Thursday, April 17, 2008

Document Behaviors

With my use of wikis and the adoption of wikis at The Firm, I have been focusing a lot of attention on the behaviors towards documents. After all, a wiki page is just another type of document. When producing documents, I have noted five types of behaviors: collaborative, accretive, iterative, competitive and adversarial.


With collaborative behavior, there are multiple authors each with free reign to add content and edit existing content in a document, and they do so.


With accretive behavior, authors add content, but rarely edit or update the existing content. Accretive behavior is seen more often in email than documents. Each response is added on top of the existing string of information with no one synthesizing the information in a coherent manner. I have seen this in wikis as well where people will add content but not edit others content.


With iterative behavior, existing content is copied to a new document. The document stands on its own as a separate instance of content. The accretive behavior is distinguished from the iterative behavior by the grouping of similar content together. With accretive behavior the content is being added to the same document, effectively editing the document. With iterative behavior, the person creates a new document rather than adding to an existing document.

With competitive document behavior, there is a single author who seeks comments and edits to the document as a way to improve the content. However, interim drafts and thoughts are kept from the commenters. The transmission of the content to a client or a more senior person inside the firm will result in a competitive behavior.

Adversarial behavior is where the authors are actually competing for changes to the content for their own benefit. Although there may be a common goal, the parties may be seeking different paths to that goal or even have different definitions of the goal.

Collaborative, accretive and iterative content production are largely internal behaviors. Competitive and adversarial are largely external document behaviors. Of course, a document may end up with any or all of these behaviors during its lifecycle.

I have an article coming out in KM Legal and Inside Knowledge magazine that further discusses these behaviors in more detail and in the larger context of wikis and document management systems.


  1. I'll be looking forward to the article, Doug. This is a great little taxonomy.

  2. This is really interesting, Doug. Are these different behaviours: that is, due to the personalities of the individuals involved? Or are they a reflection of the different purposes to which documents are put? There seems to be a bit of a mix in your taxonomy.

    For example, the most collaborative and sharing individual might sometimes want to create a document that nobody else should change. A senior partner may use such a document to communicate firm policy, perhaps.

    I think it may be useful to classify documents by reference to what they are intended for -- collaborative and consensual drafting, communication of difficult messages, and so on -- and then look at the authors' behaviours.

    I wonder also whether some of the behaviours are driven by the mechanism. A document that clearly has an owner (such as is the case in most DM systems) may cause people to be more deferential to the original author than one that lives in a wiki.

  3. Mark -

    I noticed the different behaviors as I introduced people to wikis. There was general reluctance to edit content that was added by someone else.

    Over the course of the last few months, I have been focusing on my own behavior and the behavior others in The Firm.

    Obviously, documents are subject to the various behaviors throughout their lifecycle.

    I reach a conclusion that people were just not used to the true document collaboration that you can get with a wiki. They seem more comfortable with the iterative behavior of email.

  4. Do you think many people have fully bought into the wiki idea? Without having had the opportunity to test it out, my feeling is that collaborative and accretive behaviours are broadly acceptable, but the other three are likely to have a more negative impact on acceptance of wikis as a tool.

    I notice from the screenshots in the post following this one that Sharepoint adds the name of the most recent author at the bottom of the page. I wonder if that strengthens the sense of ownership that might give rise to iterative, competitive and adversarial behaviours.

  5. Mark -

    We will see what happens. I expect that wikis will start largely as a lightweight content management system and that most of the wiki content will be entered by my KM team.

    We also plan to have our summer associate class "wiki-fy" lots of our high-level content.

    My KM team has adopted collaborative behavior and will jump right in and edit content.

    I just see the true collaborative behavior of wiki to be a very different way of working.

    As our wikis spread, I will certainly put up future posts.

  6. This is very interesting analytics with respect to document behavior. Indicative of how group behavior changes the way content is developed. Probably not very visible when there are only one or two content authors but scale content development beyond that and the dynamic changes; probably along roughly similar vectors as described here.

  7. What a fascinating taxonomy. And since the types were presented in seeming order along an apparent axis with end-points of "highly collaborative" and "highly combative," it reminds me of Maslow-esque needs hierarchies, personality typing, communication styles, etc.

    It would be interesting to see if (and how) wiki participants from different industries would cluster at different points along this axis. Would lawyers group around competitive and adversarial?

    I'd hypothesize that initial modeling of specific styles, with social rewards for behaving in collaborative modes and social punishments for behaving in combative modes, would highly influence subsequent behavior.

    I'd also hypothesize that, without any such modeling and rewards, wiki participants' behaviors would highly correlate with personality attributes.

    Ann Lee Gibson


Note: Only a member of this blog may post a comment.