General types of forecasting methods

Published on May 31, 2023   27 min
Please wait while the transcript is being prepared...
Hello, my name is Eric Wilson. I'm the Director of Thought Leadership with the Institute of Business Forecasting. Today I want to talk about general types of forecasting methods.
When you're looking at forecasting method there is a lot of different methodologies available to you that can build a lot of different models. Institute of Business Forecasting, IBF does a lot of research around what types of models people are using. Right now, about 50% of the models out there utilized by most practitioners in businesses are time series type of models. Which makes sense because a lot of the data that's available to organizations is with inside their own four walls. It's going to be shipment history, it's going to be order history. It's going to be things that are time related that inside their own four walls that they then can forecast with. That doesn't mean there's not other forecasting methods and models available. There's a lot of other types of methodologies available as well. Causal models right now make up about 17% or so of the methods that most companies and people are using. This looks at external variables or other types of relationship type variables and tries to find those patterns and be able to forecast with those. What we've seen recently with the more data available with more processing power available, with new types of methods that are now accessible to practitioners we're seeing new types of methodologies coming available as well. These are more of your AI or machine learning types of models and they actually make up approximately about 15% of the methods that companies are currently utilizing right now. There is one final type of method I do want to mention, and that's the judgmental models. Those are still relevant and important today, and I don't see them going away at any point. Judgment is still part of the forecasting process. It's still something we need to bring into a process to build better forecast, and there's judgemental type of models that you can utilize to bring in methods, different types of inputs, and be able to improve your forecasts as well.