Americanas S.A. is the leading retail innovation platform in Brazil, with over 51 million active e-commerce customers and 3,500 physical stores across the country. Its mission is to bring together the best from around the world to improve people's lives - with exceptional customer service and fast product delivery, even in the most remote areas. With a unified base of 51 million active customers, over 40 thousand members, 137 million items for sale, and 3 million partners including suppliers, vendors and merchants, Americanas is committed to providing an omnichannel consumer experience with endless possibilities.
The Challenge
With exponential business growth comes exponential data growth. To handle this increasing volume of data, Americanas built a sophisticated cloud-based data platform. But the team still faced challenges. Scaling the infrastructure so that everyone had access to the right data and managing data quality at scale was becoming difficult, especially with AI and ML becoming more important, impacting both data volume and quality. To ensure that everyone continues to have access to reliable data to make data-driven decisions, Americanas not only needed to democratize initiatives across the organization, they also needed a better solution for end-to-end observability and data quality management.
The Solution
Americanas chose Soda Cloud as its data observability solution due to its reliability, ease of use, automation capabilities, availability, and ability to integrate with existing ecosystems. Open source was a major factor in their decision making, as the company's data stack is largely based on open source tools such as Apache Kafka, Airflow and others. The team also appreciated Soda Cloud's scalability - they now use it as a single platform for their entire data pipeline.
The deployment of Soda Cloud allows every business unit and development team in Americanas to identify potential data issues before they become a problem. Notifications are sent out where needed so relevant people can take action. Data quality checks have also been applied to machine learning streams to ensure AI and ML models don't go rogue due to input data issues. The time to value with Soda Cloud is rapid - when used to develop models to ensure the accuracy and security of the data used for crucial decisions during Black Friday, an important ecommerce period, it only took four weeks to launch a pilot.
.svg){kind=avatar}
