Thursday, July 9, 2009

Change Models: A Systems Perspective

Like other people who collect stamps, I collect change models. I find it interesting and fascinating how much energy and passion people put in the development of a conclusive generalization in what happens during a process of complex organizational or political change. And most of the models make a lot of sense - in certain contexts, organizational cultures or applied to a specific issue that the organization needs to address. However, I haven't found a model yet (and I believe I will never find it) that is applicable to all different situations on the ground. Or, maybe better: although many models might be applied to all kinds of contexts, the result of this application in terms of an improved organization or political process is not always satisfactory to
either the client or the people who are impacted by the change process. I can't believe that some consultants still have only "their" model in their briefcases.

Let us first have a closer look at why no change model can be a panacea and then have a look at different change models and evaluate them in terms of popularity and applicability. There is no finally proven scientific concept of change - as their is no universally valid sociological model that explains human groups or societies - because change happens in complex social systems. The behavior of highly complex social systems can be described and modelled but not exactly predicted and it looks like that it will never be possible to do so. All change models that do not take into account the enormous complexity and final unpredictability of human systems might help to understand certain aspects of how a system reacts to a change driver but it will evidently fall short to produce repeatable explanations that are valid and accepted for all different stakeholders. Present the same model to 10 different people in an organization and you will hear 10 different interpretations of it. You might say now - if a change model fails which does not take into account complexity theory, can't we then find a universally valid change model that respects the laws of complexity theory?

Well, there are indeed a variety of complexity theory based change models. Problem is that complexity theory is a very young discipline which developed over the last 30 years, originally based on the mathematical description of theoretical or artifiically created systems. Since then, complexity science (vulgo: Chaos Theory) has been expanded but those theorems that have been proven experimentally are not that specific enough to explain every aspect of a a system's behavior. (John Smart says: "Most prediction is a predictable failure."). Let us look at what seems to be the scientifically agreed fundament of systems theory. These theorems are not taken from a text book but I wrote them up. So, as I am not a complexity scientist, please tear me apart if you can:

1. Complex systems consist of different, interconnected and interdependent parts.
2. Complex systems show non-linear behavior.
3. Complex systems are not stable; if they look stable, they are in a state of equilibrium.
4. Emergent phenomena are a common feature of complex systems not because they are magic but because we lack the computational power to exactly predict the behavior of the system.
5. Social systems are complex per definitionem.
6. You need a specific model for each kind of system to compute/predict its behavior. The more complex the system the more we depend on empiric findings to establish a specific model. Any prediction can only be an approximation.

Because of theorem No. 6, no model for organizational or political change can be universally applicable to more than a limited set of change processes. Practice will show its validity.

In the next weeks, I will publish a series of blog posts that describe different change models, their background, theoretical foundation, applicability, field validation.


  1. Looking forward to the series Holger. I'm also interested in these models, and find it funny how some are so proud and protective of their models when they are so similar to others, and as you've mentioned; not universally applicable.

    While none are the perfect fit for every scenario, I think to have a model as a frameworks to build a plan around is a positive. It just remains important that each phase of the model is questioned.

  2. Holger,

    Change, by definition, is not a static thing. My background is change management in IT and over the past 30 years I have seen the evolution of change management as a discipline that attempts to bring some order to the management of change. We do this to create traceability and accountability, so that changes can be undone and processes refined to avert future incidents.

    The big lesson I have learned is that you cannot assure repeatability of the full change process. You can create some macro level checkpoints and track what happens, but as you note, the human factor has to judge how individual changes are handled at a micro level. That way you can assure auditability and accountability, but predictability is still wishful thinking in most cases.

    I too am looking forward to your future articles.