Forecasting, as is known in the economic environment to the process of forecasting sales or demands, is defined as art and science to predict the future for a particular good, component or service, based on historical data, marketing estimates and promotional information, through the application of various forecasting techniques.
Demand forecasts are a key part of planning systems and hence the economy in general. Demand forecasts have a great influence on the determination of key process factors, factors such as installed capacity (equipment, warehouses, plants), financial requirements (inventory, cash flow), organizational structure (people, systems, services), contracts with third parties (purchases, operators), etc.
Due to the extensive influence of forecasting on any production system, demand management is considered to be a fundamental factor for the success of any organization.
“Every activity requires some volume estimation system to be handled within it. Estimates are the result of predictions and forecasts”
Forecast Planning Horizon
One of the most common questions when generating a forecast is what demand period we need to calculate. That is, if we want to calculate the demand of a month, a quarter, a semester, a year… The time period that will cover the forecast is known as the planning horizon, and its suitability depends on our goal when using demand forecasting. It is very common in demand management to establish planning horizons no larger than 18 months, since it is considered that depending on the changes that constantly affect processes, systems and environments, a longer period would yield very unreliable results.
Implications of the forecast error
Although within the most common errors in demand management is first the failed selection of the forecasting method, there is a no lesser problem consisting in the elaboration of different forecasts by each functional organ of the organization, that is, the forecast is usually developed by the commercial area and clashes against the planning carried out by the production area.
Developing a forecasting process
- As the “Demand Planning” team
- Set the policy of the “Demand Planning” process
- Group sku’s into families
- Identify patterns of product demand
- Identify the stage in the lifecycle of each product
- Sort items in A, B, or C
- Implement forecasting software
- Load product demand history
- Examine identify and eliminate irregular claims
- Run the forecast module
- Get the “Demand Planning” team forecast
- Generate forward-looking forecasts
- Monitor the forecast
- Work with the forecasting system, not against it
AMERICAN PRODUCTION AND INVENTORY CONTROL SOCIETY- APICS
Demand Planning Team
Objective: Ensuring that a forecast is agreed
Focus: Align material availability and capacity with expected demand volumes.
There are currently a number of forecasting methods that can be considered as standard. There are two large groups covering all standardized forecasting methods, these are qualitative and quantitative. Another great categorization, has the forecasting methods in three categories, these are qualitative, historical projection (quantitative) and causal (quantitative).
Simple Average: This method is to attenuate the data by obtaining the arithmetic mean of a certain number of historical data to get with this the forecast for the next period. The number of data to consider to calculate the average is a decision of the person making the forecast.
Moving Average: Each point in a moving average of a time series is the arithmetic mean of a number of consecutive points in the series, where the number of points is chosen in such a way that seasonal and/or irregular effects are eliminated.
How to improve the prognosis?
According to a study by Marshall L.Fisher, Ananth Raman and Anna Sheen McClelland, an organization can substantially improve the accuracy of its forecasts by executing the following activities:
- Updating forecasts based on initial sales data.
- Analyzing the accuracy of your forecasts, identifying errors and understanding when and why they occur.
- Testing acceptance of new products before and after launch.
- Using different methods of forecasting approach, so that it allows to understand the different assumptions implicit in the different techniques.