Forecasting and estimation techniques

Abstract

It is very common in any project or program that deviations occur on different scales between what is expected and reality. On many occasions, these deviations could have been reduced by statistical analysis of past examples and their respective forecasts. Throughout this article, we will focus on studying the main statistical methods used when making forecasts and how, by using them, organizations could optimize their supply chain or the management of their projects by forecasting the quantity of units or resources that the market or project in question will demand, based on historical data from previous seasons. As an output, the advanced management of the project will be optimized, resources will be assigned to the tasks or processes that really require them, unnecessary costs will be eliminated, and the time used to complete the project will be reduced.

From the beginning of the 20th century to the present day, the importance of forecasts has grown exponentially, since it gives the company a competitive advantage, since it will be able to plan its actions and take measures based on expectations. The term "forecast" can be associated with different fields, such as business, engineering, politics, project management ... In this article we will approach the term forecast from an operations management perspective, defining the term as an accurate prediction of the future, based on past events and whose objective is to give the company valuable time to face future events, to be able to respond to the market demand and allowing the company to adapt capacity of production to fluctuations in demand: raw material consumption planning, supply chain management, quantities to order from suppliers, work shifts...

Forecasting approaches

Demand classification

Demand can be classified in three ways: stable, trending, or seasonal.

In the first place, stable demand is characterized by following a common sales or production needs pattern during the time analyzed. That is, even with maximums and minimums, it can be assumed that during the set period the level of sales (and therefore production) will not fluctuate from the expected standards, that is, it will have a constant average.

Secondly, a trending demand pattern is defined as a constant or systematic increase or decrease in demand as time progresses. An example of trending demand could be the sale of electric scooters, whose demand has increased day by day in recent years.

Thirdly, seasonal demand is one that during specific periods of time the market is stronger and experiences an exceptional peak demand. An example of seasonal demand could be the sale of ice cream during summer periods or the increase in demand for bouquets of flowers on Valentine's Day.

Forecasting classification (short, middle and long term)

The forecast o estimation process can be structured in different levels, which are defined according to the objectives and the time horizon of visualization (short, medium and long term). These must be carried out under a continuous improvement model, executed periodically and their performance should de measured, in order to keep improving the quality of forecasts.

Short-term forecasts are generally made with a vision of three months and with a maximum of one year, whose main applicability at the operational level is the planning of tasks, assignment of workers or immediate material needs.

Medium-term forecasts are intended to anticipate events six months ahead, and up to a maximum of three years. Their main operational functionality is to make a sales forecast, from which a bill of materials (BOM) and a production planning will be made.

Finally, long-term forecasts are made with a time horizon longer than three years and their main purpose is the development of new products or the planning of the use of facilities.

Forecasting methods

It can be projected into the future in two ways: through qualitative methods, in which we rely on past actions or on the implicit knowledge of the subject to intuit future actions. Or through quantitative methods, in which through the use of statistics or mathematical models, historical data are projected into the future.