PREDIKSI PENJUALAN OBAT PADA APOTEK RAISHA FARMA MENGGUNAKAN METODE TRIPLE EXPONENSIAL SMOOTHING

Advances in information technology require companies to have computerized sales forecasting systems to minimize losses due to insufficient or excessive drug stocks. Raisa Farma Pharmacy still uses manual recording, which is prone to errors in managing drug availability. This study developed a website-based drug sales forecasting system using the Triple Exponential Smoothing (TES) method, which is capable of handling time series data with seasonal patterns and trends. Testing was conducted by measuring accuracy using Mean Absolute Percentage Error (MAPE) and Mean Percentage Error (MPE), as well as functional testing using the Blackbox Testing method. The results show that the system is capable of producing a MAPE of 7.16% at α = 0.9, which is classified as highly accurate, and an MPE of -0.25%, indicating a slight overestimate. Evaluation of various α values shows that α = 0.9 provides the lowest prediction error, indicating that the predictions follow the actual data patterns well. Functional testing proves that all features work as needed, including sales data input, prediction calculations, and automatic evaluation result display. This system provides practical benefits for pharmacy owners in effectively controlling drug stocks, minimizing the risk of loss, and supporting data-driven decision making.