Customer Stories / Hospitality
2021
Deliveroo Makes Data-driven Decisions Based on AWS Machine Learning
Deliveroo uses machine learning (ML) to make better decisions. Those decisions can be automated or they can be made by humans supported by rich data, says Mahana Mansfield, data science director at Deliveroo.
Automating dispatch of riders is an example of an area where using ML produces a better, faster result than would be possible with a human in the loop. Full automation benefits everyone, from restaurant, to rider, to consumer.
On the other hand, deciding which types of marketing campaigns to run is an area where data science can help humans make the best choices. “People used to make those decisions based on their past experiences. They’d try to look at results but were limited in how deep they could go,” says Mansfield. “Now it’s possible to look at different campaigns and analyze what the responses were and to control for variables not related to the campaign, such as seasonal variations or which style of cuisine is most popular in a given delivery area.
“The final decisions are still made by people, but now they can base it on a deep understanding of data, rather than trying to guess which factors were most important and just going on gut feeling,” says Mansfield. “They’re able to perform at a higher level because they access the data they need.”
How ML Delivers Better Service
Deliveroo has embraced Amazon Web Services (AWS) to help it use ML more quickly and efficiently. That’s important, says Michael Sprague, machine learning engineering manager at Deliveroo, because ML touches so many aspects of the business.
"A lot of the things that the customers see on the site are machine learning. The restaurant recommendations. The estimated delivery times. Everything you see on the screen is ML. We wouldn’t be Deliveroo without it,” he explains.
“On the delivery side, Frank (Deliveroo’s rider dispatch tool) tries to predict how long it would take the rider to get to the restaurant, how long for the food to be cooked, the delivery time, and so on. That would be a lot of variables for human dispatchers to consider. ML can refresh its view of the world every few seconds and humans can focus on higher-value tasks,” says Sprague.
Deliveroo also uses ML and artificial intelligence (AI) tools powered by data to provide a better experience in its contact center. Amazon Connect, an omnichannel cloud contact center, integrates with Amazon Transcribe to automatically convert speech to text. That text can then be analyzed using Contact Lens for Amazon Connect to better understand the reasons customers are calling and any snags in the Deliveroo care experience, allowing the company to improve its practices.
In addition, Amazon Translate gives Deliveroo’s agents the ability to serve customers speaking a different language using real-time translation. This means that if there is a demand spike in one region, agents from another market can help, even if they don’t speak the language. Data science is the foundation of so much of what makes Deliveroo special.
Making the Right ML Choices
Communication between data scientists and other parts of the business is crucial to the success of any ML project, says Mansfield. AWS has the tools and resources ready, but a business needs the right people to use those to achieve business goals. The best ML results start with the best human results.
“If a company wants to get into machine learning, they should be looking for the right people. Trust the people you can understand, the ones who keep it simple and speak in plain English. They should be able to speak in terms of impacts and how to measure that impact,” Mansfield says.
“Communication with your ML team is essential. It’s their job to support the people in the business to automate what can be automated, make better decisions faster, and deliver the most value."
AWS Services Used
Amazon Connect
With Amazon Connect, you can set up a contact center in minutes that can scale to support millions of customers.
Amazon Transcribe
Automatically convert speech to text to extract key business insights.
Amazon Translate
Amazon Translate is a neural machine translation service that delivers fast, high-quality, affordable, and customizable language translation.
Explore Deliveroo's journey of innovation using AWS
More Deliveroo Stories
Get Started
Organizations of all sizes across all industries are transforming their businesses and delivering on their missions every day using AWS. Contact our experts and start your own AWS journey today.