Methods of Forecasting the Sale of Metal Products on Local Markets as a Component of the Enterprises’ Economic Security

The article investigates forecasting methods for the sales of metallurgical products in local markets, with an emphasis on their role in strengthening the economic security of enterprises. This problem is especially relevant for Ukrainian metallurgical companies, which face increasing global competition, market volatility, and the need to adapt production and sales strategies to unstable local demand. Reliable sales forecasting is a prerequisite for efficient resource allocation, production planning, and building resilience in uncertain environments. The aim of the study is to determine the most effective models for forecasting the sales dynamics of metallurgical products in local markets. For this purpose, four forecasting approaches were evaluated: Simple Moving Averages (SMA), Classical Seasonal Decomposition (CSD), Seasonal Autoregressive Integrated Moving Average (SARIMA), and its extension with exogenous variables (SARIMAX). The empirical analysis was based on an anonymized dataset covering a 57-month period, which enabled the detection of seasonal patterns and the assessment of forecasting accuracy. The results demonstrate that different models provide varying levels of precision depending on the characteristics of the data. On average, the forecasting error measured by Mean Absolute Percentage Error (MAPE) for a 9-month horizon amounted to 0.2403 for SMA, 0.2308 for CSD, 0.2174 for SARIMA, and 0.2074 for SARIMAX. Thus, SARIMAX proved to be the most accurate model, while SMA, despite its simplicity and minimal data requirements, showed higher errors and a systematic lag in capturing seasonal changes. CSD performed well for product categories with strong seasonal demand, whereas SARIMA and SARIMAX provided robust results across both short- and medium-term horizons, with SARIMAX offering additional benefits from incorporating exogenous factors. The study concludes that no single model is universally optimal, and the choice should depend on data availability, the presence of seasonality, and the planning horizon. Nevertheless, SARIMAX can be considered the most promising method for practical application, offering the lowest forecast errors and the highest adaptability to complex market conditions. The findings contribute to improving sales forecasting practices and provide a methodological basis for enhancing the economic security of metallurgical enterprises operating in local markets.