What a firm does is more revealing than how much it makes, but firms are often described with metrics for economic size. Instead, we characterize what firms know in terms of what they read, the information footprint, using a data set of hundreds of millions of records of news articles accessed by employees in millions of firms.

We discover that the reading habits of firms are of limited diversity. This observation suggests that information constraints act on firms. To understand how, we relate a firm’s information footprint with economic variables, showing that the former grows superlinearly with the latter. This exaggerates the classic Zipf’s law inequality in the economic size of firms and reveals an economy of scale with respect to information.

Second, we reconstruct the topic space firms inhabit, finding that the space resembles a tangled “hairball” with a few dense knots of topics and many specialized strands sticking out. Some of the topics are ubiquitous, indicating inescapable demand regardless of firm size. Finally, we connect these pieces in a model of how firms grow in the space of topics.

We show that diversity in firm reading habits can be explained by a mixed strategy of local exploration and recurrent exploitation on the topic graph. This shows that the constraints that the space of ideas imposes on firm growth provide a useful and new perspective on firm development.

E. Lee, A. P. Kwan, R. Hanel, A. Bhatt, F. Neffke, Following The Information Footprint Of Firms, pre-print (2022).