Library and museum staff members are faced with having to create descriptions for large numbers of items found within collections. Immense collections and a shortage of staff time prevent the description of collections using metadata at the item level. Large collections of photographs may contain great scholarly and research value, but this information may only be found if items are described in detail. Without detailed descriptions, the items are much harder to find using standard web search techniques, which have become the norm for searching library and museum collection catalogs. To assist with metadata creation, institutions can attempt to reach out to the public and crowdsource descriptions. An example of crowdsourced description generation is the website, Flickr, where the entire user community can comment and add metadata information in the forms of tags to other users' images. This paper discusses some of the problems with metadata creation and provides insight on ways in which crowdsourcing can benefit institutions. Through an analysis of tags and comments found on Flickr, behaviors are categorized to show a taxonomy of users. This information is used in conjunction with survey data in an effort to show if certain types of users have characteristics that are most beneficial to enhancing metadata in existing library and museum collections.