Current U.S. federal mapping standards call for the use of the Geographic Names Information System (GNIS) point layer for placement of all United States populated place labels. This point layer contains both limited and duplicate classification information and therefore complicates the balanced labeling of place with congestion at smaller scales. In turn, this affects map quality for database ‐ driven, multiscale, reference mapping, such as maps served on The National Map Viewer from the U.S. Geological Survey (USGS; a U.S. national mapping agency (NMA)). Using heterogeneous and multiple data sources, our research investigates alternatives for the automatic generalization and creation of label hierarchy within NMA reference maps in order to accurately reflect the relative importance of point features such as populated places and airports.