Measuring and tracking building energy use and other data is essential for cities, utilities, and service providers. Currently, buildings are conceived as "places/locations" and therefore are identified with a street address or with a single geographic coordinate point (latitude and longitude). However, the address-based system is insufficient to uniquely identify buildings, and to append and join data from disparate sources due to address name duplicates, incorrect entries, and lack of an input format standard. As a result, essential building energy data for energy efficiency policy decision making is difficult to leverage. The Unique Building Identifier (UBID) was developed as a natural key to match building energy and attribute data between databases and to facilitate matching across datasets. Our methodology converts a two-dimensional building footprint into a unique alphanumeric code based on an open source grid reference system. The UBID is designed to be easily overlaid on publicly available, web-based mapping platforms and enable fast and accurate spatial matching with minimal computational resources. In this paper, we present the UBID methodology, discuss the use cases in energy efficiency, and present the results of tests in three major US cities (San Francisco, Chicago, and New York City) to verify the uniqueness and effectiveness of the proposed building identifier.