The interlocking world city network model and other office location approaches (OLAs) have become the most widely used empirical models of the world city network (WCN). Despite numerous methodological improvements, they continue to rely on a legacy of using data on office locations of firms to indirectly estimate intercity business flows. To advance the dialogue about how to improve on existing empirical models of the WCN, we examine the content, construct and structural validity of OLAs. We analyze the link between the theoretical construct of intercity business flows and network projections obtained from office location data and uncover evidence that calls into question the validity of OLAs as empirical models of the WCN. In the spirit of no deconstruction without reconstruction, we then develop an alternative empirical model of the WCN, based on directly observable relational ties among APS firms, which are formed through co-production of complex services. We call this the inter-organizational project approach (IOPA). We argue for IOPA's construct validity as an empirical model of the WCN and offer empirical evidence for its structural validity. We demonstrate it using a global sample of 161,114 investment bank syndicates in the 2000–2015 period.