Sales on e-commerce websites cannot increase if visitors leave the websites without purchasing. To better explain and increase buyer conversion and user satisfaction, it is needed to better understand why Internet shoppers give up and exit websites without achieving their goal of purchasing desired products. First, a compendium of website features which are thought to improve buyer conversion is provided. Second, the current theory, minimal cognitive processing theory (MCPT), is summarized. The MCPT applies Newell and Simon's cognitive processing theory from human computer interaction to user behavior on retail websites. It proposes that two latent variables, clickstream cost and personal cost limit, are responsible for a user staying on or leaving a website. Clickstream cost is the amount of cognitive processing required to reach a given page in the site, which is operationally defined by counting the number of site features and applying the Hick Hyman Law. Personal cost limit is the total amount of cognitive processing that the visitor is willing to invest in the site. When the clickstream cost exceeds his/her personal cost limit, the visitor exits unless he/she has already found and purchased the desired product. From this it is deduced that one of the major latent variables responsible for website abandonment is website complexity (amount of cognitive processing required by the site). Third, an empirical study evaluating the MCPT is summarized. In this experimental study two websites were created: a high-complexity website, with all of the features provided as defaults by the site building software osCommerce, and a low-complexity website, with minimal features. The websites were otherwise identical. In total, 120 participants (each buying two products on both sites) were used. Four major hypotheses were deduced from the MCPT: the higher the complexity, (a) the lower the number of visitors making successful purchases, (b) the greater the amount of time required for purchase, (c) the greater the number of visitors voluntarily exiting the sites before completing any purchase, and (d) the lower the user satisfaction. All major predictions from the MCPT were confirmed with p = .000. Secondary hypotheses suggest the effect of knowledge of B2C site design, self-reported experience in Web purchasing, and learning across websites, as suggested by the MCPT. Fourth, plans for generalizing the theory and further empirical research are summarized.