From Dine-in to Takeaways – Surviving the Covid Disruption
Monday, April 12, 2:40-3:20pm EDT
Kedai Kopi is a café, established in February 2018 and located in Taman Melawati, Selangor, Malaysia. It serves typical Kelantanese cuisine. Kelantanese is a state in North East Malaysia with a unique culture reflected by their art, dialect and food. Kedai Kopi has 186 food and beverage SKU’s and faces constant excess or shortages of SKU’s on a daily basis resulting in losses and wastages. Before the Covid-19 pandemic, majority of the sales came from dine-in customers. However due to the lockdowns and implementation of emergency in Malaysia Kedai Kopi has to focus on items that are more likely to be ordered as take-away items. Hence today, we need to understand the current demand to have a monthly plan on the SKU’s optimal offerings removing certain SKUs and minimizing inventory levels. Historical sales data was collected from point of sale system. Using segmentation SKU’s contributing to 80% of the revenue were prioritized and their demand understood. Since these items were served as a la carte other B and C segment items that were strongly correlated with A items were also mapped and a new menu was prepared combining these items as a group (combo) while keeping the a la carte option. Simultaneously, cost of raw materials, preparation lead time, optimal purchasing frequency and period for stocking was calculated while forecasting the demand for the next period. This forecast was converted to raw items for inventory where items used across different SKUs were combined together and forecasted on aggregate to increasing the forecasting accuracy and lower shortage and wastage. News Vendor model for daily SKU Forecast for the Café and Time Series methods for forecasting kitchen inventory were used. Finally we tested the robustness of our model using simulation and the results have predicted significant increase in profit and reduction in wastage and stock outs. The model is being actually implemented as of March 2021 in Kedai Kopi and we plan to fine tune the model further with more data in hand.