A Multi-Branch Retrofit IoT Smart-Shelf System for Indonesian Restaurant SMEs: Cloud-Based Inventory Visibility to Cut Food Waste

Indonesia’s food-service sector is dominated by SME multi-branch chains and cloud kitchens driven by online food-delivery platforms. Many businesses rely on manual or semi-manual inventory practices with weak integration between sales and stock, contributing to avoidable food loss and waste (FLW) and excess inventory. This study proposes a Multi-Branch Retrofit Smart Shelf system that uses low-cost IoT nodes and cloud computing to provide real-time visibility of stock and expiry risk across branches. ESP32-based nodes with load cells and temperature sensors are attached to existing fridges and racks; batch identities and expiry dates are linked via QR or RFID/NFC tags, and telemetry is sent via MQTT to a cloud backend. Using a design science approach, we design, prototype, and test the system and simulate its use in a 10-branch chain. Lab results show that the retrofit nodes detect stock events and environmental conditions with sufficient accuracy and latency for kitchen use. Simulations suggest that local “consume soon” alerts can significantly reduce waste by 1.5–2% points, while combining local optimisation with cross-branch redistribution may cut waste by 3–5% points overall.