The diagram above was used by John Poindexter when he introduced the technology session last week at the IRAHS Symposium. This event brought together a really interesting and diverse range of people from all over the world. John used the diagram to provide a general context for the presentations and discussions that followed. I scribbled it down, produced the slide during one less than interesting presentation, then got permission over lunch to use it for this blog. Now my first reaction negative; I have a long standing hostility to the Data-Information-Knowledge-Wisdom hierarchy. However this time, there was no use of the dreadedwisdom word, a pretension that reaches its worst excess with people claiming to be experts in wisdom management. I was also curious about the linking of sense-making and path-finding, and if I needed no other reason to pay attention, then my considerable respect for John’s thinking provided the motive. Listening and thinking made the model more attractive and also suggested some variations that I will work on over the next few months so consider this an initial reflection.
So let’s work through the stages:
Analysis: data to information
Here the model follows convention: myself, Prusak and many others have used this definition. We have a mess of unstructured data to which we apply structure or interpretation in order to inform others, we put the data in context. Raw accounting data lacks context, until we put it into the form of a report. Prusak used to have a good way of explaining this; he talked about messages. If I structure data through process of abstraction and possibly codification then I create messages with which I seek to inform someone else. If that person understands the message they are informed; however if there is no shared context between message creator and message receiver then we are left with data, no information is created.
So we can take that as read, however the interesting addition here is to identify that process as analysis. In the intelligence world this is the process of classifying raw data to provide a context in which the data can be used. The analyst is immersed in the data itself. They are dealing with material at a low level and if the data fits their expertise, and the people they have to brief share the same context then things work. The problem (well one of the problems) is that the context is rarely shared. The ability to work with raw data, is not necessary the same as the ability to see the bigger picture and weak signals, seemingly insignificant material, is easy to ignore.
Sense-making: information to knowledge
Now I have historically argued that knowledge is the means by which we create information out of data. Given that this only happens when shared context then knowledge management can be defined as providing shared context. Now this is not too far away from understanding what the information means and my general definition of sense-making, namely how do we make sense of the world so we can act in it. I think I might be prepared to refine my original opinion here. In the majority of cases in KM, decision makers are making sense of information rather than data and a degree of common context is assumed. I like the idea here of asking what the information means, avoiding assumptions. However I also think that it will be important to get back, in context to the original raw data. That is something we have worked on with the SenseMaker™ software, moving from representation to originating data without intervening stages. As I play with this model over the next few weeks I am going to look at some extra arrows and labels to make this and other points.
Path-finding: knowledge to options
Now the model starts to get interesting, and I sense a trace of one of the best decision models of recent years, Boyd’s OODA loop. Once we have a sufficiency of information, or rather comprehension of information they we are in a position to determine options. I like the idea of calling this path-finding. It has that sense of experimental journeys, the fail safe experimentation that is for me a necessary feature of decision making in a complex space. This path finding creates options, which will have different risk factors and will represent different levels of threat and opportunity. Of course path-finding will require mini-cycles of analysis and sense-making as we start the journey. I am not sure that we can fully separate path finding from execution (or from sense-making for that matter).
Execution: options to action
Neither am I sure that we can completely separate action form path-finding or sense-making. In a more complex space the micro-loops around the model are going to come into play as each step will produce more data, create different contexts for existing information and give us additional insights. However in a formal decision environment these stages are separated. Options are presented to senior decision makers who determine actions. As the model indicates, actions initiate new data and we have to iterate the model. Now, as in market options, the knowledge that an option is being considered, or that an action has been taken changes the space. Given that multiple actions and options from multiple players will all be interacting at any point in time the situation is complex and we will have to move beyond a linear model.
WHERE IS THIS GOING
Now I think the way I want to go with this is to take the operators (analysis, sense-making, path-finding & execution) and arrange them as a type of OODA loop with different input/outputs. I also want to look at creating more direct linkages between the output/input types (data, information, knowledge, options, actions) as interacting with the operators, but with the operators intersecting with more than one input and one output. The amount of intersections would of course depend on the degree of complexity in the decision environment. In the mean time I think the model is a considerable improvement over the D-I-K-W chain and offers more possibilities. I offer it for comment.