The Cost Impact of Spam Filters: Measuring the Effect of Information System Technologies in Organizations

Abstract

More than 70% of global e-mail traffic consists of unsolicited and commercial direct marketing, also known as spam. Dealing with spam incurs high costs for organizations, prompting efforts to try to reduce spam-related costs by installing spam filters. Using modern econometric methods to reduce the selection bias of installing a spam filter, we deploy a unique data setting implemented at a German university to measure the costs associated with spam and the costs savings of spam filters. The applied methodological framework can easily be transferred to estimate the effect of other IS technologies (e.g., SAP) implemented in organizations. Our findings indicate that central IT costs are of little relevance since the majority of spam costs stem from employees who spend working time identifying and deleting spam. The working time losses caused by spam are approximately 1,200 minutes per employee per year; these costs could be reduced by roughly 35% through the installation of a spam filter mechanism. The individual efficiency of a spam filter installation depends on the amount of spam that is received and on the level of knowledge about spam.

Keywords: propensity score matching, treatment effects, spam filter, spam

JEL Classification: M12, M15