A significant part of data and content management in e-business scenarios deals with exchanging product-related data between business entities, and integrating them into target applications (e.g. ERP systems) or target documents (e.g. e-catalogs) at the recipient?s side. Content integration tasks can be much better automated if the textual descriptions are augmented by a machine-readable representation of the semantics. For this purpose, categorization standards for products and services, like UNSPSC, eClass, eOTD, or the Rosettanet Technical Dictionary (RNTD) are widely in use. Existing research, however, has focused on the architecture and structure of such standards, and did not investigate their actual content. In this paper, we present a framework of metrics for the quality and maturity of categorization standards, and apply these metrics to eClass, UNSPSC, eOTD, and RNTD. The results clearly show weaknesses which hamper the use in many application domains. Also, we can reveal that only some of these standards are actually maintained and updated, while others are rather inactive, dead collections.