AI Case Study

Otto reduces the rate of product returns by predicting sales for the next three months with 90% accuracy using machine learning to understand consumer preferences

Otto, the German e-commerce giant, saved millions of Euros by accurately predicting customer demand and using it to plan inventory. They were looking to reduce losses caused by product returns and found that customers prefer to receive all the items in one shipping, within two days. The only way they could do this was by predicting which items would be sold next month accurately. Otto's solution has been able to predict demand with 90% accuracy.

Industry

Consumer Goods And Services

Retail General

Project Overview

"Otto discovered that the rates of return is higher for products for which the shipping time is more than two days and when there are multiple shipments. Applying Blue Yonder's deep-learning algorithm, around 3bn past transactions and 200 variables (such as past sales, searches on Otto’s site and weather information) were analysed to predict what customers will buy a week before they order.

The AI system predicts with 90% accuracy what will be sold within 30 days—that Otto allows it automatically to purchase around 200,000 items in a variety of products, colors and sizes that the machine orders a month from third-party brands with no human intervention. Customers get their items sooner, which improves retention over time, and the technology also benefits the environment, because fewer packages get dispatched to begin with, or sent back."

Reported Results

According to the company:

* 90% accuracy in predicting products that will be sold within the next 30 days enabling Inventory optimisation
* Inventory surplus is reduced by 20%
* Number of returns are reduced by more than 2M

Technology

Blue Yonder's deep-learning algorithm

Function

Strategy

Data Science

Background

Otto was losing millions of Euros every year in shipping costs due to customers returning products frequently.

Benefits

Data

"Around 3bn past transactions and 200 variables (such as past sales, searches on Otto’s site and weather information)"

"Delivers probabilistic forecasts based on hundreds of different variables including weather, promotions, and holidays. This allows the business strategy to automate millions of daily replenishment decisions across products and stores."