Forecasting Daily Sales of Coffee Shop Products: A Comparative Study of Moving Average, ARIMA, and SARIMA

The coffee shop industry in Indonesia has experienced rapid growth, with local brands like Kopi Janji Jiwa and Kopi Kenangan expanding aggressively. However, coffee shops often face challenges in demand forecasting, leading to inventory mismanagement and operational inefficiencies. This study focuses on forecasting daily sales of Ice Afecto, a popular coffee shop product, using time series models to enhance business decision-making. This study compare the performance of three forecasting methods Moving Average (MA), Autoregressive Integrated Moving Average (ARIMA), and Seasonal ARIMA (SARIMA) to identify the most accurate model for short-term sales prediction. Utilizing transaction data from a coffee shop, processed and analyzed to capture daily sales trends. Results indicate that ARIMA (0,0,1) (0,0,1) outperforms both MA and SARIMA, achieving the lowest Mean Absolute Error and Root Mean Squared Error. While MA provides simplicity, it fails to capture rapid fluctuations, whereas SARIMA underperforms due to weak seasonal patterns in the data. The study highlights ARIMA’s effectiveness in optimizing inventory management, workforce scheduling, and promotional strategies, offering practical benefits for small-to-medium coffee shops. This research contributes to the industry by demonstrating how data-driven forecasting can minimize waste, improve resource allocation, and enhance profitability. Future work should incorporate external factors such as weather and promotions to further refine predictions. The findings provide valuable insights for coffee shop operators seeking to implement predictive analytics in their daily operations.