Last month in this column we began the design of a core-competency knowledge management system, and developed the principle design requirements. We also introduced the components of a design architecture that fulfills those requirements, which we will explore here. The context of our design is a GM Metal Fabrication plant with exceptional competency at creating high-variety, low-volume production processes, and their desire for training both new employees and a broader base of existing employees in this competency.

Any knowledge management system must recognize the rapid change in knowledge value today, and provide active means for evolving the knowledge base under management. Even leadership core competency knowledge, if it remains static, becomes irrelevant in the churning competitive environment. The design requirements (Dec. 97 essay) focus on the dynamic nature of knowledge capture, dissemination, renewal, and creation; and recognize the need for transparent training that doesn’t interfere with daily employee productivity.

As a start, we need to build one process that can capture the innate knowledge and insights that a few people possess and another process that can effectively plant that knowledge and those insights in the minds of others. And because the value of knowledge and the nature of its application changes constantly, these processes must be highly change proficient. Consider the alternative: if we should actually succeed in capturing and packaging the insights of a few people today, and also succeed in feeding this boxed wisdom to everyone else, there is both a risk that the contents are incomplete and a time when they become stale—better to let things compete for acceptance than to institutionalize rigor mortise.

Leading management sage Tom Peters says it well: “I’m totally opposed to the learning organization idea. I argue for the forgetting organization. You get droids when you have too much training and too many people thinking and learning in the same way.” (Wired magazine Dec. 97).

So we need a process that captures wisdom from those who have it, even if they can’t articulate what it is; a process that seeks wisdom wherever it may be at the moment; a process that actively renews (improves, upgrades) its content; a process that creates and accepts new knowledge when it’s there to be had.

Reasoning our way through this, our knowledge capture process has grown into a capture, renewal, and creation process: the activities that identify and package the right stuff. But we’re still going to have problems if this stuff is simply put in a box and handed over to a separate and dedicated “teaching/training” process: for one it’ll go stale, for another the teachers will not be quick to change what they teach.

A little more reasoning is needed. New employees come in the door and existing employees change job functions constantly throughout the year—frequent events that trigger a need for training activities. On the other hand, deliberate knowledge generation typically relies on the slow-to-admit and finally undeniable failure of existing knowledge as its triggering event. This tells us that the knowledge generation activities are better tied to the training triggers and leads us to the conclusion that we don’t want separate generation and dissemination processes but, rather, one integrated system, a system that generates/reaffirms knowledge in the process of teaching it. The implication here is that the people being trained will be the agents of knowledge generation as well as the triggers.

We don’t want off-the-shelf knowledge to feed to people, but a training process for discovering and reinterpreting appropriate knowledge and its application.

We’ve addressed the stale knowledge problem and the stuck-in-a-rut teacher problem. Now reality bites: performance pressures preempt time off for training and postpone dollar commitments for training resources.

Rather than problems, these are godsends in disguise. We don’t want dedicated training resources—they institutionalize the rigor mortise. Instead, we want a rotating mentor-student relationship that exposes the wisdom of real workers, and challenge them to explain their insights explicitly. And we don’t want time-off for training—that encourages the wrong knowledge focus. Instead, we want training to occur during the process of solving real problems—with solutions that provide real value in real time to the organization. We’ve called this employment of real people solving real problems in real time Realsearch. (Details can be read in “Realsearch: A Framework for Knowledge Management and Continuing Education” available on the web site at www.parshift.com.)

So we know that we need a highly change proficient process. The cases analyzed here in the last four essays show that we can gain this with a framework/module architecture based on RRS (Reusable, Reconfigurable, Scalable) principles (Sept./Oct. 97).

We define the architecture’s framework as a set of evolving standards that both constrain and enable the interactions of compatible system modules. There is both an implicit and an explicit framework.

The implicit framework is present whether we design it or not. Examples include the local corporate culture, global corporate policies and strategic plan, regulatory and legal constraints, the union contract and work rules, communication infrastructure (e.g., electronic distance learning technology), and skill sets and capabilities of the existing workforce. Though these are all real parts of the framework, practically speaking we can have no immediately effective hand in their redesign—they are the “givens” of the framework, and for the most part are the constraining portion; they limit what is possible.

We will focus our design effort on the enabling portion of the framework—that part which provides the required adaptability to changing knowledge values and application requirements as well as changing personnel priorities and profiles.

Conceptually, we design the architecture’s framework after we establish the change issues it must accommodate (refer to Dec. 97). The reasoning process we went through here earlier coalesces these issues around strategic themes that emerge as our key framework elements.

The accompanying diagram shows these strategic theme elements as orange bubbles connected to each other as well as to functional activities that support these themes. This “activity map” depiction is adapted from Michael Porter’s strategic positioning work with purpose: the framework themes constitute a strategic mesh that forms the core “standards” of the business practice we are designing. By standards we mean the (relatively speaking) unchanging part of the knowledge management practice that defines plug compatibility for the changeable modules (which are not to be confused with the functional activities in the diagram).

The connecting lines convey a strong support relationship which is generally neither uni-directional nor strictly hierarchical. Thus we have themes supporting themes as well as activities supporting themes. More connectivity indicates a tighter weave of mutual support, leading to a more consistent, more compatible, and higher leveraged set of elements.

A Quick Look at the Framework

1) Solve Real Problems: When learning time is focused on solving a real problem for the business, the time spent has direct and immediate payback, and the relevancy of the knowledge is self ensuring. In our GM context, topics for analysis and solution work are chosen for their abilities to shed new light on existing processes and/or develop new processes with superior characteristics.

2) Fundamental Principle Based: The problems associated with static knowledge are greater to the extent that knowledge is specific and narrow, and lesser to the extent that knowledge is fundamental: in physics, the Heisenberg uncertainty principle is less likely to change than is the location of a particle it governs. Thus, we focus on principle-based knowledge with examples of good application. In the GM context, the RRS principles themselves are appropriate as the core competency GM wants to manage is about adaptable systems. The RRS principles become more accessible, however, when they are translated into local rules, using the vernacular of the plant, its processes, and its people.

3) Metaphor Model Packaging: Analysis of existing processes and development of new processes will both be packaged by the students as modified metaphor models (Oct. 97); providing representational consistency and illuminating the key competency features. In the GM training context the standard metaphor model is modified to exclude the need for written descriptions and include a section on pertinent value metrics.

4) Students Renew Knowledge: We want a fresh look continually at the knowledge base. Students provide this when they build and refine metaphor model “candidates.” The models remain candidates until the “QA committee” (mentors and prior students) decides they are worthy of entry into the official case study library. Students are also responsible for identifying and adding applicable outside information and readings to the review library. With less historical investment in the status quo students are more aggressive in their outside information considerations.

5) Capture Equals Mobilization: In our earlier discussion we arrived at the need for an integrated process, one which utilizes those being trained to develop the very substance of the training material. The act of capturing core competency knowledge is the same act that also disseminates it. Mobilization is further enhanced by using the metaphor model as the knowledge representational vehicle.

6) Value-Based Application: Finally we come to value—important in both the student and the company context. Developing new knowledge is not easy, and developing knowledge at the depth of insight is tougher yet. To this end a portion of the preparatory work is aimed at relating mastery of the fundamental principles to motivating personal values. On the company side, and in the GM context, knowing how to analyze or develop a highly adaptable process is not necessarily good if no value accrues to the company. To this end all process analysis and design work is accompanied with a performance value analysis.