Principles that shape the design of our marketplace

Grab’s matching and pricing algorithm is the backbone of our deliveries and rides service. Our goal is to efficiently match drivers and consumers, at an optimised and competitive price for the supply and demand at that very moment. The fare is also balanced to help encourage the timely supply of drivers, and needs to dynamically adjust to changing demand and road conditions, among other factors.

Here are some principles that guide the design of our transport and deliveries allocation system.

Allocating the best suited driver-partner

We search across our diverse network of vehicle types, and take into account additional factors such as training and experience when sorting for a match. For example, the algorithm’s logic will assign larger vehicles to larger orders, and will set shorter trips to riders on bicycles.

The backend is also able to identify partners that have shown repeated success in navigating particular areas, and will select these drivers on trickier routes that might need to pass through one-way streets or back alleys. 

This search is neutral, and we don’t discriminate based on demographic characteristics. In some markets that have cultural sensitivities, however, we try to match female passengers with female driver-partners. Some Muslim food operators prefer delivery-partners that are trained to handle halal products, so we try to match those drivers in these cases, too.

Balanced earnings for all

Our algorithms also look across the board, and make sure driver earnings are fair for the service provided. 

We also want to incentivise high service levels, by rewarding partners who show a strong track record with good reviews and a high trip completion rate. We offer priority allocation to minimise their time driving around looking for a booking.

We also prioritise trip allocation to partners who turn on auto-accept on rides. This is done to prevent rider disappointment and unnecessary delays in the system that may arise from partners who cherry pick bookings and turn down requests, resulting in the system needing to reallocate. 

But we don’t want drivers to sacrifice safety if they need a break. We don’t penalise drivers who turn off the app. They’re free to do so, and resume when they’re ready. 

The system also allows drivers to finish working near home; they can set an end-destination, and Grab’s algorithm will match them to a booking that ends nearby, when it’s time to call it a day.