Combating Empty Shelves: Machine Learning to the Rescue

Stock out forecasting has troubled consumer goods and retail analytics teams for decades. Out-of-Stock (OOS) events are not easy to predict. Aggregate sales data is insufficient for demand forecasting. In addition, many factors can contribute to Out-of Stock / Stock Out events, e.g. promotions or under stocking shelves to keep inventory low for the retailer. Traditional techniques outlined in this Inventory Forecasting Guide do not resolve the perpetual tension between retailer and consumer goods company. The retailer wants to keep inventory low, throughput high, and drive repeat traffic in the store. The consumer goods company wants to keep volume and price high. 

Traditionally, demand forecasting leverages both qualitative and quantitative methods. Qualitative methods include collaborative forecasting, market surveys, or delphi forecasting which rely on interviews and to some extent informed “opinion” from experts. Analyst teams then buttress the qualitative analysis with statistical methods.

A survey of corporate retail professionals conducted by Wakefield Research and Bossa Nova Robotics found

73% of respondents consider inaccurate forecasting 'a constant issue' for their store. Another 66% said the same for price inaccuracy. 65% said they struggle with the ability to track inventory through their supply chain. Click To Tweet

The increase in online channels vs traditional Big Box retailing complicates forecasting models even further. Do online channels cannibalize or impact sales in brick and mortar retail? If retail inventory is kept low does that drive traffic online? For example, if a store is out of an item, with ease of mobile shopping like Amazon or Target, do customers order right in the store from their phone? Are there ways to encourage this behavior to solve the stock out problem?

Machine learning has now joined the party to bring a fact-base alignment to the dance between product manufacturers and retailers or ecommerce channels they sell-through. Google Cloud for Retail provides an entire platform built in the cloud for retailers to benefit from an army of PhD data scientists for real-time inventory management and intelligent analytics tools.

In this podcast interview, Prabhu Palanisamy, SpringML co-founder, shares how we are working with major consumer brands and retailers to enhance forecasting and inventory management with AI and Machine Learning.

Join Prabhu’s webinar, “Retail Use Cases for Machine Learning” March 19th, 2020. Prabhu will cover how retailers are adopting AI to redefine core processes such as forecasting, product recommendations, product search, and pricing to improve operations.

If you are growing your business and want to stay competitive in this rapidly changing marketplace with the latest in analytics, Machine Learning, and AI, contact us at info@SpringML.com or tweet us @springmlinc.