Interactivity is an important feature of retail websites. Some research suggests that interactivity can be increased by implementing additional features while other studies point to the fact that “objective” interactivity - as measured by the mere number of features - is quite different from perceived interactivity (PI). Interestingly, while there seems to be initial evidence about the consequences of PI, a comprehensive assessment of attitudinal and behavioral outcomes of PI as well as the impact of PI on (objective) website and firm performance using field data has not been offered. Our study attempts to fill that void. Based on objective performance data of the top 500 retail websites over two years as well as ratings on 35 interactivity features, we find the majority of objective interactivity indicators to be unrelated to PI. Perceived responsiveness and perceived control relate positively to retailers’ sales differences. Perceived communication and objective responsiveness show negative indirect effects on sales, giving rise to the assumption that unobserved heterogeneity of the retailers’ websites might explain these counterintuitive findings.