Our Blogs  


What is thought to be is rarely real

By Dave Snowden  ·  November 5, 2014  ·  Reflections

As I described yesterday, my use of narrative started with the need to discover decisions and their nature as part of creating a bottom up approach to knowledge management strategy.  Decision mapping has high utility in its own right as, done properly, it reveals the true nature of activity in an organisation.  That can then be contrasted with the formal understanding of the organisation, normally represented in a process map with supporting procedures.   In all my years of doing this work I have yet to see the two aligned.  Contrasting what is with what is thought to be with a representation of reality is a key complexity technique.  Presented that contrast without evaluation an essential aspect of changing attitudes by enabling descriptive self-awareness.  By that I mean allowing decision makers to come to conclusions for themselves by presenting evidence of dissonance; the role of the consultant being to enable this rather than to determine the answer.

I should say that this type if mapping is not just of value in knowledge management.  It's a key element in complexity based strategy, scaling Agile development and lots of other things!   Decisions are at the heart of human activity and mapping gives you the basis for lots of other understanding and intervention design.

So how does it work?  Well the process is fairly simple and can be done manually, or more effectively with a simple SenseMaker® deployment.  I'll describe it with SenseMaker® but also indicate the manual alternative stage by stage in italics where it is different.  For those interested we will have a preconfigured version of SenseMaker® ready for this later in the month.  It's one of a series of standard applications of SenseMaker® that use predefined signifier sets that we are now launching as a lower price entry point requiring less training and understanding. So lets run through the stages:

  1. All staff or a representative sample of staff are asked to download SenseMaker® to their smartphones and are asked to make a note of every decision they make no matter how trivial. This can be done by typing, recording or taking a photograph or using a drawing/picture or the like.
    Without SenseMaker® this is done through workshops and interviews with a sample of staff focusing them on what decisions they make daily, weekly, monthly, annually or some other event based stimulus.
  2. For each decision they are are encouraged to comment on the decision but in particular are asked to identify information sources used in making the decision, the manner of its communication and the resources or more precisely artefacts used to support the decision. This can use predetermined categories relevant to the organisation or can just be left open.   Often its worth capturing for a week or so, then creating a drop down set of options as the journaling continues.
    Capture is more limited manually, but does have the advantage of letting the interview start to construct an emergent taxonomy of information and resource use which will reduce the time in analysis.
  3. At the same time they have the opportunity to identify how the decision could better be communicated, to who and with what improved information sources and resources.
  4. If using SenseMaker® each decision is mapped onto triads and stones (if you don't get that look some of the project examples or watch some of my YouTube videos - the ones from State of the Net in Trieste all contain the basics).  That allows us to look at underlying decision types and then cluster them for evaluation.
    Without SenseMaker® the process of clustering is workshop based and manual in nature; having a lot of junior analysts helps here.
  5. Now, in a workshop with some preparation we present the decision clusters with summarised information flows and start to link and connect decisions.  Communication from one will be information in to another and so on.  This will also show up gaps that require investigation.   At the end of this you end up with a wall of hexagons with lots of links between them and it is always plain bloody messy!
  6. The map is transcribed into concept mapping software which optimises the representation and you end up with something that looks like an extended network model.   It's messy but it is coherent and it bears little relationship to the formal process map.
  7. Of course it would be better to start process mapping with a decision map and then modify it over time.  But in most cases that is too late so you are forced to make senior executives very uncomfortable when they see just how little the neat and tidy constructs created by those pervasive consultants a few years back match the reality of day to day work.   The gaps between then become specific change projects that will feed into the wider picture.

Ideally you would run this over a few months in background and then leave it in place for subsequent monitoring.  However I have done it in two days in a large multi-participant workshop for the Treasury in the UK and Education policy in Australia.   It's an open source method but I'd recommend mentoring the first time you run it.  You can end up with tens of thousands of decision clusters and then you need to create fractal or scaled representations.  Concept mapping software is good for that as you can compress and expand.

Of course you can also ask questions of the decision clusters - more on that tomorrow - and that also feeds into the dependency matrix that is at the heart of this approach to knowledge strategy.