Forecasting Catering Requirements for Improved Passenger Experience

Challenge

The effective planning of onboard logistics for train operators is critical to delivering their customer proposition, particularly for premium seats.

Forecasting food requirements considering limited storage space, onboard team requirements and on the day schedule changes presents several challenges, especially when there is little visibility of future passenger levels on individual services.

This challenge becomes more complex with additional factors such as the train model and layout, staffing models and grades, together with the fact that food orders need to be submitted to suppliers several months in advance of any service running.

Decisions have significant impacts on the customer offering as well as the smooth running on services on the day. Underordered consumables and top- up orders of hard to predict quantities need to be delivered during the journey. Overorder and supplies are inevitably wasted. Incorrect numbers of onboard teams can lead to inefficiencies or insufficient numbers of staff to deliver the service. Not addressing challenges in accurate forecasting and responsiveness have a damaging effect on a smooth-running service.

Solution

Working with a rail catering team, we have supported this challenge using mathematical modelling.

The solution predicts the required stock levels and sales of different food categories throughout the journey, using historic data. Onboard staffing requirements are calculated down to the journey leg, rather than the individual service. This gives greater confidence in the decisions being made and ensures that even during the busiest sections of a journey, there is enough team members onboard.

Using a combination of historic passenger data combined with operator service reports, open data sources from Network Rail and National Rail statistics, the tool gives an accurate picture of future scenarios. Passenger numbers, service uptake rates and staff requirements can be forecast, whilst simulation enables evolving scenarios to be rapidly evaluated, accounting for uncertainty. Both planning and onboard teams can make more informed decisions about the effective use of resources, reducing food waste and costs.

Impact

Outputs helped to reduce the uncertainty present in the forecast of passenger numbers for a given journey, contributing to a more accurate placement of orders at relevant depots. The increased granularity in uptake for food categories also meant these orders were more accurate, further reducing waste and costs.

This approach has also allowed for more accurate requirement forecasts during special events such as festivals or sports events, where uptake can be four to five times more on specific sections of the journey. Further refinement and development of the solution is planned to accommodate both long-term planning and short-term disruption handling, where logistics become more challenging. Integration with onboard systems and electronic point of sale (EPOS) platforms will allow for easier transfer of requirements and visibility on stock requirements.