Clusters are geographic concentrations of industries related by knowledge, skills, inputs, demand and/or other linkages. There is an increasing need for cluster-based data to support research, facilitate comparisons of clusters across regions and support policymakers in defining regional strategies. This article develops a novel clustering algorithm that systematically generates and assesses sets of cluster definitions (i.e., groups of closely related industries). We implement the algorithm using 2009 data for U.S. industries (six-digit NAICS), and propose a new set of benchmark cluster definitions that incorporates measures of inter-industry linkages based on co-location patterns, input–output links, and similarities in labor occupations. We also illustrate the algorithm’s ability to compare alternative sets of cluster definitions by evaluating our new set against existing sets in the literature. We find that our proposed set outperforms other methods in capturing a wide range of inter-industry linkages, including the grouping of industries within the same three-digit NAICS.