<![CDATA[Product portfolio optimization is essential in automated retail because vending machines operate under severe storage constraints, requiring every Stock Keeping Unit (SKU) to contribute effectively to overall sales performance. However, conventional portfolio management approaches, including the Boston Consulting Group (BCG) Matrix, rely on market-level indicators that are difficult to apply in vending machine environments where external market share data are unavailable. This study proposes a Modified BCG Matrix by replacing Relative Market Share with Relative Sales Performance (RSP) and Market Growth Rate with Sales Growth Trend (SGT) to enable SKU-level portfolio optimization using internal transaction data. A quantitative longitudinal design with an Interrupted Time Series (ITS) approach was employed using 167,018 valid transactions collected from 11 smart vending machines operating in Cikarang, Indonesia, between September 2025 and June 2026. The Modified BCG Matrix guided operational decisions, including facing allocation, replenishment priority, product rotation, replacement, and product delisting. The implementation reduced the active product portfolio from 253 to 146 SKUs (42.3%), increased average daily transaction volume by 28.53%, increased average daily sales revenue by 22.19%, and produced statistically significant immediate intervention effects based on segmented regression analysis. These findings demonstrate that data-driven product portfolio optimization not only improves operational efficiency and sales performance but also reshapes consumer purchase behavior. This study extends the application of the traditional BCG Matrix from the Strategic Business Unit (SBU) level to the SKU level and provides an evidence-based decision-support framework for automated retail portfolio management.]]>
Copyrights © 2026
