Journal rankings are widely used and are often based on citation data in combination with a network perspective. We argue that some of these network-based rankings can produce misleading results. From a theoretical point of view, we show that the standard network modelling approach of citation data at the journal level (i.e., the projection of paper citations onto journals) introduces fictitious relations among journals. To overcome this problem, we propose a citation path perspective, and empirically show that rankings based on the network and the citation path perspective are very different. Based on our theoretical and empirical analysis, we highlight the limitations of standard network metrics, and propose a method to overcome these limitations and compute journal rankings.