Wing & Prayer Forecasts
Have you ever looked at a sales forecast or other financial forecast and thought to yourself “this looks like it has been done on the back of a postage stamp”?
Indeed there are people in organisations who rely on forecasts of this nature that depend primarily on high level estimates and intuition. Conversely there are many other people who go to extremes and build large models and forecasts that require loads of data and complex excel formulas.
So it is natural to wonder which type of model is better and it is therefore interesting to note the results of a study undertaken by the University of Southern California of 350 sales managers which concludes:
Building models and forecasts that can be adopted and applied as useful business tools is as much an art as it is a science. There are many steps, processes and features that should be applied to modelling to make sure the final model is useful.
One very important feature that good models include is the dynamic capacity to generate ‘what if scenarios’. This is how to best feed ‘intuition’ into the analysis. For example, what will the result on unit sales be if the price of the unit decreases by 5% or what will happen to profits if the $A appreciates in value.
Practically speaking, the extent to which someone should invest time and energy into preparing a model or forecast really depends on what the model is to be used for and what critical decisions need to be made as a result. The level of logic, data and intuitive input should then be matched accordingly.
Reference: “What type of forecaster are you”, Harvard Business Review, March 2016
Have you ever looked at a sales forecast or other financial forecast and thought to yourself “this looks like it has been done on the back of a postage stamp”?
Indeed there are people in organisations who rely on forecasts of this nature that depend primarily on high level estimates and intuition. Conversely there are many other people who go to extremes and build large models and forecasts that require loads of data and complex excel formulas.
So it is natural to wonder which type of model is better and it is therefore interesting to note the results of a study undertaken by the University of Southern California of 350 sales managers which concludes:
- Forecasts that are based mainly on intuition tend to be least reliable.
- Forecasts that rely mainly on logic and data but are low on intuitive input tend to be moderately reliable.
- Forecasts that are based on a good balance of logic, data and intuition tend to be most reliable.
Building models and forecasts that can be adopted and applied as useful business tools is as much an art as it is a science. There are many steps, processes and features that should be applied to modelling to make sure the final model is useful.
One very important feature that good models include is the dynamic capacity to generate ‘what if scenarios’. This is how to best feed ‘intuition’ into the analysis. For example, what will the result on unit sales be if the price of the unit decreases by 5% or what will happen to profits if the $A appreciates in value.
Practically speaking, the extent to which someone should invest time and energy into preparing a model or forecast really depends on what the model is to be used for and what critical decisions need to be made as a result. The level of logic, data and intuitive input should then be matched accordingly.
Reference: “What type of forecaster are you”, Harvard Business Review, March 2016