Challenge
Our client owns a Supply Side Platform that processes millions of ads on a daily basis. With ad campaigns coming from multiple sources, having different priorities, types and schedules, managing pacing of such campaigns is a non-trivial task. The client was looking for a partner that could help them understand their own data and enable their AdOps team making sound decisions on scheduling a growing number of campaigns, increasing revenue without risk of them not being fulfilled in time.
Solution
We performed a thorough analysis of raw client data and created a data aggregation pipeline that cleans and extracts most valuable information for further processing. Based on that data, we provided our client a thorough documentation and new insight into their own data and traffic patterns. Going forward, we used both simple statistical models and neural networks (DeepAR) to create a forecasting engine that can be used by AdOps for predicting future campaign spend based on historical data.
Results
Our client gained additional insight on their production traffic and secured further growth thanks to new forecasting solutions that aid decision making and operating their business at even larger scale.