Optimizing volume and frequency forecasts for an online video advertiser

Abstract

Videology is an online advertising company with a targeted advertising platform that more efficiently connects brands with target consumers. By obtaining and utilizing user data, advertisers have targeted specific groups through a process known as behavioral targeting. This process increases the odds that a user will click on an advertisement and reduces the odds that a customer will encounter irrelevant advertisements. While Videology has grown their market share in this space, inefficient forecasts have cost several hundred thousand dollars in lost opportunity costs. This paper addresses this problem by leveraging a systems engineering approach to suggest procedures for optimizing the validity and performance of two forecast variables for Videology. The paper first analyzes a volume forecast variable: the expected number of visitors, and second, a frequency forecast variable: the number of times a visitor comes back to the same website. Research used existing data to construct a pseudo-process to replicate Videology's algorithm in order to test the validity of and make enhancements to forecasts. Videology will then utilize findings from this process to continue forecast improvements.