we had picked up around 71% of the total rooms we were going to sell for that arrival day. This year two weeks prior to arrival we already have 169 rooms, so we can assume that 169 is 71.2% of the total rooms that are going to be requested this year. Now, it is simply a case of dividing the current on the books of 169 by the 71.2, which gives us a number of 238. In this example, we have used the hotel total figures to demonstrate the process. This is, in reality, not really ideal. Really, you should be doing this kind of forecast for each major market segment. The reason why will become more evident once we discuss the decisions in more detail. It is also important to note that this is definitely not a model to be accepted blindly. For many specific days, there will be some variance that makes the pace based forecast unrealistic. This is one of the reasons why we have revenue managers, to adjust a forecast based on experience and market knowledge. However, this is a great way to speed up a daily forecast and improve accuracy by starting with a simple, but powerful model. For those who want to understand this in more detail, please try the optional monthly forecasting exercise. There are no wrong answers, and it can be actually a lot of fun. Although our colleagues require performance forecasts, an unconstrained demand forecast is really the tool that helps the revenue manager make the right decisions. In the next session, we will discuss specific decisions based on the forecast we've created.