Saving 250 Million Meals from Waste Using AWS Managed Services with Too Good To Go

Founded in 2015 in Denmark, Too Good To Go has grown its operations across 17 countries. Customers can purchase and pick up “Too Good To Go Surprise Bags” of leftovers, which participating restaurants offer at deeply discounted prices around closing time. “At first, some people would set their alarm clocks for late at night to get these bags as fast as possible,” says Robert Christiansen, vice president of engineering at Too Good To Go. “That success meant the service sometimes went down during these traffic spikes.”

Bursts of customer traffic can vary by a factor of 50 between off-hours periods and peak hours of 16 purchases per second. Rather than invest in costly hardware that would drain energy as it idled for much of the day, Too Good To Go uses AWS managed services to scale sustainably in the cloud. “We use the elasticity of AWS,” says Christiansen. “We don’t want to be polluting and are using everything we can to scale as we need to.” The company has grown from saving around 480,000 meals in 2016 to nearly 79 million meals in 2022. Its total 250 million meals saved as of 2023 equate to 625,000 tons of carbon dioxide avoided.

To build further efficiencies into its operations, Too Good To Go went through an AWS Well-Architected Review, which uses a framework of six pillars: operational excellence, security, reliability, performance efficiency, cost optimization, and sustainability. “We check in with our AWS solutions architects and account manager to determine how we can take advantage of new managed services to be more efficient,” says Christiansen. “Using AWS managed services makes it a lot faster for us to make product changes. They quickly scale up to accommodate high traffic load, and we don’t have unused servers standing around. AWS services are super great for us not only from a cost standpoint but also from an environmental perspective.”

To work toward carbon neutrality within its own operations, Too Good To Go uses the Customer Carbon Footprint Tool from AWS to track, measure, review, and forecast the carbon emissions generated from AWS usage. “We see which Region has the lowest carbon footprint,” says Christiansen. “We think about how to optimize so that we don’t burden the environment.” Too Good To Go also seeks to work with businesses that prioritize sustainability in their operations, an important factor in its choice of AWS and recognition of its sustainability initiatives. “We make the green choice for our shipping partners, warehouses, and packaging because we don’t want to shift the environmental impact of our operations somewhere else,” says Christiansen.

Consumers and business partners use the Too Good To Go application, which handles more than 100 million API requests per day. The API calls access backend logic hosted on Amazon Elastic Kubernetes Service (Amazon EKS), a managed Kubernetes service to run Kubernetes on AWS and on-premises data centers. Using Amazon EKS, the company uses only the resources that it needs, and clusters scale automatically to accommodate bursts of user requests. “There’s a limit to how many people we have available, and we didn’t want to put our efforts into a big tech stack and have to manage uptime, quality control, and security ourselves,” says Christiansen. “Using AWS, we concentrate on innovating and building an actual product instead of managing infrastructure.”

As it launched in the United States at the start of 2021, Too Good To Go recognized the need to keep user data safe across different continents. Its service operates in multiple AWS Regions so that it can comply with Europe’s General Data Protection Regulation (GDPR) and other local data privacy regulations. “Using multiple AWS Regions, we reduce latency for users and keep their data local,” says Christiansen. “We can be compliant with data privacy requirements as we continue scaling our business worldwide on AWS.”

Too Good To Go uses a breadth of AWS database and analytics services. Databases scale automatically using Amazon Relational Database Service (Amazon RDS), a service that makes it simple to set up, operate, and scale databases in the cloud. For its primary data store, Too Good To Go uses Amazon Aurora, which—as part of Amazon RDS—provides built-in security, continuous backups, serverless compute, up to 15 read replicas, automated multi-Region replication, and integrations with other AWS services. Customer and restaurant information from the databases feeds advanced search capabilities within Amazon OpenSearch Service, which lets organizations securely unlock near-real-time search, monitoring, and analysis of data. “Using Amazon OpenSearch Service, we offer customers very localized, very narrow searches for what they want,” says Christiansen. For operational data analytics, Too Good To Go uses Amazon Redshift, which offers optimal price-performance for cloud data warehousing.

The company plans to use built-in machine learning capabilities on AWS to drive further operational efficiencies as the company grows, such as gaining near-real-time insights into what customers want and eventually developing chatbots to improve customer service. Too Good to Go hopes to achieve impact at scale, expanding the app’s success beyond the retail segment to address food waste further along the supply chain, including food production and distribution.

“Using AWS, we address the problem of food waste for our customers,” says Christiansen. “There’s an economic interest. But in general, people want to do better. We all are very much aware that we only have one planet, and we need to do better than what we do now.”