Sunday, January 27, 2008

Everybody wants an enterprise-wide information model - so where is it?



Every project that aims to improve how information is used and managed within an enterprise (be it related to ECM, BI, SOA, or whatever) needs to do the following:

  1. Ensure that key concepts have the same meaning and are represented by the same terms everywhere they exist and are to be used. This is to avoid basic misunderstandings that might put the project in the wrong direction.
  2. Ensure that the information resources that are to be used and/or managed are identified, defined and values (so that key assets can be identified), that their relationships to each other are known, and that every instance of these information resources has the same - or at least a similar - structure and attributes. This is to be able to focus the improvements where they give something back (ROI) and to simplify integration and exchange of information resources between people, activities and applications.

After having realized that there are no updated conceptual and logical information models readily available, most projects start to develop their own. Thet start developing their models from their own perspective, covering only their areas of interests. And if they succeed in developing such models and making them usable for the purpose of their specific projects, they soon falls into after the projects have ended.

Now, imagine if the project would have access to an information model describing all information and content resources - structured as well as unstructured - that exist within the enterprise. First of all, it would mean that each project would not need to spend energy, time and resources on developing their own model. Each project would be able to make a jump-start and focus on identifying areas improvements and developing efficient solutions instead of being occupied with trying to describe how the world looks like. And even if they would still look at the world from their own specific view points, they would be starting from the same foundation as every other initiative. They would speak the same language and see the world in a similar way as other initiatives taking place elsewhere in the enterprise, and it would be less likely that they would misunderstand each other.

An enterprise-wide information model is the dream of most Information Architects. The challenge is to develop an information model describing how all information resources within the entire enterprise are related instead of only doing it for a specific domain or application. The information model could then be made available to all initiatives within the enterprise, as a shared service.

Since developing and maintaining an enterprise-wide information model must be a never-ending task and requires a holistic approach and lifecycle perspective on all information and content resources, it cannot be accomplished by a project. The only reasonable approach would be to set up a team that is dedicated to define and maintain representations of the enterprise information architecture. This team would also be assigned the responsibility to continuously refine and assure the quality of the information and content resources and the relationships they have to each other (the architecture). The team must consist of skilled information analysts, information architects and information stewards. However, to be successful the team must be operating throughout the enterprise, helping out different initiatives. Otherwise it will end up doing like most enterprise-wide initiatives – developing something that is neither usable nor based on an understanding of how things really are in the different corners of enterprise.