Content management… what is this thing? Probably the first place is to get a handle on some of the primitives involved. Clearly, there are two aspects to the problem: the social and the technical [although technic ultimately emerges from social process]. In defining these things, I perhaps need to establish primitives that span both. Based on my training, I'll start with the basics: documents and people.
Structured document-- What exactly is structured content? The most obvious example is a formatted XML form. I suppose that EDI communication would also be highly structured. In this regard a number of other documents could be considered to be highly structured. A purchase order or a claims form would be highly structured. There are two other considerations here: how do different types of people interact with the different types of documents, and what happens when we try to squish these documents across their boundaries.
Unstructured document-- So if purchase orders, etc. are structured documents, what are unstructured documents? I suppose that they are the "other". Everything else in unstructured: emails, memos, etc. There could by hybrids. For example, metadata applies structured principles to unstructured documents making them Franken-documents. The notion of "Franken-document" (i.e., part unstructured, part structured) needs to be addressed more thoroughly elsewhere.
Corp-- Now we move on to interpretive communities. The first level is the corporation... too big. I'm going to start with the individual and come back and then look at interaction with documents, etc… After puttering around for a little while with individuals and groups, I’m back to the corporation. It seems that pretty much everything has already been accounted for within the other two groups. The corporate demands are different. The same factors apply: production and consumption. Consumption happens in a broadcast manner i.e., everyone in the organization needs the same type of information. This information could be of either a public or a private nature. Private information may be tax forms. In this case we’re dealing with some sort of content management. Public information could be stuff like press releases or stock prices. This information could also be dealt with by a content management system. The production of information is likely related to fairly structured information about the overall functioning of the organization such as tax forms or SEC filings. This type of information may be analog or digital.
Group-- After thrashing around a bit with individuals I can move on to groups. So how are groups different from individuals? Well, let’s take the case of the insurance worker as a start. They process single things and—in most cases—assume all responsibility for that one thing (of course, individual documents may be promoted to management for review but this contingency is probably involved in the workflow). Other documents are the result of composite efforts by a number of individuals each of whom may have their own information needs. The group may produce either structured or unstructured documents. Whereas an individual worker likely produced largely structured documents (perhaps the level of structure determines the level of authorship?!?), group authors may produce something considerably less structured. Perhaps the role of group authorship is to actually create some sort ad-hoc and temporary structure for documents. Of course, there are genre limitations but I’m digressing… So what kinds of things are important to group authors? Well, workflow control and shared access to documents. Another important consideration may be content management. The same considerations apply for groups and individuals with respect to workflow management and the control of contingencies.
Individual--An individual is, well, an individual. Perhaps the important thing to remember about the individual is that they have their own tasks and do their own things: they are creators and users of information; they search and they write; they annotate and they delete (sometimes irrevocably). The important thing about individuals is that they have particular roles with respect to usage and production; for example, an insurance claims processor (as immortalized in Wenger's "Communities of Practice") has a constant stream of current live information from a client on the phone in addition to information from existing data sources such as client files and manuals. In this case, the processor gets information from a data source as part of their workflow. They may also get a type of "meta-information" about how to do their jobs. Both production and access is locked up tightly in their workflow. What becomes more interesting is how they deal with things that are outside of the typical workflow: how do they handle exceptions and how do they communicate lessons learned? What documents do they actually produce? Is it about workflow management? Is it about capturing exceptions?
From the context of the individual, let's move to some scenarios of how they may actually use technology. They would probably have some sort of work-flow tool i.e., a CRM system or some sort of SFA or claims-processing thing. From these workflow tools they have access to the relevant structured documents. They may also have some need for unstructured information. For example, they may have to conduct some sort of web search to get background information on a client or access to other correspondence. Of course, some unstructured information may be associated with the structured information through a type of cataloguing or semantic analysis. These documents may be Franken-documents or they may consumed by some sort of technephagos (i.e., instantiation of skill or practice that eats other things, kind of like an email attachment). The unstructured information could come from technologies like desktop search or perhaps enterprise search. Documents may also be available through more structured means. However, while people may use Franken-documents through a customized user interface, I can’t imagine that these users would have the level of bibliographic skill to fully tap into systems based on bibliographic surrogates.
So the individual may get information of a structured or unstructured kind as part of their workflow. But how do they get information related to “meta” functions or other aspects of their jobs? Perhaps standard information is available through some sort of Intranet or similar but there is an entire other class of information out there. From a management perspective, the information user is assumed to be a fully trained and competent employee. What’s missing is that sense of becoming i.e., what information do they need to become a fully functional employee? Similarly, what happens when the position changes? How do they evolve as fully functional employee? A managerial whim may conclude with a profound change in processes that may not be supported by the affordances of the existing technology. How do we take account of these changes in practice? Perhaps this is where knowledge management enters the perspective.
Lessons learned: 1- employees are part of a workflow. This workflow has two components: a- info access, and b- info production. For many workers—we could say Taylorized workers—these processes will be integrated into their tools. Problems occur when this workflow breaks-down. To use a Schlumberger expression described by Bowker, problems occur when parasites get in to the system. To communicate these problems is exceptionally difficult.
Some conclusions—There seem to be a variety of dyads for evaluating information management systems. Each of these factors is addressed to one extreme or the other by any of the systems:
- Workflow vs. off-workflow information
- Public vs. private information
- Information access vs. information production
- Structured vs. unstructured information
I think that these axes pretty much sum things up. Now, how to create a map that accurately describes these things! The question that emerges is: which tools best serve corporate needs? It seems that a lot of companies are positioning along the structured/unstructured and access/production but I suspect the workflow/non-workflow may be a more important driver.