How to reduce stockouts by more than 80%: the new case for Profit Whales

The Profit Whales’ marketing agency works with a sophisticated ML-powered framework that predicts the supply and demand for Amazon's fastest-growing brands.

Background

Profit Whales is a full-service marketing agency for Amazon brands. It accelerates sales growth using on & off Amazon traffic to give brands an edge over competitors, diversify revenue streams, and optimize the process of evaluating a business’s performance.

Profit Whales achieves growth for clients via its omnichannel approach, A/B tests, analytics, and diversion. The fundamental goal is to build and implement a tailored marketing strategy for Amazon e-commerce brands.

In 2022, Profit Whales reached out to Datrics to help develop a full-featured machine learning system for sales forecasting. The solution became part of the Profit Whales service portfolio after proving its worth by leading to customer success.

“For a long time, we had been thinking about expanding our portfolio of services by adding shopping activity forecasting to it. Today, artificial intelligence technologies have made accurate forecasting possible, which significantly reduces stockouts. For this purpose, we needed a trusted partner with expertise in ML technologies”, says Olha Fedirko, Lead Data Analyst at Profit Whales

The Challenge

Demand forecasting is an essential problem to solve for a thriving retail business. Increasing the accuracy of predictions ties into performance improvements in critical business processes:

  • financial planning and goal setting
  • assortment management
  • promo campaigns and pricing planning
  • prevention of stockouts

Profit Whales wanted a solution to predict the demand for goods 3-4 months into the future. Based on this information, their clients would be able to prepare an accurate logistics schedule and stock the right amount of inventory in each warehouse.

Challenges for building the systems encompassed requirements from the side of business processes as well as from the ML model itself:

  • Clients had to have enough data to train the ML Model
  • High requirements for the ETL process: cleaning, structuring, and transferring data to be fed into the Model
  • The ML model had to account for dozens of factors affecting buyer behavior. To name a few: seasonality, price changes, holidays, and most importantly, marketing campaigns types and budgets

The Solution

Based on the client's requirements, Datriсs helped develop an ML model that integrates with the client’s database, executes data analysis and processing, forecasts future sales, and stores the results in the database for further visualization and reporting.

The ML model accounts for factors such as seasonality, the demand impact of discounts or price surges, increases or decreases in demand for certain groups of goods, trends, marketing campaign types, budgets, and much more. The model can perform various scenario analyses and provide optimal approaches to budgeting, forecasting, and stocking. Such predictions allow retailers to plan warehouse deliveries within a one-quarter horizon, significantly optimizing marketing costs and, most importantly, eliminating out-of-stock situations.

The presentation of the data obtained as a result of the work of the model takes place in the Profit Wales interface. The client can plan deliveries to a warehouse along with other marketing activities.

The Results

Based on Datrics’ consultancy experience and tight cooperation with business owners from the Profit Whales side, an advanced ML model was developed and put into production. The model’s forecasting error margin stays within 10% across different products when calculated on an aggregated monthly basis. The actual monetary savings varies from customer to customer. Yet it is indisputable that decreasing stockouts while optimizing marketing approaches and budgets is crucial for any e-commerce business.

As for Profit Whales, the company reports that their clients see an average decrease of stockouts of over 80% with Datrics' ML model compared to a more traditional approach.

“The decision to work with Datrics came quite easy to us. As part of a risk-free pilot program, these guys helped us develop an ML system in the shortest possible time, helped with integrations, and streamlined and systematized the collection, segmentation, and transfer of data between databases. We highly recommend the Datrics team as professionals in the data and AI niche”, sums up Olha Fedirko.

Do you want to discover more about Datrics?

Read more

A Bank’s Journey to Simplified Data Analytics

Our Client, a leading European bank, faced a challenge. Despite having a large analytics team, they had only a handful of data scientists.

Zedosh's Attention Exchange® Engine Powered By Datriсs For Purpose-Driven Advertising

Zedosh and Datrics are changing the landscape of online advertising to get ready for a post-cookie internet, with their targeted ad algorithm.

Streamlining LTV Calculations for a prominent European Bank

The Client, an European bank faced a significant challenge. Their existing process for calculating the LTV of customers was cumbersome, manual, and reliant on tools like Excel.
BACKED BY