Designing a Demand Forecasting Service in a Food-delivery Platform

Abstract: This thesis presents a preliminary design of a demand forecasting service using a service design approach. The service aims to provide a better user experience when forecasting demands for the Operational Managers (OM) in an online food-delivery platform. Based on the internal user interviews, demand forecasting is essential to plan the right yet efficient balance between order demand and delivery supply. However, some limitations were discovered in the existing process that creates unnecessary manual work, and therefore less time productivity. This thesis explores whether to create a better digital and centralized forecasting service and can be introduced to reduce the manual tasks as much as possible using Machine Learning models. The research methodologies used in this thesis are the user-centric design methods, for example, semi-structured interviews, Affinity diagrams, Stakeholder Mapping, Persona, User Journey Mapping, and Service Blueprint. Moreover, the research highlights the current gaps in the forecasting process and presents comprehensive suggestions in designing the forecasting service. The results also combined the stakeholder aspirations to ensure operational efficiency and user-centric design methods to solve those gaps.