Exponential smoothing, take for example ARM & HAMMER baking soda. It's a product you can use for baking, cleaning, or even brushing your teeth. Therefore, we expect demand to be fairly stable over time. We don't expect people all of a sudden to use much more or much less. This makes it a great candidate for exponential smoothing. Exponential smoothing, similarly to the moving average, is a very versatile method. But actually, I like it even better because it is much more elegant to implement. By changing one value, you can make it more reactive or more stable. Now let's take a look behind the math of the exponential smoothing formula. So our forecast again is denoted by F sub t. And that is equal to alpha, and I'll explain what alpha means, times, Our demand at t sub 1 + (1- alpha) times our previous forecast at t- 1. So, what we're doing here is we are balancing how much weight we give to our previous or most recent demand and how much weight we're giving to all of the forecasts. And through this part everything is being linked back all the way to the first information available. And this alpha and the 1- alpha are setting the weights of our weighted moving average. So there you have exponential smoothing. So how do you know what value of alpha to pick? For moving average we said you have to do trial and error. And this is slightly true for exponential smoothing as well. However, rather than redoing the forecast over and over, all we need to do is change the value of alpha from larger to smaller. And it makes the forecast either more reactive or more stable. Now, when you change these different values for alpha, you have to keep looking at your accuracy. If your accuracy is improving, then you're going in the right direction. If your accuracy's getting worse, you're going in the wrong direction. Luckily, because we have two cells in a spreadsheet that we need to monitor and adjust, it's something that we can automate. And I will show you this later. There are a number of extensions to the exponential smoothing formula. You can use the basic principle to not just forecast stable demand, as we've done so far. But you can take into account trended data, which means you have a consistent increase over time or decrease over time. Or you can estimate demand that is seasonal. That is, we have a consistent pattern, or demand data that is both trended and seasonal. So the options are limitless by using just this basic exponential smoothing formula, making some adaptations to it, and using it for a completely new purpose.