Complexity and Knowledge Management Navigators…
Over the next four weeks I am going to speaking at KMAsia, K-Net and the European Training Foundation KM Seminar. I thought that I would share with you the core of my discussion; all three talks will revolve around the idea of the ‘future of KM’.
My argument is that organisations exist as model dependent entities. Everything the individual and the collective does within an organisation is either informed or influenced by a model; from the mental models the individual constructs to make a decision to the organigraph, a simple model that illustrates the structure of the organisation. Managers and executive teams love models, from PEST/PESTLE/PESTLE(C) to SWOT, to Porter’s Five Forces, to Force-field Analysis, to Cultural Webs, to Lean, to Action Learning.
Take a look at the foundation of any knowledge-intensive organisation; the operational aspects of the business are built upon a simple model of input (transformed resources) applied against transforming resources to produce an output that then links back to the input process through a feedback loop – simplicity itself! Consider the macro implications for KM…Information is the input, people then transform that information – with knowledge as the output, expressed through services or products – which is then assessed against the boundary requirements of the organisation.
Models, models, models! We even persist with models when they are clearly not functioning correctly! Take KM and Nonaka’s SECI model, it clearly is not an operational model for Knowledge Management. It is an interesting conceptualisation of how KM could work, but it certainly is not an operational model to be applied in an organisation; for a start, the model is incomplete – it might tell us the ‘what’ of certain preconditions, but it most certainly doesn’t show us ‘how’ those preconditions work together to produce an outcome. Regardless, many organisations persist with the model, reciting its mantra of Socialisation. Externalisation, Combination and Internalisation, patching over its shortcoming with their own ‘work-around’ – even Nonaka had to develop a work-around to SECI, incorporating ‘Ba’ – all the while believing in the core tenet that Tacit can be converted into Explicit knowledge, which I would strongly argue that it can’t. A good model is said to be defined through its ability to pass tests, such as ‘Comprehensiveness’, ‘Correctness’, ‘Usefulness’, ‘Clarity’ and ‘Conciseness’. I would suggest that SECI fails badly when assessed against these criteria; for those interested my argument for this is set out in other blogs, such as ‘Nonaka, the wonderful wizard of KM‘.
Now, if we can agree that organisations, and the individuals that make up the collective, are model informed/dependent, then we have to accept that KM will only ever be as good as the models that inform the process. This leads us to the hundreds of models that are available to the organisation. Which is the correct one to choose? The wrong choice can lead to dissatisfaction and, while many will argue that it is dependent on the situated need, my argument is that an organisation needs a general model to work from, a high-level systems overview, through which they can develop evidence-based operational models as a response to that need. To do this there is a need to understand and map the general preconditions for knowledge driven processes. Here’s the scary thing; for four years we have coded and mapped the functions and dimensional constructs that inform KM processes in organisations – some of the findings have been made available in this blog. We have then compared over 100 models against that coding process and none of them have met the preconditions of the four functions and twelve dimensional constructs that we have identified. In fact, on average, models are generally only 62.5% complete; this is a reflection of only one aspect of what constitutes a ‘good’ model and, if we can accept this, then we also have to accept that these models will provide a distorted output that could serve to disrupt the environment they are constructed to serve, contributing to failure and breeding dissatisfaction. People can, and have, rightly raised the argument that their models are ‘proven’ to succeed, but they do not profess them to be general models and, as I have already argued, they are often ‘adapted’ to overcome deficiencies in their design according to the situated need.
This is not to say that we are right and everyone else is wrong. This is us saying that actually there are too many incomplete models out there and, if we can agree that organisations are model dependent, then we also have to agree that a lack of ‘completeness’ when it comes to preconditions for KM systems will result in system failure. Ultimately, we need completeness when dealing with KM. Without digressing too far into the theoretical aspects of our field, the KM system is complex by nature, which, accordingly, requires an approach to optimise the whole instead of a reductionist approach that focuses on the individual variables.
What I am suggesting, and something we are working on with Perigean Technologies in the US, is the need for the M-Model, a meta-model if you will, for KM. A model that sews together overlapping theories, concepts and models to develop a single M-Model for KM system design. We accept the argument that situated need of any given organisation will distort the intensity in the preconditions and the programmed response. However, we also contend that the preconditions remain the same, what we are ultimately speaking of are general preconditions that are adapted at an operational level; think of it in terms of granularity and the pixels that inform the picture as a whole – the picture is clearer at the higher level of magnification and becomes more distorted at the lower level of magnification, regardless, the pixels combine to make the whole.
This blog, by nature, will never be able to cover off all the arguments that inform this debate. Also, I accept that I am opening a can of worms here by suggesting that ‘incomplete’ models are causing system failure and probably dissatisfaction. I accept that people will put forward evidence to support their personal models. I am only positing that our evidence base suggests the a lack of completeness to be a core issue for the model-dependent world that exists within public, private and third sector organisations.
Well, that’s the core of my Keynote addresses…with a lot of evidence sprinkled in for effect…hopefully it is seen as engaging and stimulates though provoking responses…