3. The Naive Forecast

The naive forecast. Let’s start with the most simple form of forecast. It is called the naive method. It’s naive because all we’re doing is, we’re saying what we sold yesterday, that’s how much we’re going to sell today. What we sell today is the number we’re going to project for tomorrow.
It doesn’t take a lot. And it’s very simple and easy to use.
If we sold 20 croissants yesterday, we’re going to sell 20 croissants today. If we sell 20 croissants today, we’re going to sell 20 croissants tomorrow.
The naive forecast works very well in certain situations.
And it sometimes works even better than other more complicated methods.
The naive method can be represented mathematically as well. So our forecast, we denote that again as F. And we’re going to forecast into the current time period, so it would be F sub t. And we set that equal to our demand at t- 1. And that means the previous time interval. If we look at it on a timeframe then we are trying to forecast here and we are going to use
this level of data demand. At minus 1 as our forecast at period t. We can make several adaptations to the naive method. For example, if we have our forecast, and rather than using the previous time period as our demand, we have a forecast that is very seasonal. So it’s more applicable to use the same month last year then we would do something like this. So we’re recording time in equal size monthly buckets, and we’re going to say t-12 because that is the same month 12 years back or one year back. And that would be an example of a seasonal naive method.
Another version of this would be if we use
Our demand at the previous time period, but we do know that we are growing every month, and we know for example every month we sell T units more, and therefore we add that T into our forecast. The naive method is clearly simple, but in situations where other more complex forecasts are not doing very well, it may be equally good and remember what we said earlier. If you can use a simpler method and get the same level of accuracy, it’s better than using a more complex one.
Now the naive forecast is very noisy because it doesn’t filter out any noise whatsoever. That makes it very nervous but also responsive to changes in demand. Nervous means we have a very volatile forecast.
That may not be in our best interest when we’re trying to plan for a smooth production. Therefore, we have to understand that typically, only the last period is used for our forecast. That is a big assumption, and it may not be very realistic.
However, the naive method can be used as a benchmark to more complex methods and it tells you whether a more complex method is clearly superior.

Jim Rohn Sứ mệnh khởi nghiệp