Deliveroo HK speeds delivery time with machine learning

Deliveroo HK speeds delivery time with machine learningDeliveroo Hong Kong has improved delivery time for meals by 20% using a customized predictive algorithm, making Hong Kong one of the most efficient Deliveroo markets globally.

The real-time dispatch algorithm is able to evaluate the most efficient way of distributing orders based on the location of restaurants, riders and customers.

The algorithm, named Frank after a character from US sitcom It's Always Sunny in Philadelphia, came online in early 2017.

Deliveroo GM for Hong Kong Brian Lo said the system uses machine learning to predict the preparation time of each meal based on the specific order and time of day.

“With the help of Frank, restaurants have been able to improve efficiency and save an average of 13 minutes preparation time per meal, which is of great help, especially during peak hours,” he said.

Deliveroo's algorithm also bases its decision process for selecting drivers on factors including weather, distance and location. As well as improving delivery time, the machine learning algorithm results in customers getting an ever more precise estimate of the time of arrival of a delivery.

“We’re building an algorithm which can cope with millions of simultaneous orders, while continuously refining itself, improving the experience for our riders, restaurants and customers,” Deliveroo CTO Mike Hudack said.

“The results in Hong Kong show this is working and as we grow our tech team and continue to refine our trailblazing technology, I look forward to seeing further improvements for customers in the region.”

The company has dedicated a substantial portion of its recent US$385 million funding injection to improving its technology capability.

Image from iStockphoto