Forecastability Analysis

What is forecastability analysis?

In essence, this refers to appraising your company’s demand scheduling challenges while helping employees focus on addressing demand variability. This process entails segmenting items in various unique ways so that you can establish and apply the appropriate supply chain policies for production and inventory management and inventory planning.

You can also use this process to evaluate how profitable the company’s products and product segments are. The result is a set of defined rules to be applied to accomplish dramatic improvements in the bottom line.

Some of the key indicators that point to the need to conduct this analysis in your demand forecasting include:

• Lack of precision
• Tedious forecasting
• Too much inventory
• Stock-outs

Why is it essential for statistical forecasts?

Determining where to focus your limited forecasting capacity among the multiple products and customers in your portfolio – both your demand planning and sales/marketing resources – is crucial. Understanding the forecastability of your portfolio makes this easier.

It helps to make important decisions:

• What is the lowest level in the hierarchy you should forecast at?
• What should be the time granularity of your forecasts and the proper forecast offset (lag)?
• Choose statistical forecasting algorithms correctly based on the
forecast profile

How Does it Work?

There are 3 essential steps to a good Forecastability Analysis.

• Demand History Analysis
This initial step involves evaluating your actual sales history for each product. You’ll also begin segmenting the products based on their relative value. Analyzing your demand history will also reveal the different factors that affect the demand in your company.

New = not enough history to generate a good forecast
Emerging = able to generate forecast but not detect seasonality
Mature = best picture of demand history
Sporadic = high percentage of historical values are zeroes
Delete = no sales history, obsolete

• Selecting a Practical Forecast Strategy
At this stage, the different approaches you use in forecasting will be evaluated to determine the ones with the highest accuracy. The
process entails checking out statistical and non-statistical methods, various product hierarchies, promotion/event modeling, and product lifecycle curves.

• Quadrant Analysis
This final step gets to the actionable results of the analysis. The
different products will be grouped and segmented based on their
relative value and ease of forecast. You’ll end up with a final set of practical policies for driving inventory, capacity building, and
prediction.

For instance, if you have products with higher forecastability, you’ll
require an automated forecast, and responsible staff should be there to monitor any exceptions. On the other hand, those with low value and low forecastability require management with inventory policies instead of wasting time carrying out direct forecasting.