As we begin looking at the different types of forecasting methods, one of the first things that people think of is the quantitative type of forecast. They take historical data or time series or correlation data and project what they think is going to happen. There are however, quite a few instances where we see a lot of qualitative forecasting approaches. Qualitative approaches are based on opinions from experts, decision makers, or customers about what you would expect to see in a specific situation. In this overview we’re going to give an overview of the different qualitative and quantitative methods to forecast. You can see how to put each of these methods into practice in the other forecasting methods articles!
The quantitative forecasting methods in supply chain approach can be broken up into 4 different methods:
Naïve Approach: Looking at what happened in the previous sales period and saying that’ll happen again. (i.e. I sold 100 widgets last sales period, therefore I’m going to sell 100 widgets again.)
Moving Averages: Taking the average of previous sales periods and applying it to upcoming periods (i.e. If the average of the last 3 sales periods is 130, then the next period will be in that range.)
Exponential Smoothing: Taking a weighted average approach when looking at moving averages. (i.e. If i’m selling ice cream I may weight January-March differently than July-September.)
Trend Projection: What is the trajectory, based on our data, of what will to happen. (i.e. If we are increasing every period we should raise the forecast. Or, maybe we have 2 increases, a small decrease, and another increase so we should adjust the forecast to gradually increase accordingly.)
We’re going to look at each of these methods in individual videos and then we’ll talk about linear regression and how to think about it when forecasting.
The qualitative forecasting methods in supply chain approach can also be broken up into 4 different methods:
Executive Opinion: A group of executives making a decision on what will happen in the next period. (i.e. The CEO, COO, VP of Sales, and VP of marketing meet to decide, based on their experience, where the company sales are headed.)
Delphi Method: Trusted advisors in the industry give an opinion about what they believe will happen, then another group compiles and interprets the analysis to give to decision makers.(i.e. A group of experts decide how many widgets they would buy or how many they think will be sold in different markets and then sends that to an internal group within a company to interpret and, in turn, relay to the company decision makers.)
Sales Force Estimates: Individual sales people make their own sales forecasting estimates based on their experience selling the product. (i.e. the sales team believes they will close a deal this month with a large retail company so the forecast is adjusted to reflect that.)
Consumer Surveys: Asking consumer for their opinion on products (i.e. Customers are given a survey on a new product to be released to gage the behavior of other consumers.)
Again, in each of the articles we will get into greater detail on each of the specific forecasting methods in supply chain and how you can use them to forecast. Avercast, our forecasting solution, makes good use of these approaches and is powered by 250+ algorithms.