AI for demand forecasting

By Christine Babington Smith on July 23, 2020

Business decisions rely on being able to forecast the future. The ability to accurately predict demand can help CPGs achieve efficiencies in inventory, warehouse and logistics optimization.

Unfortunately, we humans are not born with the power to see into the future. Forecasting is a very unpredictable science. Get it right and you achieve competitive advantage through cost leadership. Get it wrong and you tie up working capital in inventory; understocking leads to lost sales whilst overstocking ends up in unnecessary discounting, wastage and lost margin.

If the world was stable, this would not be such a problem as it would be relatively easy to predict – but it’s not. The reality is that we live in a VUCA world. with increasingly volatile demand affected by changing consumer habits and attitudes. The shift to the digitization of marketing channels, coupled with consumer preferences for socially responsible products and personalization, has already increased the complexity of forecasting. And this was before the black swan of COVID-19 made its appearance.

These trends have been amplified by the pandemic, with 62% of consumers now saying ESG concerns are more likely to influence their buying behavior, according to EY. It’s also more important than ever to have the right product, in the right place at the right time given that consumers’ expectations of convenience have increased. An out-of-stock is a frustrating experience for both CPGs and consumers, and does not engender brand loyalty.

How do CPGs adapt to this fast-paced world? By adopting the predictive qualities of AI, CPGs can become more agile. With this in mind, we’ve teamed up with Amazon Web Services (AWS) to produce a solution that AWS claims is 50% more accurate in predicting demand than traditional forecasting techniques. We’ve integrated AWS’s Forecast algorithm (which powers Amazon’s own retail offering) into our Decision Intelligence platform, making it easily accessible so you don’t need to be a data scientist to use it.

The flagship tool is an easy-to-use application for demand planners which uses an ensemble of techniques, including sophisticated deep learning algorithms, to accurately forecast demand. We created it as part of our mission to develop accessible AI that’s easily adoptable for everyone, so all companies can benefit from the technology. It’s an opportunity to inject the power of Amazon’s formidable 20 years’ experience and expertise in AI into your own demand forecasting abilities.

Our customers are already seeing the benefits of the predictive ability of AI, with Speedy Hire achieving an 18% reduction in inventory levels based on the ability to predict what, when and where their inventory is needed.

Peak’s AWS Forecast is designed to help inform decisions on inventory optimization, manufacturing scheduling, allocation and assortment, demand and labor planning, warehouse optimization, logistics optimization and Distribution Resource Planning.

Book a demo of our Dynamic Inventory AI application to find out more about AI for demand forecasting, and the role it could play in your business. Check out our demo to learn more about AWS Forecast and our platform ?