How online beer retailer BEERWULF safeguards optimal stock levels, using AI & advanced analytics for inventory demand forecasting.
Beerwulf's challenge as a fast-growing online retailer in food & beverages
Since its launch in 2016, Beerwulf has experienced tremendous growth.
The company was in 2019 named the fastest-growing international e-commerce retailer in the Netherlands. In Heineken's 2020 half-year results, the company noted that its daughter company 'had more than 3 million visitors, half of them new. Online sales of our home-draught systems like the Sub and Blade have more than doubled during the lockdown.'
Initially launched in the Netherlands, the company quickly expanded to other countries such as Belgium, Germany, France, and the United Kingdom. Currently delivering to more than 10 countries, Beerwulf is on its way to becoming Europe’s leading online beer store.
Supporting this rapid growth from an operational perspective is an extremely challenging task, even for a company that is being backed by a behemoth such as Heineken. Beerwulf offers over 1.000 different beers in bottles, cans, packs, and draught systems. Managing inventory for all of these SKUs across various warehouses and markets is tough, especially given the described growth scenario.
Beerwulf contacted Crunch Analytics to help tackle this challenge. The key objective was to build a data-driven solution that can optimize and automate a large part of the inventory replenishment process, using an advanced demand prediction/demand forecasting model
More accurate demand forecasting using big data & AI-techniques
Within this domain, the retailer's key objective is to optimize the balance between inventory cost and the opportunity cost of lost sales. So ordering the right amount at the right time is crucial. Such depends on being able to estimate what is going to sell. So, in essence, it's 'an estimates game.'
This is exactly why the latest advancements in analytics, using big sets of data & AI-techniques, are so important for retail & e-commerce. It allows for the creation of so-called 'prediction machines.' Tools that can embrace and process significantly more relevant information, therefore allowing for significantly more accurate predictions.
What tool did we create?
Working from the business user's perspective, we created an easy-to-navigate Power BI interface that combines a number of actionable dashboards. The main dashboard provides an overview of recommendations on when to replenish a given item and what quantity to reorder. Other dashboards allow that same user to dive deeper into details & insights.
In the background, a demand forecasting model and optimization model - that takes into account specific business rules and operational constraints - work their magic.
In a 4-step journey, we went from an inspiration & discovery workshop to a fully-functioning tool. One that enables the team to cut repetitive tasks and flawed decisions that led to overstock or out-of-stock situations. A single user interface, a few clicks of a button, and both tactical & strategic insights now ensure that the team can excel at what it does best.
Download the detailed client case here!
Learn more about how we helped to solve Beerwulf's challenges in managing inventory as an online retailer amidst rapid growth. Find all the details in this extended client case.