ASSESSMENT OF FORECASTING METHODS TO REDUCE THE MARGIN OF ERROR IN ELECTRONIC COMPONENT SALES

• Ricardo Daniel López TECNOLÓGICO NACIONAL DE MÉXICO/CAMPUS CD. VICTORIA http://orcid.org/0000-0002-2662-6103
• Araceli Maldonado TECNOLÓGICO NACIONAL DE MÉXICO/CAMPUS CD. VICTORIA https://orcid.org/0000-0003-3585-8034
• Humberto Jasso TECNOLÓGICO NACIONAL DE MÉXICO/CAMPUS CD. VICTORIA https://orcid.org/0000-0003-4565-190X
• Gerardo Huerta Tecnológico Nacional de México/Campus Cd. Victoria, México. Division of Graduate Studies and Research https://orcid.org/0000-0003-0863-8159
• José Amparo Rodríguez Universidad Politécnica de Victoria, México. Department of Engineering and Manufacturing https://orcid.org/0000-0002-6226-4242

DOI:

https://doi.org/10.7166/33-1-2553

Keywords:

Forecasting, Qualitative methods, Combined methods

Abstract

International competition in the electronic market requires that organisations use their resources not only to manufacture high quality components, but also to adopt or develop appropriate sales forecasting methods that could adapt to their needs and guarantee their economic development. From an industrial engineering perspective, keeping balanced orders and healthy safety stocks is required for such organisations. These two metrics play a significant role in their economic growth and development, because any disruption results in high costs throughout their manufacturing processes. Thus significant resources are spent by these organisations to develop information systems and logistics skills in order to implement more reliable and precise sales forecasting methods. Nevertheless, planners and forecasters constantly face different challenges such as sudden demand changes, seasonality, products with a short life cycle, a lack of historical data, and swings in the world economy. The objective of this research is to determine the most convenient demand forecasting method for the manufacturers of electronic devices that target a specific market. Twenty-seven months of sales data were analysed and different quantitative forecasting methods were tested and analysed using statistical tools. From the results obtained, the combined forecasting method appeared to be the most suitable since the least amount of forecasting error is obtained when this method is applied. The results of this research could be adopted by other companies to forecast the future sales of any items with a similar pattern to that used in our study. This has significant implications for their decision-making processes and inventory planning.