A geodatabase doesn't provide much flexibility in organizing your data hierarchically. All feature sets live either at the top level of the database, or one level down in feature datasets. A feature dataset can't, in turn, contain another dataset. I'm interested in hearing how others organize their data, especially in environments where they must use a geodatabase and are stuck with these limitations.
Typical organization methods that I've seen are by prefixing. In an enterprise geodatabase, this can be done with an owner/schema name (e.g.
UTIL.WMETER or by prefixing the featureclass/table name itself with a prefix string, then underscore, then name
UTIL_WMETER). Prefixing the featureclass/table name also works in personal/file geodatabases.
Because of editing overhead for feature datasets, they aren't usually used as a primary organizational mechanism.