Accurate forecasting of insurance policy sales is essential for strategic planning and performance evaluation in the insurance industry. Fluctuations in monthly sales create uncertainty that may affect marketing decisions and the achievement of operational targets. This study aims to determine the most appropriate AutoRegressive Integrated Moving Average (ARIMA) model for forecasting monthly insurance policy sales at PT Avrist Assurance KPA Lampung and to evaluate its forecasting accuracy. The study applies the Box–Jenkins framework, including model identification, parameter estimation, and diagnostic checking, using monthly time series data. Model selection is based on statistical significance, residual diagnostics, and forecasting accuracy measures. The results indicate that the ARIMA (3,1,0) model provides the best fit to the observed data, with a Mean Absolute Percentage Error (MAPE) of 12%. The selected model is subsequently used to generate five-step-ahead forecasts for short-term planning purposes. These findings suggest that the ARIMA framework can provide reliable short-term sales projections and support managerial decision-making in developing more effective marketing and operational strategies.
Copyrights © 2026