Friday, October 5, 2012


In my last post Don't Panic i touched upon complexity, a topic which seems to be pretty hot at the moment by looking at various twitter messages and fairly recent blog posts. Richard Veryard has touched the topic in a quite interesting way in his post On The Causes of Business Complexity. What I particular liked in his post was his four causes of complexity:
  • Emergent Complexity - consequence of many small and unrelated decisions and actions.
  • Perverse Complexity - consequence of clumsy attempts to reduce complexity.
  • Contrived Complexity - consequence of deliberately creation to benefit some stakeholders.
  • Irreducible Complexity - consequence of real complexity of the demand environment.
 I like to reorder that slightly without really changing the context, because working in complexity is epic.
  • Emergent Complexity
  • Perverse Complexity
  • Irreducible Complexity
  • Contrived Complexity
To tackle the EPIC Complexity I can now use the SCAN Framework. and by that create sense, aim for decisions and look for the right skill and mindset. A topic which is also tackled by Shawn Callahan in his post When Should We Collaborate, where he also refers to Cynefin and links a way of working to complexity:
  • coordination for Simple Problems
  • cooperation for Complicated Problems
  • collaboration for Complex Problems
  • Any method for Chaos to shift to one of the other three complexity domains
I agree with his judgement of the first three levels of complexity but I am more willing to follow Tom Graves approach in the fourth domain. In his post Sensemaking - Modes And Disciplines he sees working and acting as an Artist (inner value) as the answer to Chaos. In this one I full agree, because I believe the other three working models (and roles in Toms post) will only find unknown areas of their own domain in the unexplored space of Chaos.

So I call that now the EPIC SCAN.

So how do I link this knowledge to GLUE. First of all I look into optimizing or fixing the flows in the GLUE Space. Here applying the right approach to way of working and finding the right skill and mindset is in most cases more important than finding the perfect answer. The people will find the right answer inevitable if their skills and mindset fit to the complexity domain. If the fit is not given then they will try to shift the complexity into another domain or start being frustrated. A typical statement here is: "If X would just understand me".

And unfortunately there is no Silver Bullet to it. Not even Ken Schwaber markets the methodology SCRUM that way. In a recent post he clearly links SCRUM to the complex domain (where the unknown is greater than the known). In spaces where the known is greater than the unknown SCRUM creates more waste than needed and other methods (if applied correctly) will produce results in a less wastefull way. The challenge is to find the right method to maximize Value Add. And for those who are still unsure: the complexity area where I love the Domain Enterprise Architecture most is in the Chaos, no matter where in GLUE.


  1. Kai

    I'm not sure what you mean by "tackling" complexity. In the case of what I call Perverse complexity, the correct action may be a negative one - avoiding or unpicking counterproductive action. Contrived Complexity calls for various forms of competitive or regulatory action, for example either exploiting or challenging the contrivance of complexity. And Irreducible Complexity calls for acceptance. Thus in most cases it is not the complexity itself that we need to tackle, but our own limited response to this complexity.

    1. I am not specific with my answers here for good reasons, but I believe that the answer is either simple, complicated, ambiguous or not known. And in GLUE people is always the centre of the action, so in most cases (80/20) it is not a technical answer even though it might be given in the context of technical discussion. And in my GLUE thinking we (or our own limited response to the complexity) is part of the system or at least the greater context.

  2. I've tried to explain this to Tom several times without success. But to make things clear - if you check the Cynefin articles and material (ones with myself as author) you will find that the main system types are:

    Order - constraints make agent behaviour predictable
    Complex - constraints and agents co-evolve
    Chaos - agents unconstrained

    Order is then split into Simple and Complicated and the central disordered space added/

    Tom's description of Chaos are to my mind what the literature describes as Complex; I suspect he is using a common language definition rather than the scientific one which is fine but which I also think explains his confusion.

    The general principle of Cynefin is that Chaos is a good space to enter deliberately for purposes of innovation or exercise of wisdom of crowds (where the agents have to be unconstrained by interaction). Entered accidentally then one should move into another domain, ideally the complex one.

    Otherwise your taxonomy of complexity is interesting - I will pick that up and blog when I have time.

    1. Hi,
      thank you for the compliment and please share that I can read your blog and try to understand more. From my point of view (and that is more the practical than the scientific approach) there is no real difference in GLUE between Chaos and Disorder, because People matter. So something which could be seen as perfectly simple can easily turn out to be not-known in my GLUE SCAN, because the people at hand to solve the problem (complexity) are not able to find a relationship in the simple, complicated or complex space. Disorder of course can be tackled with any of fit-to-purpose methods for simple, complicated and complex, because then some of the disorder turns out to be ordered.

      So yes, I confess, I use common language and not scientific language, but for whatever reason that inaccurate common language adds better to my toolchain to deliver to the demand. (And I also use emotions to get my message delivered).

      With respect to a more accurate use of Cynefin I like to work in Chaos and disorder. Both spaces are of big joy for me, but for simplicity and appliance I fold them into one.

    2. To correct Dave's usual misapprehension above:

      a) My usage of 'Chaotic' is the same sense as in chaos-mathematics or the science of uniqueness and unique-events: in that sense it is every bit as 'scientific' as Dave's.

      b) I do not have 'confusion' - I am very clear in my thinking on these points. My practical problem is that I have to do a lot of work with my enterprise-architecture colleagues to sort out the mess caused by Dave's technically-correct yet highly context-specific usage of terms, relative to the standards used in this industry (which align more to the colloquial usage and/or the chaos-mathematics usage).

      I would _really_ appreciate it if Snowden backed off from his much-repeated innuendo and insult of implying that my work is 'unscientific' and 'confused', because it is neither. Enough said, really.

    3. First of all: I am not educated enough (or in the position or willing) to judge who of you two is right with respect to scientific or not scientific. I apply common language in my daily work, because it works for me. If that is aligned with scientific language and thinking even better, but nothing i spend a second of thinking about.

      What I know is, that I have read Cynefin with interest, but was not able to really get it to work for me and therefore I kind of invented my own approach to complexity handling. SCAN helped me to formalize my thinking, so I can only second the statement in b with respect to my own work. I am not sure if there is really a mess created, confusion, not working or just not understanding, but again that does not really matter.

      For me there is clear value in both of your approaches and I don't want to spend much time on trying to sort out some chaos which seems to be more of a tribal (or religion) war including all the emotional aspects of throwing nasty words at each other (or the desire to hear the nasty words between the lines). So please, if possible, carry out your private fight somewhere else. I still appreciate any words of wisdom on my thinking, because that helps me to dig deeper into my thoughts.