DETECTING POOR-QUALITY WEB TRAFFIC IN VIDEO AD MARKETING CAMPAIGNS BASED ON ANALYTICAL METHODS

In digital marketing, performance data can be used to determine which variables affect the quality of web traffic generated. There exist different ways of collecting data, which enable the data enrichment and give us a lever to generate new ways of looking at and dealing with information. And yet, a large amount of data, the daily variation and the different levels of granularity require dedicated data preparation. Our data needs prior preparation and cleaning to be able to develop the multivariate and correlation analysis through Cramer's V-diagram. Our objective is to analyze and predict the traffic quality score, a parameter that is not only used by companies to verify the suitability of the audience concerned but is also directly related to the bandwidth awareness and profitability of video advertising campaigns. Analytical identification of the variables that most affect traffic quality as well as the quality prediction can be used to optimize both marketing campaign structure and bidding strategy.