This is a summary of the talk that David Perkins provided. 

The idea for these next few minutes is the 10,000 foot view. It’s not an effort to re-summarize Katie’s synthesis or restate Zollo or Vera’s work. It’s not an effort to touch on everything somehow. The idea is to provide a wrapper for some of the things that were said, at this gathering and the last one.

This talk has 4 themes: defining flexpertise, understanding it (why would we want it), valuing it (what is it worth), and fostering it (what do we do to get more flexpertise, when we want more of it).

Defining flexpertise

A quirky aspect of expert knowledge is that it has vertical and lateral potential.

• Vertical – direct application to the kinds of problems for which we have it. The circuit expert designs the circuits.
• Lateral – more surprising applications, off to the side. It can connect with other areas, inform them, yield unexpected insights, address problems that require perspectives from multiple disciplines.

Almost all knowledge has lateral potential (not just vertical). But there is a catch: knowledge gets stuck in its vertical potential. So we can define flexpertise as the art and craft of exercising the lateral potential of knowledge, of getting it unstuck so it works on the sides and not just straight down.

Understanding flexpertise

We have to ask ourselves…. Why do we end up needing flexpertise? Why is it that knowledge gets stuck?

We use the phrase “drill down.” That is what tends to happen to the knowledge we have: we drill down, go deeper. What keeps the drill in the hole? It’s useful! That is what the knowledge was first and foremost for! But the lateral potential is a bonus that we don’t want to miss.

What keeps us fixated “downward”? At the individual level, we may cling to past success (“it worked before, let’s do the same thing again”) or become risk averse (“it’s safer to do the same thing again”). Perhaps the deepest reason is the anesthesia of familiarity. People don’t know what they know. Expert knowledge is tacit. You don’t know it articulately. It comes to mind in contexts. This is no problem for drilling down; the expertise you need gets evoked by the context, through pattern recognition. But that is a huge problem for lateral application. It will not be automatically evoked by context. You have to dig it out, try it on. So that mechanism of automaticity that empowers the drilling down gets in the way of lateral potential.

At this gathering, we are exploring the flexible firm, so we also have to consider what keeps us drilling down collectively. Sometimes, this is an organizational version of the same factors. Organizational level knowing is also tacit; the organization doesn’t know what it knows. There is also risk aversion, compartmentalization, etc. that collectively undermine our exercise of lateral potential.

We might think that individuals and organizations are stuck with inflexibility. Not so! Good news comes in the form of a paradox. Inflexibility is flexible. Huh? We know where inflexibility comes from and what can be done about it.

• The individual: In individuals, inflexibility and flexibility reflect habits of mind, mindsets, and skillsets. It’s clear from psychological research that those habits of mind can be nudged in different directions.
• The individual in the organization: Moreover, how an individual behaves in an organization is not just a matter of individual psychological profile, but is also very much in response to what the structures of the organization make easy or make hard, invite or discourage.
• The organization in general: Taking this further, organizational routines, structures, and culture can discourage or encourage the flow of information and ideas across boundaries in ways more or less likely to mobilize the lateral potential of expert knowledge. That’s what dynamic potential is all about!

So we are not stuck with inflexibility at any of those levels. Flexpertise could be seen as expertise in exercising flexibility.

Valuing flexpertise

What is the value of flexpertise? Can you go too far? How much of it do we want?

Flexpertise does have its costs: costs in effort, time, and risk. These costs need to be recognized and balanced against the gains. So valuing flexpertise cannot be categorical; it needs to be measured. There is a sweet spot, a balance, that we need to find.

We don’t all have to be Semco. There is sometimes no reason to be. The most extreme models aren’t always the best.

Should we be scared of too much flexpertise? It does scare people! But remember the simple fact that in general, we undershoot lateral potential rather than overshooting it. When we reach for flexpertise, we have to remember its costs are costs that can be controlled. We can use quick prototyping. We can use managed risk. Smart flexpertise would manage the downsides (dumb flexpertise might not, but who wants that?).

So, on the whole, don’t be scared of too much flexpertise. Those worries are, on the whole, a defense mechanism. They are a manifestation of “immunity to change,” or shyness in the face of the unknown.

Fostering flexpertise

There are lots of ideas on the table about fostering flexpertise, such as dynamic capabilities or improvisation. We can look to foster flexpertise on four fronts:

• Lateral mindsets
• Lateral skills
• Lateral structures
• Lateral leadership.

These characteristics should pervade much of the organization; but again, not always everywhere. We need to find the sweet spot.

As Maurizio Zollo noted, dynamic capabilities should not bee seen only as routines to change routines, but also routines to change culture, values, etc. In other words changes in mindset, skills, structures, and leadership.

As Dusya Vera discussed, improvisation uses what you already know. A fundamental problem of flexpertise is that it starts with tacit knowledge. Improvisation mines and utilizes that tacit knowledge, which is great. It helps us churn up our expertise and spread it around. We can foster skills of improvisation, structures that allow time and space for improvisation, and styles of leadership that evoke it and model it.