Linking through Search

A colleague of mine made and interesting statement about the sprawl which results from SharePoint site proliferation. He said that sites should be allowed to become whatever is right for the business while being tied together by search. This got me to thinking about the concept and the types of things which would be needed from a corporate perspective with this philosophy.

Corporate Document Control

In environments where there are audit and security requirements -- like ours -- you would need some policies and procedures to control and track controlled information (I'll refer to them here as documents just to keep the blog shorter). Access to this information would have to be controlled and, often tracked, for compliance reporting.

Classification

In order to find something, it would have to be tagged in some way with a consistent set of terms so that I could find it. This taxonomy, in my experience, it the place where many organizations fail. It takes a good deal of continuing maintenance to assure that it is consistently enforced, updated, and applied.

Another aspect of classification is knowing what is actually real at the time of the search. Let's say you're looking for the "jury duty" policy (something I'm currently involved in). You enter the term into your handy search bar and get back several documents. You really need to be able to know which is the one governing the corporation at the moment vs. the one which was presented to the committee as a white paper, but rejected.

Your search needs to give you some locational elements to know where this document is in order to know its relevance to your current need. This can be difficult for a search engine. Take a look at this site as an example. It contains 25,000+ documents. Sure, you can find a document based on your term -- often hundreds of them. But, what is the reference you are searching for? Where are these documents located so you can get some context for them? We implemented the concept of Abstract files to allow you to find that information. It's maintained by our publishing system and allows the user to go to the page(s) on which the document is located to understand more of the context in which it was produced.

Document Life Cycles

People tend to be pretty good at putting things on sites. They are less than good about getting old and irrelevant things OFF the site. Thus, you can end up with a mass of old, outdated information on your site. Using search an reveal this information and, being presented out of context, it can lead to confusion, misunderstanding and frustration. It can be worse to get too much information than too little. This isn't really a problem for your search, but for your governance policies, procedures and life-cycle automation to help keep the sites clean. Just finding old and irrelevant information can be a major undertaking without some policies and tagging procedures and the involvement of the owners of the sites.

Administration and Maintenance -- the Human Factor

One problem of proliferation of sites is the human factor. The scenario is that some super user on your team establishes a site, maintains it, makes it elaborate and hard to maintain and then gets transferred to another department. Their replacement (if there is one and this wasn't a downsizing), is not so interested in the site and things languish. This can occur because job functions change, personal inclinations, or just the press of other duties (their real job). The site becomes outdated and irrelevant, even blatantly wrong and there is no one to clean it up, so it's documents contaminate search results.

With inevitable reorganizations comes the interesting phenomenon of information which is completely unowned or ends up being owned by multiple organizations. Now, when the system administrators ping a group about their site which isn't used, that group doesn't even exist and the new groups won't take responsibility for the information on the site. It might still be valuable corporate information, but there is no one to actually maintain it -- an interesting situation for the system administrators and compliance staff.

Conclusion

I think my point is that the concept is good -- sites should be useful for the organizations which use them. In fact, they HAVE to be or they won't be used. But, there has to be come corporate framework and human administration and maintenance to keep them relevant and useful at a corporate level also.