Demand forecasting at scale with Foresient on AWS

For future-forward business success, enterprises must leverage current data to prepare for the future. With unprecedented access to data, the key is flexibility—understanding current insights effectively and accurately predicting future trends. Equipped with the right forecasting tools, enterprises can gain a significant edge over competitors by anticipating and adapting to upcoming changes.

Tetra Pak, a world leading food processing and packaging solutions company, was looking for higher category forecasting accuracy, with a solution that would provide insights across multiple categories across E-Commerce markets, over a period of five years.

E-commerce growth and sustainability

With E-Commerce channels promising to deliver high growth in line with their sustainability ambitions, they wanted to know more about how this could evolve alongside other developments in the Fast-moving Consumer packaged goods (CPG) industry.

Accurate data interpretation

Our knowledge of developing bespoke solutions through our integrated cross-market AI-based suite of algorithms meant we were ideally placed to work and help them understand the stories their data could tell them.

We started by running discovery sessions to look to understand the client requirements. This helped us to build our knowledge of the business and the challenges they faced. The pre-processing work we do, including sanity checks and data harmonisation, gives our partners extra confidence in the inputs than they may have been used to with simple or manual forecasting.

Unlocking business insights with Foresient on AWS

Fractal’s Foresient, a state-of-the art platform on AWS enabled high accuracy forecasting at speed and scale. By combining historical reporting and other associated data with the Foresient engine we were able to generate business-relevant forecasts. Optimised solutions via end to end forecasting operations, helping organizations to better navigate their markets.

Service deployed

• Amazon Web Services EC2
• EMR based on Spark
• Amazon Relational Database Service
• S3

Objective

• Provided scalable compute capacity for hosting the front end, back end, and APIs of the application.
• Used to process and analyze huge volumes of data.
• Provided a cost-effective database.
• Stored information in the form of forecasts and models.

Outcome

• Scalability and flexibility
• Efficient data processing
• Cost optimization
• Reliable data storage

Foresient generated forecasts across multiple categories across E-Commerce markets, over a period of five years, with improved accuracy.