Advertisement Sales Revenue Prediction Forecast
Successful in-house collaboration between Gruner + Jahr and Arvato Systems
Gruner + Jahr contacted Arvato Systems to create their vision to use Machine Learning (ML) and other new technology to develop two forecasting prototypes. In a fantastic joint partnership between the two Bertelsmann subsidiaries, both prototypes were successfully developed in a short time of only six weeks. Our sister company G+J, as well as the Arvato System development team, are both pleased with this project as both teams focused on an open dialog and agile approach.
Working with Arvato Systems went very well. We exchanged ideas regularly, so we could combine expertise and AI-knowledge to get results quickly. The forecasts help us to predict revenue trends, and we want are planning to go beyond the AI Lab as the next step and further optimize the forecasts.
The first developed prototype was out of five ideas to use ML in their daily business, and it focuses on the prediction of advisement revenue to smartly leverage advertisement sales on their increasing online presence. With this prediction prototype, G+J can now foresee potential revenue streams for each media type, e.g., per magazine brand for their online presence as well as traditional print. Using historical data to build the ML, which is used on regression and rule-based holistic, created a very accurate prediction model.
The second prototype idea was also part of the first brainstorming of how to improve daily operations with the use of new technologies. Similar to the first prediction model, this prototype also focuses on advertising revenue, but instead of predicting existing client revenue, this prototype focuses on new customer potentials. With interpreting current and historic external market data, such as how a potential new client uses the current market for advisement of their brand, this prediction tool points out potential new customers. By matching either the customer segments/target groups or comparing existing advertisement strategies of companies, this prototype can identify what brand may be a potential customer for a specific product.
This project was a great example of agile communication and teamwork extended through Bertelsmann divisions.
After a very successful prototype phase, both prediction models are going to be further improved before they are going to be implemented in their daily business in the upcoming year. The constant and agile exchange between both Bertelsmann companies showcases a fantastic collaboration, which is going to be developed further in 2020 and, hopefully, even after that.
Your Contact for Our Insights Digital Innovations Newsletter