bonprix: Enhancing Shopping Experience with Scalable ML and AI Automation via Google Cloud

• Analytics & Modeling - Machine Learning
• Functional Applications - Warehouse Management Systems (WMS)

• Consumer Goods
• Equipment & Machinery

• Product Research & Development
• Warehouse & Inventory Management

• Predictive Maintenance
• Time Sensitive Networking

• Cloud Planning, Design & Implementation Services
• Data Science Services

bonprix, an online international fashion and homeware retailer, was facing challenges with its traditional on-premises data warehouse setup based on Teradata and a data lake based on Cloudera Hadoop. As the amount of raw data the company collected grew, storage costs rose. The company also had to deal with the complexity of managing two different systems for structured and unstructured data. This not only made data comparison difficult but also hindered the enhancement of their machine learning and artificial intelligence capabilities. bonprix wanted a solution that would allow all their data to be in the same place, be more cost-effective, and enhance their machine learning and artificial intelligence capabilities.

bonprix is a Hamburg-based online retailer that sells quality affordable fashion and homewares. The company operates across Europe, Russia, and North America. Founded in 1986, bonprix initially focused on mail-order fashion before launching its first online shop in 1997. Today, the company has a 3,000-strong workforce that supports more than 35 million global customers, with 90% of its revenues coming from ecommerce. bonprix has strong in-house development and data engineering teams that are always looking to maximize the potential of available and developing technology to optimize its business goals.

bonprix decided to migrate to Google Cloud, with the help of Google Cloud partners KPMG and Aliz. The migration began in March 2020 and was completed 12 months later in March 2021, on time and on budget. With Google Cloud, bonprix now has all its structured and unstructured data in one place, making it easier and faster to analyze and integrate the data for machine learning. The company also found the new setup better suited to testing new use cases. The time it takes to develop new ML models has been significantly reduced, now taking just two weeks. bonprix has also seen its data storage costs drop. Looking ahead, bonprix plans to integrate Google Cloud with the external workflow solution Apache Airflow and use Google Cloud management tools to get insights into its data analysis, datastore, databases, and developer services.

• The migration to Google Cloud has allowed bonprix to shift its IT team's focus from infrastructure management to service creation. The company has been able to test and deploy new machine learning models more quickly and efficiently. It has also been able to build up its machine learning and artificial intelligence capabilities, creating a recommendation engine that recommends similar products to customers based on what they click on, a fraud-detection system, and a programmatic advertising system that forecasts the return on investment for different advertising spots. The company is also developing other machine learning-based use cases. With Google Cloud taking over the management of much of bonprix's infrastructure, the company's developers and engineers can now focus on creating new data-based products and getting them to market faster.

• Migration to Google Cloud completed in 12 months, on time, and on budget

• Significantly reduces time to develop new ML models to just two weeks

• 40% to 50% more cost-effective than on-premises solution