Mind the (knowledge) gap

A crucial process in the continued success of any enterprise is that of identifying and bridging gaps in organisational knowledge. Joseph M. Firestone outlines the importance of enhancing problem-recognition, idea-generation and error-elimination capabilities if businesses are to fulfil their potential and achieve the results their stakeholders demand.

Knowledge gaps and their elimination drive progress and competitive advantage. The better we are at recognising them, the more likely it is that the knowledge we use will get us the results we need. They are the first step in innovation. The motivation they elicit provides an important and continuing incentive for the innovative process. In the next few pages I’ll provide an account of:

How and why the recognition of knowledge gaps arises out of everyday business processing and decision making;
How knowledge gaps are bridged through successful knowledge production;
How knowledge management can help to enhance problem recognition, as well as other aspects of knowledge processing, such as creating tentative solutions and testing and evaluating them.

How knowledge gaps occur

Case 1: suppose you’re a medical doctor attached to a healthcare centre. You want to treat a patient with an infection. You believe you can cure the infection by prescribing Ampicillin. You set out to order it through the centre’s order-entry system. The system confirms your recommendation. You order the Ampicillin, which is effective when administered to your patient.

Figure 1 outlines a conceptual interpretation of this very common type of sequence. It begins with a gap between what you want to achieve (to cure your patient) and the present state of affairs. I call this the instrumental behaviour gap. To close it, you use your previous knowledge about infections to decide to prescribe Ampicillin. You take action, by ordering Ampicillin, and after having it administered to your patient, you monitor the result and evaluate the outcome. If the result isn’t satisfactory, you can begin again by planning and deciding on a new treatment.

Figure 1 – the decision-execution cycle

In this scenario, you haven’t produced any new knowledge beyond that of particular facts accompanying acting, monitoring and evaluating the immediate results of your actions. Rather, you used previous knowledge along with that of particular facts. There were no knowledge gaps affecting what you did. You performed a number of sequential tasks that we may think of as an operational business process or work flow, and also used your centre’s order-entry system. You learnt some facts during this process and the success of your prescription reinforced the previous knowledge you had about what should be prescribed in situations like this one. In other words, you engaged in single-loop learning while you performed an uneventful decision-execution cycle.

Case 1a: let’s vary case 1 bit. Suppose that after you log on to the order-entry system, it reports a previous allergic reaction of your patient to Ampicillin. Now your previous knowledge, ie, the knowledge available in your memory, and the knowledge available in your organisation’s information system, suggest different conclusions and you are no longer sure whether you should be prescribing Ampicillin or some other therapeutic measure for your patient. You have now recognised a problem: specifically, you don’t know what therapy to prescribe and must therefore find out if you are to treat your patient. That is, your problem is an epistemic problem, a knowledge gap between what you know and what you need to know. Figure 2 illustrates the change in the decision-execution cycle when a problem arises.

Figure 2 – adding problems to the decision-execution cycle

In cases 1 and 1a, I gave you an example of how a problem may arise from a business process that you carry out yourself. This example was abstracted from the Partners HealthCare case reported on recently by Thomas Davenport and John Glaser in the June 2002 issue of The Harvard Business Review.[1] However, knowledge gaps may also arise out of changes in business processes imposed by external authorities.

Case 2: a clear example from the government sector occurred some years ago under the Carter administration in the US. During this period there was great concern about bringing deficits under control and distributing scarce US-government funds according to need. The Farmers Home Administration (FmHA), a major loan and grant agency in the US Department of Agriculture, became caught up in this concern and mandated that its Program Evaluation Office provide funding formulae for allocating its loans and grants according to need. This left the Program Evaluation Office with a knowledge gap. It had no idea how to construct such formulae.

How we close knowledge gaps

Put simply, to close knowledge gaps, one needs to solve problems. In the Partners HealthCare example, it was necessary for the doctor involved to decide whether the system’s recommendation to avoid using Ampicillin would produce a greater benefit/cost outcome than any other competing therapy including his own initial prescription of Ampicillin. In order to do that he needed to consider the alternatives, evaluate them and learn for himself what the best choice of treatment was. The decision he made, after using the system to determine that the patient’s past allergic reaction was a minor and easily treatable rash, was to reject the system’s recommendation and stick with his original prescription of Ampicillin.

In case 1a, therefore, the doctor had to (1) recognise his problem, (2) seek and acquire information to help him solve it, (3) think about the alternatives and perhaps formulate them more clearly either in his mind or on paper. Then he had to (4) evaluate the competing alternative solutions, select the one he thought would cure the infection with acceptable and treatable side effects after rejecting the other alternatives as errors, and finally (5) use the knowledge he learnt by overriding the system and ordering the Ampicillin.

Three of the most important steps in this five-step pattern have been expressed in various writings by Karl Popper, who called the problem-solving pattern illustrated in figure 3 the ‘tetradic schema’. Popper did not explicitly distinguish information acquisition as being a separate step in problem solving, but there is no harm in calling explicit attention to it as an activity that often precedes formulating alternatives. Popper also did not mention the step of ‘using’ knowledge as part of the tetradic schema. But I have gone beyond the schema and included it here to emphasise, first, that solutions are made to be used and, second, that when a problem is solved, the knowledge is used in the business process that produced the original problem (and awoke the learning incentive that, in turn, drove the problem-solving life cycle). Finally, Popper included a second problem at the end of the pattern to indicate that solutions to problems usually suggest new problems (further knowledge gaps), and that life involves continual problem solving.

Figure 3 – Popper’s tetradic schema

In Case 2, the FmHA Needs Assessment Capability (NAC) project, the same pattern of problem solving I described for the doctor at Partners HealthCare can also be observed. During the FmHA project, there were extensive efforts at information acquisition. Literature reviews, interviews with personnel of other federal agencies, facilitation sessions with national, state and local-level decision makers were all used as sources of information external to FmHA to provide a foundation for formulating tentative solutions. In turn, the alternatives were arrived at using group decision-making processes involving the aforementioned decision makers, along with psychometric scaling techniques, mathematical modelling, and multivariate statistical analyses of census, expert judgement, programme and other federal data to generate alternative formulae for consideration. Finally, the same data and analytical techniques were used to eliminate errors in the formulae and to control other factors biasing the effort to measure need across states and counties.

In the final analysis, we close knowledge gaps by following the pattern expressed in the tetradic schema. When we look at this schema closely, there are at least four important things that we must understand about it in an organisational context:

The tetradic schema is a theory about how knowledge is produced by humans either individually or in systems;
According to this theory, there is nothing deterministic about how we close knowledge gaps and solve problems. Rather, we do this by the application of trial-and-error elimination. The processes of formulating trials and eliminating errors is an emergent one, which at the level of groups and organisations involves self-organisation;
Closing knowledge gaps is, in the end, about adaptation. It is, to paraphrase Popper, about killing our bad ideas before they kill us;
At the level of groups and organisations, problem solving and knowledge use are not enough to both bridge knowledge gaps and bring knowledge into use. In addition, we must recognise a process, called knowledge integration, that links problem solving, also called knowledge production, with knowledge use. That process, which involves knowledge sharing and other activities facilitating knowledge distribution throughout the organisation, joins the knowledge-production process in forming the knowledge life cycle (KLC), a construct that my collaborator Mark W. McElroy and I have discussed at length in our books and other publications. The outcome of any instance of the KLC is a contribution to what we call the distributed organisational knowledge base (DOKB): the products of past KLCs distributed across the various information systems and minds in the enterprise. The DOKB, in turn, provides previous knowledge (as identified in figures 1 and 2) as the knowledge that gets used in operational business processes.

KM and enhancing problem recognition

A short definition of knowledge management is: a set of activities we use to enhance the processes that identify and bridge knowledge gaps. The first of these processes is problem recognition. It is at the boundary between operational business processes and knowledge processes, and also at the boundary of the knowledge manager’s authority, since people performing business processes under the direction of operational managers play an important role in recognising everyday problems in business processes. So how can KM enhance problem-recognition capabilities when much problem-recognition activity is beyond its authority?

First, the importance of problem recognition as the first step in adaptation must be emphasised throughout the enterprise. This won’t happen without agreement among key executives that problems must not be swept under the rug, but confronted in order that they may be solved. Even then, we must expect that many operational managers and operational business-process participants will not want to ‘see’ that the gap between expectations and outcomes is serious enough to justify recognising that a gap between what they know and what they need to know exists.

Knowledge management can assist in moderating the natural fears of people by offering problem-recognition and communication workshops to employees. The objective of these workshops should be to train people in:

Understanding why problem recognition, in the sense of pointing to knowledge gaps, is important for competitive advantage, organisational effectiveness and job performance;
Self-evaluating the results of their activities;
Recognising when outcomes are falling short of their expectations;
Recognising what knowledge and capabilities they need to overcome the performance shortfall;
Communicating about the problems they recognise.

The workshops should use case-study, knowledge-café and storytelling techniques since an important goal should be to provide participants with a variety of interpersonal perspectives on the areas to be covered. It is also important that sharing perspectives in a workshop environment begins to create a community that will reinforce the idea that problem recognition is important. This community may then be organised as a community of practice after the workshop is over. Knowledge management can make a further and continuing contribution by moderating the CoP.

Second, an important aspect of problem recognition is securing feedback on the results of employees’ activities, in order that they are able to monitor and evaluate the consequences of their decisions. In organisations with active quality-management, balanced-scorecard or other performance-monitoring programmes, there is a great emphasis on measuring outcomes and on reporting, and this provides a good foundation for recognising knowledge gaps where they exist. Knowledge-management programmes should support metrics development and implementation activities throughout the organisation, and should support the development of a metrics programme covering KLC processes, KM activities and knowledge outcomes.

Another aspect of providing feedback to people so they can recognise problems is to use IT to provide relevant information (and sometimes knowledge) that is ‘baked’ into the jobs of knowledge workers, to refer again to the Davenport and Glaser article and case 1a. In this instance, the doctor perceived the existence of a problem after the order-entry system reported the previous allergic reaction of the patient, an example of timely feedback tied to the doctor’s role of ordering prescriptions, which in turn stimulated problem recognition and an individual-level KLC.

More generally, I have written extensively about enterprise information portals and their capability to provide alerts to knowledge workers that are relevant for their jobs.[2] As portal technology continues to develop into a full-fledged DKMS, integrating a variety of individual tools into true composite applications, it will be possible to provide baked-in information to decision makers throughout their decision-execution cycles and work flows. In the meantime, KM initiatives should include portal solutions that provide alerts in key work flows of the kind illustrated by the Partners HealthCare example.

Third, the most important way for KM to enhance the problem-recognition capacity of an organisation is to persuade it to accept a policy of openness. That is, a policy of maintaining freedom for all participants in business processes to state that a knowledge gap affecting performance exists and to communicate that view to as many others in the organisation as they care to, without fear of reprisal. Openness here will produce distributed problem recognition and increase the probability that problems will be addressed provided that the enterprise has the capacity to address them. In addition, this type a policy requires KM to receive and allocate resources for an IT infrastructure that will empower staff to exercise this freedom. In practical terms that means, these days, a portal system that will allow the free publication of newly identified knowledge gaps in the context of the business processes, work flows and types of decisions that generate them.

KM and enhancing our capability to form, test and evaluate solutions

What can knowledge managers do to enhance the processes and work flows knowledge workers use to arrive at tentative solutions, and to test and evaluate them in order to eliminate errors? The short answer is that there is a great variety of things they can do, including social, technical and combined interventions aimed at policies, social attitudes, competencies, IT infrastructure, recruitment, social networks, the DOKB, and many other targets. The motto for knowledge managers in this area for bridging knowledge gaps has to be ‘let me count the ways I can help’. Due to space limitations, I cannot present any specific proposals for KM interventions here, but you can find 16 of them (eight on generating new ideas and eight on eliminating errors) in an appendix to this article available at www.dkms.com.

The open enterprise and other issues

In explaining how to recognise and bridge knowledge gaps (see the appendix), I’ve enumerated a number of initiatives that knowledge managers could take to enhance problem recognition, generate new ideas and eliminate errors in these new ideas, as well as old ones. I’ve also offered these in an optimistic tone and with the expectation that if you take these initiatives you will indeed enhance your capability to bridge knowledge gaps and solve problems. Now come the cautions and qualifications.

This set of initiatives for bridging knowledge gaps isn’t complete for two reasons. First, I’ve certainly made some errors of omission. Second, space limitations (even in the appendix) prohibit me from covering a number of promising interventions in the categories discussed, as well as whole categories of interventions targeted at individual and group learning, various areas of knowledge integration and the DOKB.

Further, when you evaluate my proposals, you can regard them individually or more holistically as a pattern. If you read Knowledge Management regularly, you’ll remember my collaborator Mark McElroy’s article in the September 2002 issue introducing the idea of the ‘open enterprise’.[3] The holistic view I’m proposing here is a series of interrelated proposals that together amount to an effort to establish openness in problem recognition, knowledge-claim formulation and knowledge-claim evaluation, and in that way move towards the open-enterprise pattern of organisation (a pattern optimised for innovation and corporate transparency). But getting to the open enterprise is not as simple as instituting a particular set of programmes and policies.

Organisations are complex adaptive systems, subject both to management interventions and to self-organisation at every level of organisational interaction, including that of the individual. The state of the organisational system emerges from the interaction of its management and self-organising behaviour and predispositions. While the number of states of any organisational system is theoretically infinite, when we view real organisations as systems in a phase space defined by their attributes, we believe that they do not constantly change their states. Instead they stabilise in self-reinforcing patterns. We refer to these patterns as attractor basins in phase space. The theory of the open enterprise conjectures that there are four important attractor basins associated with types of knowledge-processing patterns.

The politics of openness in knowledge processing is one of these attractor basins, the one associated with the open enterprise. To get to the open enterprise from one of the three other attractor basins (which I will not provide an account of here), the policies and programs of knowledge managers can, at best, enable or empower transitions. They can’t determine them, due to the emergent character of the patterns and their dependence on self-organising interactions that managerial policies and programmes can influence but not control.

As such, the initiatives I have suggested will not necessarily produce the intended effects in enhancing your capability to bridge knowledge gaps because they may run afoul of the self-organising tendencies of existing patterns. The chances of success will be greater if they are initiated:

As part of a pattern of change initiatives that are synchronised with self-organising tendencies supporting the transition to the politics of openness in knowledge processing (the policy synchronisation method), rather than piecemeal;
On an organisational pattern whose attributes (such as a moderate level of trust among individuals) provide fertile ground for these initiatives in that they enable them to reinforce one another;
In an organisation that is already moving towards the politics of openness.

To summarise, we bridge knowledge gaps when we solve problems. To increase our capacity to solve problems, however, we need to enhance our problem-recognition, idea-generation and error-elimination capabilities. In this article, and the appendix, I’ve presented a ‘laundry list’ of initiatives for enhancing the above capabilities. This list, however, is not ad hoc. It focuses on interventions that might be expected to increase openness in knowledge processing in the enterprise, and to move organisations towards the open enterprise, a type of organisation characterised by an enhanced capability to bridge knowledge gaps (solve problems, innovate), and greater organisational transparency. But if one wants to travel to the open enterprise, the way to get there is not simply a matter of implementing one or a few of the interventions presented. Rather, it is important to develop a pattern of interventions that is synchronised with the tendencies toward self-organisation that typifies the transition towards the open enterprise, or to implement interventions in an organisation that is ready for this transition or is already part way through it.

In short, though enhancing the capability to bridge knowledge gaps is fundamental to an organisation’s ability to cope with its continuing challenges, it is not an easy thing to achieve. It can be done piecemeal with the risk that intended impacts will be negated by pre-existing patterns, or it can be done more comprehensively, but in such a way that it works against pre-existing patterns. Alternatively, it can be done more comprehensively with the intent of synchronising with self-organising tendencies. This last choice is of course the road to travel, provided we can see it clearly.

References

1. See also Davenport, T., ‘Making knowledge work productive and effective’ in Knowledge Management (Ark Group, November 2002)

2. See Firestone, J., Enterprise Information Portals and Knowledge Management (KMCI Press/Butterworth-Heinemann, 2003) and various papers on my website at www.dkms.com)

3. McElroy. M,. ‘Ethics, innovation and the open enterprise’ in Knowledge Management (Ark Group, September 2002)

Joseph M. Firestone is chief knowledge officer for Executive Information Systems and executive vice president, Education, Research and Membership at the Knowledge Management Consortium International. He can be contacted at eisai@comcast.net