 Hi, this is Dr. Don. I want to give you a review of exponential smoothing, doing it both using basic Excel functions as well as using the data analysis tool pack. Many students make a mistake when they're doing exponential smoothing and I don't know for sure whether it's because they don't recall the formula or whether they're using the data analysis tool pack. Let me show you how to approach the problem. Here we have some time series data for two years, 15 and 16, and we need to forecast a value for January of 17. To do that using the exponential smoothing, we use this formula. The forecast for time t is equal to alpha, our smoothing constant times the actual for time period t minus 1 plus 1 minus alpha times the forecast for t minus 1. If alpha is set up to be 0.2, then here's how we would build our forecast. We have to start with a seed value because the formula is looking back one period, so you can use a number of ways you can do this. I just like to use my first actual period there to start. Now for my first forecast, I want to go equal my alpha times my actual from t minus 1 plus 1 minus alpha, close the parentheses, times my forecast for t minus 1. And that gives me, as you will see, it's the same value and that makes sense if you just go through the algebra. You can see that we're taking, in this case, 20% of this plus 80% of that added together, which of course gives you that. So then what we need to do before we drag this down, we need to lock it so we're always referring back to alpha. So I'm going to go up here to D1, hit my function 4 key to lock that down. There's my other D1 function 4 key to lock that down. Now I can drag this formula down to get my forecast for the actuals and I'm going to drag it one more place to get my forecast for January of 17. Pretty, pretty straight. If we look at the formula here, up here you can see it's alpha, which is D1, times B3, B is my actual column, 1 minus alpha, which would be 0.8, times C3, my forecast column, which is what we want. Now let's do it using the data analysis package. We go up here to data, data analysis, bring this up, and there's our dialog box and we pick exponential smoothing. I'm going to click OK and it'll sometimes offer up stuff. I want to delete that and I want to make sure that I have my actual values there in the range. Here's the problem that I see a lot of students get into. Excel here is asking for a damping factor and many students will plunge ahead and put in alpha there, but that's not right. The damping factor is equal to 1 minus alpha, so if our alpha that we're interested in is 0.2, 1 minus 0.2 is 0.8, so that's the damping factor we want. We do not have labels this time included. The other mistake I see students make is where to have the output range. You want it to parallel the other data there, so you should select D3 in this case. I'm not going to get the chart output to standard errors here, but I'm going to click OK and you can see that it enters data there. I'm going to format that so it looks the same, so it matches. You can see that the forecast are the same. I forgot to change that one. As we have with the manual method of calculating this, we have an NA put up here in that first cell instead of just copying over the seed forecast. The reason for that when we calculate our error metrics, our mean squared error, root mean squared, our mape, we don't want to use that first row anyway, so that's the reason Excel just puts the NA in there because we're going to ignore that first value. But anyway, that's how to do it using both methods and I forgot I've got to drag this down to get my forecast of January 17th. That's how to do an exponential smoothing using both the native Excel as well as the data analysis tool pack. Hope this helps.