Picnic delivers e-groceries to thousands of customers every day. To do so both efficiently and effectively, Picnic’s distribution system needs to be as smooth as possible. One part of this system is the vehicle routing model, which determines which routes are optimal given the customers and their placed orders, amongst other factors. An important input into this model is the drop time: how much time we expect delivery to take for a given customer.
A new Picnic-blog covers why drop times are so important and how we, the Data Science team within Picnic, developed a model to better predict this amount of time. There are two contrasting objectives here. On one hand, we want to have a high on-time percentage, which requires us to plan enough time for each and every customer. On the other hand, we want to increase efficiency and serve as many customers as possible within the duration of a single delivery trip made by a runner.
Estimating delivery times
To make sure that we can increase efficiency while maintaining the same on-time percentage, Picnic needs to have a better estimate of how much time is ‘enough’ to deliver a given order. When drop time estimates are inaccurate, a planner needs to plan a lot more redundant time on top of our estimate as a safety buffer to maintain a high on-time percentage. When this estimate becomes more accurate, the planner can reduce this safety buffer, increasing the distribution system’s efficiency, without risking any late deliveries.
Factors affecting the duration of one single delivery can be grouped into four areas:
New planning model
Picnic developed a new planning model; starting off with a couple of hubs and gradually rolled out to more and more cities. Instead of immediately starting with the new model’s predictions, Picnic added a safety buffer on top of the planning model predictions because of the decrease in on-time performance they initially noticed during the pilot. Picnic then steadily decreased this safety buffer over the weeks following the roll-out to increase the efficiency of the system, while keeping a close eye on the on-time performance. This set-up gave Picnic the tools to play around with the trade-off between efficiency and on-time performance for each individual hub. Fast-forward a couple of months and Picnic now has the new model running in production for each hub in both the Netherlands and Germany. Picnic found that efficiency, measured in the number of deliveries per trip, increased by roughly 20%.