By: Brian J. Stewart
An effective taxonomy design is critical to the success of any enterprise document management strategy. This article will focus on several do’s and don’ts for an effective taxonomy design. These guiding principles will promote data consistency, data integrity, and data interoperability across Enterprise Document Management Systems (EDMS).
|Consider taxonomy beyond application boundaries – Too often taxonomy design is viewed as a solution implementation activity on a project level, rather than a strategic architecture activity with cross-system and cross-project impact. Rather than develop taxonomies in isolation, organizations should develop taxonomies using consistent content classification and metadata models across enterprise systems.|
|Harmonization of all dictionary data – Organizations need to move to a taxonomy design that leverages an effective Master Data Management (MDM) strategy. In addition to a common taxonomy, enterprise systems need to use a consistent vocabulary or leverage the same data dictionaries to enable Enterprise application integration (EAI) and Service-Oriented Architecture (SOA). For example, all enterprise systems should use the same 'Therapeutic Area' values to ensure interoperability and enable application integrations.|
|Consider specific needs of all content consumers – It is important to consider all consumers of content when developing a taxonomy. Each consumer has their own set of unique requirements to locate and leverage content stored within an EDMS repository.|
|Integrate with other systems when metadata also resides in other systems – Effective taxonomies consider the flow of information through the entire business process which generally encapsulates multiple systems. An effective design eliminates duplicate data entry while allowing information to flow through systems.|
|Ensure all documents have a unique (non-system) identifier – All EDMS implementations have a unique system identifier or ‘Object Id’. These identifiers are unique within a repository, but organizations should employ a strategy to ensure all documents have a unique business identifier for documents. This provides many benefits for organizations, however the key benefit is to facilitate enterprise application integration. The business identifier enables documents to be migrated to new repositories without requiring each reference to be updated.|
|Test the taxonomy and metadata model with real business data prior to starting development – An effective taxonomy requires rigorous testing prior to even starting development. This ensures the taxonomy and metadata model meets all strategic objectives, as well as system requirements.|
|Leverage mind mapping software to develop taxonomies – Mind mapping software is an excellent tool to facilitate the organization and classification of content, as well as data relationships.|
|Don’t define a granular taxonomy – Granular taxonomies are difficult to maintain and more difficult for content creation and consumption. A design that is too granular leads to uncertainty and inconsistencies when creating documents or trying to locate documents through browse or search. Unless the differences are too significant, it is better to segregate content through additional metadata fields than separate content types.|
|Don’t treat taxonomy design as a ‘one-time activity’ – Taxonomy and metadata model design should be viewed as an evolutionary process where attributes are added and removed as business requirements and priorities shift.|
|Don’t retain attributes if they are no longer needed or used – As taxonomy design changes, it is important that obsolete attributes be removed rather than hidden from view. This ensures data accuracy and integrity in data reporting and enterprise system integrations.|
|Don’t use free-text attributes unless absolutely required – It is best not to use free-text fields unless absolutely required; where possible common dictionaries should be leveraged for selection lists.|
|Don’t duplicate metadata across systems – Do not implement a solution that requires duplicate data to be entered into more than one system. If similar metadata is required in multiple systems, the population of this data should leverage EAI to propagate data to downstream processes and systems or display external system data when necessary.|
An effective document management strategy is dependent on an effective taxonomy design. Successful taxonomy designs promote data consistency, data integrity, and data interoperability across Enterprise Document Management Systems.