Marketing & Sales
Unlocking sales insights through audience modeling in gaming
In the fast-evolving gaming industry, understanding the impact of streaming on sales is critical for strategic decision-making. Agilytic partnered with a leading video game industry company to develop audience modeling techniques, enabling data-driven sales strategies and rapid market adaptation.
To protect confidentiality, we may alter specific details while preserving the accuracy of our core contribution.
Context & objectives
Agilytic collaborated with a company that specializes in:
collecting information related to the video game industry
providing B2B e-commerce solutions to support their operations
The primary goal of this contract was to gain a deeper and more comprehensive understanding of how streaming activities impact sales performance in the gaming sector through audience modeling. They needed effective strategies that would enable quick decision-making processes and facilitate rapid adaptation in their selling approach and methodologies.
Approach
For this project, we used a Python library called "Darts: Time Series Made Easy.". Darts provides various models and utilities to work with time series data, making it easier to perform tasks such as forecasting future values, evaluating model performance, and manipulating time series datasets.
With this tool, we proceed in 3 steps :
First, we validated the scope by assessing the quality of data and identifying any time series that may have lower significance, which we then excluded.
Next, we developed a promotions model. This model is designed to forecast sales and predict the potential impact of promotions, allowing us to identify any unexplained spikes in sales that may be caused by external factors.
Finally, we investigated sales peaks that could not be solely attributed to promotions and incorporated audience data features into the model.
We also created a machine learning algorithm to consider the impact of promotions on gaming sales, along with a report detailing our approach to forecasting video game sales over the long term using audience modeling techniques.
Results
Key deliverables included:
A machine learning model to control for promotional effects on gaming sales
A tool to detect streaming peaks that influence sales
An analysis of how streaming impacts sales during pre-orders and the launch phase
A report outlining the steps for long-term video game sales forecasting
Reports identifying key streamers with significant influence on video game sales
Through effective audience modeling, the company can now adapt its sales strategy and gain deeper insights into industry-wide sales performance.
To safeguard confidentiality, we may modify certain details within our case studies.
