Demand forecasting and dynamic pricing for e-grocery fulfilment

E-fulfilment, especially grocery delivery, is time windowed – customer’s choices of time windows affects the efficiency of the delivery routes. Nevertheless, customer’s choices are not unchangeable, especially when financial incentives are available.

This KTP project aims to use Big Data, Forecasting and Machine Learning approaches to understand customer shopping behaviour, in order to proactively steer customers’ selection of time windows towards the most efficient delivery routes.

For example, if the delivery system identifies six households within a 5 mile radius who have booked their delivery between 10am and 12pm on a Tuesday, then customers with accounts registered within that same radius could be offered a cheaper delivery price in return for booking their delivery on the same date and timeslot.

The project benefits customers by providing customised delivery time window suggestions for their convenience at a potentially lower price (with the improved efficiency), and the environment by reducing carbon footprints of delivery fleets.