7 Replies Latest reply on Dec 10, 2008 10:47 AM by michaeljpastor

    How to Hit the Ontological Sweet Spot (Rigidity vs. Chaos)?

      I hate to go all dualist/philosophical on you in my first post, but here goes! In a utopian world, I think we want self-organizing systems in which tags always have accurate conceptual meanings and establish meaningful relationships with each other.


      In the real world, it's not so simple.  How do we deal with the following?:
      - homonyms:  Words that share the same pronunciation but have different meanings, and can be spelled the same or differently.
      - synonyms & plural/singular tags:  Different words that mean similar things (may cause dilution of tag emphasis)
      - idiosyncratic tagging


      There are established methods to make tagging more accurate and I think that's great.


      But how do we organize meaningful relationships amongst tags?  I think the philosophical answer may be to utilize organized chaos.  You could impose a hierarchical/categorical structure, but that might be too rigid.  But what if the "categories" were actually groups created by people?  Then tag clouds could be more focused and meaningful.


      Perhaps a user would search for a keyword or keywords which would result in a list of the most relevant groups that possessed information with that tag or tags.  The user could then join the group and further collaborate on that topic.


      But then there's the problem of siloization.  What if the crowd forms ten groups on the same general topic?  Whereas the instantiation of one group on this topic would cause greater synergy across the enterprise, ten groups cause the effects of siloization we see today.


      Surely there must be some way to balance the potential ills caused by both central control/categorization (rigidity) and chaos (siloization)?  And I understand that groups may collaborate on a topic at different levels (whether due to varying geography, policy or what have you) but you still want to encourage synergy at the highest level possible. Perhaps some combination of planning up front and adaptation as time goes on could help.


      Can anyone cite research or past business cases?  I'd be interested in learning about how to best solve this problem across an enterprise.  Especially one in which groups may be secured.

        • Re: How to Hit the Ontological Sweet Spot (Rigidity vs. Chaos)?

          Perhaps some people will think this is just an academic question, but with about 163,000 internal employees and many specialties it makes things a bit complex.

          • Re: How to Hit the Ontological Sweet Spot (Rigidity vs. Chaos)?

            Hi David,


            I definitely ran into this challenge when I set up my first internal community in these two areas:

            • Categorizing content - tag cloud or hierarchical set of categories
            • Employees skills - minimal questions with user tagging or a detailed skills inventory


            Because I was using the open source software Drupal for blogs, forums, and events and Confluence's wiki, I was restricted somewhat by what the tools provided out of the box. Drupal 4.7 let me do free tagging or create hierarchies of tagging content and Confluence 2.2 allows you to do free tagging with no relationship between tagging.


            I guess I may be a control freak that likes information  more logically structured so I opted to create tags hierarchically for my community of blogs and forums but that is definitely not the trend of today's social media tools.  I also provide a long list of user profile fields to more accurately target a user's level of expertise in lots of technical areas.


            I do find that our users are more likely to tag their content using our set hierarchy of categories for blogs and posts than they use free tagging for the wiki pages.


            I think there is a way to find a happy mean, but just haven't seen it yet.  If I come across any innovative approaches, I will be happy to share them.


            Personnally I rarely use tag clouds to browse content, unless that is the only form of structure that the blog or community site provides.


            But it really comes down to usability  -- both from posting the data and browsing it once it's published.  Are there any usability studies on the subject? Anyone?


            Cheers, Angelique

            • Re: How to Hit the Ontological Sweet Spot (Rigidity vs. Chaos)?

              Thanks Angelique.  In our business, we have:


              • Divisions (Matrix structure which contain Programs (Products))
              • Functions (Engineering, IT, Finance, Supplier Management, etc...)


              I think a good solution may involve orchestration from each of these entities at the various levels within each.  At each level, they could determine which groups are important to them, create, populate and link to recommended groups from their own organizational groups.  The pertinent organization could even create new groups where holes exist and communicate w/ the appropriate communities to fill them.  If they notice groups that have a very similar focus, they could encourage the groups to combine to inhibit siloization. (I think we would want a tool that made combining similarly-focused groups an easy process.)


              People could still create new groups and the above orchestration process could continue over time in a loop.  This way you still have flexibility to explore new common interests and maintain top->bottom and bottom->up feedback mechanisms.


              I hope I explained this clearly.  Comments?



              • Re: How to Hit the Ontological Sweet Spot (Rigidity vs. Chaos)?

                As usual, I've got a multi-part response.



                A typical user's propensity to use tags, and what it means to them


                First, if the majority of your organization will actually tag content, congratulations. In my anecdotal experience, the majority of people out there don't understand the value of tagging content so that OTHERS can find it more easily. Some of them do, however, understand the value of tagging content so that THEY can find it more easily.


                To me, if I'm going to tag something, I want to find it again in "My Bookmarks", a la delicious or Lotus Connections Dogear or ConnectBeam. But, for tagging to provide value to others, those tags need to live with the content, and not just in my bookmark. Until both of those needs are met, then the use of tagging will stay limp and lifeless among the majority of users.


                For those who do tag, I've seen people pick tags from a pre-fabbed list more often than they will type in their own, as Angelique describes. So, some tag seeding might be helpful, and maybe collect those tags into tag groups to make it easier for people to select them.


                Natural group formation and how they continue to "live"


                In my experience, it's perfectly fine to let people form multiple groups around the same topic. In fact, you should encourage this, because that's how pockets of people discover other pockets. The community manager should do monthly or quarterly reviews of all the groups formed, then approach group leaders when their groups might benefit from merging, or at least becoming sub-groups underneath a larger topic umbrella.


                Like Andrew McAfee says, let structure emerge


                The best txonomies emerge over time. Let the population converse and commune in general "piazzas" and periodically, take a look at what they're talking about. If there's enough volume - you get to decide what "enough" is - form a new community or group and move the content into it.

                • Re: How to Hit the Ontological Sweet Spot (Rigidity vs. Chaos)?

                  Sounds like good advice.  Thanks very much!

                  • Re: How to Hit the Ontological Sweet Spot (Rigidity vs. Chaos)?

                    We crossed this bridge a year ago, and (right or wrong) I dictated a "let chaos rule".  No predefined taxonomies, ontologies, tag clouds, whatever.


                    Why?  We were spending way too much time on the topic, and I didn't see it generating corresponding business value.


                    OK, it's been a year -- what have we learned?  For us, anyway, it was a good decision.


                    The only downside is the inevitable fact that we have people throughout the organization who can't tolerate an untidy environment.  We suggest these same people stay off the internet as well, as it too is very untidy in a similar fashion :-)


                    Seriously, though, not a month or so goes by when I have a conversation with someone who starts from the point "how can anyone get anything done in this sort of chaos?" and we once again pull out all of our examples that clearly show that -- yes -- people can get work done without having everything neat and tidy and pre-ordained.  We gently suggest the problem might be with them, and not the platform.


                    Going a bit farther -- our internal platform is focused around *conversations*, and less so content.  If we had intended in building an uber-content-repository, maybe we would have decided something different.  BTW, we do have an uber-content-repository at EMC, it has a decent taxonomy/ontology, and people continually grumble about taxonomies anyway.  For me, anyway, it looks completely unwinnable.


                    If you consider conversations, there's all sorts of ambiguity in how people use words, terms and concepts.  And, correspondingly, there's a lot of "what did you mean when you said XYZ?" just like in the real world.


                    It'd be interesting to see how others view this issue.


                    I'm glad we went the way we did -- far less effort invested, far better results.

                    • Re: How to Hit the Ontological Sweet Spot (Rigidity vs. Chaos)?

                      "We gently suggest the problem might be with them, and not the platform."


                      That seems overly pedantic to me.  Some people prefer structure and some don't, and both have their value.  Have you followed up with those who preferred the structured environment to see if they still use it, feel appreciated (both online and in your company), or compromised/capitulated and embraced the chaos?


                      And why is it an either/or proposition? Why can't it be both/a matter of degree?