Some Thoughts on Effective Enterprise Taxonomies for Content

While discussions of content management are usually referred to in the context of the needs of enterprises (i.e. Enterprise Content Management, ECM) in reality most ECM deployments start at the departmental level. This is not bad – on the contrary departmental deployments typically address specific business needs thereby sharpening their focus to improve their chances of success. However, over time as an enterprise begins to deploy ECM technologies in many departments, the benefits of an enterprise strategy to support cross-enterprise deployment become apparent.

At Open Text we have recognized that there are a number of essential elements without which transitions from departmental to enterprise-wide deployments tend to fall into a 'chasm.' It may seem obvious, but it essential to treat enterprises differently than individual departments, especially since:

• 
  
  Enterprise deployments inherently connect departmental deployments rather than replace them
  
  • If your aim is to facilitate the timely, process-dependent flow of content between departments then you have to accommodate varied work cultures, processes, content types, taxonomies and technologies already in place in order to make the necessary connections as part of a clear enterprise-level content architecture. Wholesale replacement is seldom a realistic option, but supplementing and amending can be effective.
    
  
  
  
• Enterprise deployments can also create additional value by enabling new, enterprise-spanning, content-centric processes and perspectives that cannot be achieved solely by connecting departmental deployments
  

One element to successfully implementing enterprise-level content management is an effective enterprise taxonomy.

• Enterprise or corporate taxonomies (see Wikipedia definition) are becoming increasingly critical to classify, save, access, reuse and report on growing volumes of information within an enterprise
  
• Enterprise taxonomies are particularly important for Content Management and Knowledge Management, but the needs of each of these disciplines are not exactly the same
  

While the principals of taxonomy development are relatively well understood, there are particular challenges associated with effective enterprise taxonomies. In my discussions with colleagues several recommendations have emerged:

1. Externally derived and validated taxonomies should be adopted whenever possible – this can be especially critical when you are faced with multiple, overlapping taxonomies that were adopted and typically custom built to suit the needs of specific departments
   
2. If you adopt an external taxonomy (#1 above) it is most likely that you should only implement a simplified subset – aim for 'just enough' detail
   
3. Natural language taxonomies are preferable to numeric or other code-type taxonomies to support user adoption
   
4. The appropriate taxonomy type or structure should be chosen before implementation is started, i.e. flat, hierarchical, network or faceted
   
5. Each taxonomy you deploy at the enterprise level should be orthogonal to any other
   

Typically enterprise taxonomy discussions begin when staff become frustrated because they cannot access the information, and especially content, they need from one location. In such cases an enterprise taxonomy may be seen as an access enabler (e.g. through a portal) where the focus is on information presentation.

Another driver may be the need to manage key records to meet regulatory demands. In this case an enterprise taxonomy is expected to support storage of all relevant content and effective retrieval on demand by a small set of users. Such taxonomies often don't serve the general needs of all staff.

A third scenario is becoming more common: ensuring that appropriate content items are presented in the context of business processes that already rely on structured data (e.g. in ERP or CRM systems). In effect metadata applied to content can be the effective bridge between structured data and unstructured content.

The recent application of social networking technologies within enterprises has brought with it demands to classify unprecedented volumes of small content items where much of the meaning depends on the context of discussions and may not be captured within the content itself. In this case we need to consider deriving social metadata.

These are some initial thoughts on the topic of enterprise content taxonomies. It is a rich topic that I expect to explore in greater detail in future posts, especially looking at reconciling nor only classic, mandated taxonomies, but also the more recent emerged user-managed cloud taxonomies.
Syndicated at http://conversations.opentext.com/