Since 1991, Denver Water's Planning Division has been working to build GIS databases to assist in the management of its water collection system. These management activities include State-wide and regional efforts such as surface and groundwater resource analysis, environmental planning, and facility permitting. They also include Denver Metro and urban efforts such as analyzing land use patterns and developing cooperative water planning strategies with neighboring water distribution entities. To use geographic information systems technologies in these endeavors requires the development and maintenance of appropriate GIS databases. The wide and varying spatial extent and numerous thematic layers needed to properly address all of these activities have often required that Denver Water turn to other organizations, typically federal, State, and county entities, to provide GIS data sets. These data sets can at times be used just as they are provided, but more often require additional processing to become viable GIS databases. The reasons for additional database processing are many and varied. They include procuring data that is spatially rich but informationally poor, procuring and synthesizing one GIS layer from several overlapping and inconsistent versions of the layer from multiple entities, and creating seamless layers from spatially adjacent data files when the source data is widely inconsistent. Often our experiences have led to as many questions about data reliability as they have led to the stated goal of building and maintaining appropriate GIS databases. Our successes and failures in these endeavors are discussed as well as our observations and ideas on future options for GIS data being created and managed in the public sector and provided to the GIS user community.