The geodetic profession provides geodetic data obtained in the fields of fundamental geodetic system, real estate and topography Users integrate this data with other spatial data. The quality of such applications depends on the data quality; however, the providers and the users are not sufficiently aware of the cascading transfer of errors from the primary data layers to the secondary, derived data. Geodesy has traditionally considered errors mainly as positional errors, according to the law of the transfer of errors. Since the era of widespread digitizing of geodetic data began in the 1990s, data quality in the realm of geographic information systems has been expressed with the standardized parameters of data quality and with metadata, of which positional accuracy is just one type. Unfortunately in this way, the knowledge about which data is a reference can be obscured. To many users, it is not clear what is the specified quality referred to. This article describes, and intentionally uses, the term `cascading' since for the rise of data quality one needs to upgrade data level by data level, starting first with the reference data and the semantic definitions of object types, and continuing by updating and harmonizing the secondary data, which were developed from the primary level. The reasons for the change of paradigm regarding the treatment of quality are considered. As the cascading poor quality of geodetic data can become a serious threat for the reputation of the profession of geodesy, several contemporary applications and projects are described in which this has already happened. In the conclusion, proposals are given for improving the situationion.