Every month we publish a "feature of the month" where we ask one of our consultants to highlight a feature in MILAS AX (our specific ERP modules for Microsoft Dynamics AX). This month Tim De Ryck shares his insights.
Since most businesses operate in an atmosphere of uncertainty, forecasting is considered a critical and, therefore, an indispensable step. Next to the improved purchase and sales planning, forecasting will allow you – amongst other things – to improve inventory controls and financial planning:
- Having valuable forecasting reports at your disposal will lead to improved inventory management, resulting in fewer out-of-stock and overstock situations.
- Anticipating sales will stand your business in good stead to better predict your turnover and profits.
In several industries, such as the feed and food industry, sales and purchase patterns are predictable as they recur over time (e.g. due to seasonality). In such sectors, building forecasts from history provides the business with an accurate outlook on the future. To support the forecasting process, MILAS AX offers a user-friendly and flexible tool: “Forecasts from history”.
Forecasts from history in MILAS AX
The “Forecast from history” tool allows you to generate a (purchase/sales) forecast per item based on specified transaction types within a specified time range. The following steps are taken when applying this MILAS AX module:
- Select a historical timespan;
- Select the transactions types on which you want to build your forecasts. For instance, for semi-finished products, users can select the transaction types “Production consumptions” and “Sales orders”;
- Select a calendar. Users can opt for a calendar that excludes weekends and national holidays;
- Select a site.
Applying these steps results in an average quantity per (work) day. If needed, this quantity can be adjusted automatically by a scaling factor, which can be overruled by the forecaster.Based on this quantity, a (purchase/sales) forecast can be built. To do so, the following steps are taken:
- Select a future timespan;
- Select a forecast model. For instance, users can select a model in which the first half of the timespan has a higher relative weight (due to seasonality);
- Select a calendar.
Practical example
To illustrate the theory described above, let’s take the following basic case as an example:
- Days = 30 (e.g. the last month);
- Working days = 21;
- The item was used in 15 production runs. In total = 26780kg;
- The item was 10 times sold. In total = 10380kg;
- Result:
- Quantity per working day: (26780+10380)/21 = 1769,52
- Days = 7 days (e.g. the next week);
- Working days = 5;
- Forecast model: relative weight of the first and last day = 23%; relative weight of the remaining three days = 18% (in Total = 100%) (See figure 1). In the feed industry for instance, such patterns are observed because animals need to be fed during weekends as well;
Figure 1: relative weight of the five working days
- Result:
- Forecast for the first and last day of the next week (per working day) = (1769,52 * 5)*23% = 2034,95
- Forecast for the remaining days of the next week (per working day) = (1769,52 * 5)*18% = 1592,57
Thereafter, based on an active recipe, forecasts of the semi-finished product can also be exploded into raw material purchase forecasts. This simple example gives an indication on how forecasts can be built from history in MILAS AX.
Conclusion
Despite the fact that many business people claim “forecasts are wrong” (in fact they are to a certain extent), organisations should not underestimate the noteworthiness of making accurate forecasts. To support this forecasting process, MILAS AX provides a structured, flexible and user-friendly tool.
