Structured Brainstorming

Brainstorming for Strategic Planning

This work was originally reported at the Agility Conference in Boston, 1997. This version of the report was prepared as an Appendix for the book, The Agile Virtual Enterprise, but was dropped for reasons of space. We've left the internal references in.

An edited version of this is at Rick Dove's Paradigm Shift site.

This appendix suggests a technique to feed the decomposition of the enterprise into processes to be accomplished by conversations. It is the result of work performed by ourselves, the AVE Focus Group, Sandia Labs, and the Automation and Robotics Research Institute. It has been an incredibly useful tool, so we really want to give it to you. But it is somewhat aside from the main thread, so we put it in an appendix.

Agile Strategic Planning

Businesses already have tools to support strategic planning to lower cost, increase quality, and decrease time to market. Agility is another, albeit new and different, factor that proactive managers will use in designing the future of their organizations. A central issue is how to create a strategy that has the most beneficial balance of agility with other qualities. Such decisions have a life cycle. And at the end of the planning life cycle, we have the situation where a strategy has been created. The questions are what decisions are the correct ones to support that strategy, to attain the desired agility. Our agility metrics support this end of strategic planning.

But there is an earlier phase, where strategic ideas are generated and evaluated. This section deals with the generation of novel, advantageous strategies through strategic brainstorming. Other methods will have to concern themselves with the evaluation of those ideas through simulations and associated evaluation.

The method we propose is sensitive to agility. Agility is important in situations where change is assumed, whether that change is something that merely happens or is change that has been instigated by the Virtual Enterprise (VE). Most strategic planning methods assume continuity, their worth depending on the accuracy of uniformly extrapolating from prior, real, experience. Again, we see that agility is a radical new idea, because there is no prior experience; new methods must be developed to support it. Although the brainstorming method outlined here is designed to accommodate agility, we believe it to be useful for generating novel strategies across the board.

There are two key ideas in the brainstorming method, the idea of memes and the associated idea of underlying perspectives or principles. We'll briefly introduce those two ideas, though we've been introduced to memes earlier in the discussion on French/English law. We suggest that these ideas be hosted in a role playing activity, because that is a well established productive technique where the core ideas are strong.

Memes

Certain patterns in the environment seem to pop up over and over again. Have you ever passed someone humming a tune and found yourself humming it all day? These things spread by being passed from host to host, as jokes are from person to person. They act almost like viruses, these ideas, pieces of information, and memories, spreading and replicating almost as if they were acting intelligently. Moreover, they adapt as conditions warrant; a Chuck Berry guitar riff will promulgate itself in dozens of transmuted ways in subsequent popular music.

And there are ideas that take on a life of their own, like property ownership, civil liberty, or human rights. The latter two are modern ideas, and the idea of real estate ownership (in the sense that one can sell land as they sell an object) has only a somewhat longer pedigree. It is hard to imagine the eons of human thinking that transpired before these ideas caught on. But it was so.

So powerful are these ideas that it is very difficult for historians to get a perspective on actions and motives before they appeared. Even what is taken as rational thought itself seems to change. The point for us is that as obvious as these ideas are today-perhaps so obvious that we sometimes ignore them-it would have been almost impossible to predict their appearance, understand their content, and appreciate the resulting changes in the world.

No one would claim that ideas like these are themselves intelligent entities in the way that humans are. Nor does it make sense to see the passing of an idea from one person to another as an intelligent act on the part of the idea (as opposed to the humans involved). But when viewed at a high level in aggregate, intellectual tokens (ideas, musical themes, languages) do seem to act in some ways intelligently.

Such ideas act intelligently in the same way that an ant colony or beehive, when seen collectively, seems to act intelligently. The intelligence is like that ascribed to agents of disease, the behavior of which is studied by epidemiologists, or to genes. Individual varieties of genes can be said collectively to exhibit intelligence, using life forms merely as a vector, a host in their drive to maximize benefit to themselves. Benefit to the hosts over time may coincide with the gene's interests, or it may not.

Much of modern evolutionary thinking (that of constructive evolution) is based on this idea of genes as autonomous, fine-grained agents with collective intelligence, intelligence meaning the ability to adapt to enhance certain goals, in other words, to be agile. This is a powerful idea, certain genes acting agily as a learning organization. The idea is so useful that an evolutionary scientist, Richard Dawkins, extended it in 1976 (Dawkins, 1976, 1996) from genetic entities (biology-based memory) to apply to intelligence-based entities. Such entities include songs, ideas, and the like, which are based on memory in the mind or the mind's external stores (books, records, computer memory...).

Dawkins coined the term Meme for this kind of self-replicating entity. This coining exemplifies the idea of turning the conventional relationship of (active) actor and (passive) participant on its head, thus:

"A hen is just an egg's strategy for making another egg."

The idea of memes has itself become a powerful meme, and the idea has found wide use in the artificial intelligence community and in studies of cognition and complexity. We submit the concept as one of the bases for this structured brainstorming method.

Relevance to Brainstorming

In particular, we want to understand what makes a meme tick sufficiently well in order to look at novel future alternatives. By understanding what makes memes catch on, we get closer to understanding the general shape of unexpected change, at least those changes that are humanly driven. The AVE Focus Group's survey indicated that these are the most important, and most dangerous, to businesses.

When we get to tactical agility, we consider the VE as a collection of agents which are dynamically coupled. We concern ourselves with which agents are the best to add or to take away from the VE, for certain agility goals. More importantly, we are interested in understanding and engineering the coupling that dynamically binds those agents. For the case of strategic brainstorming, we'll define the agents differently, as memes.

If we are brainstorming a Type 2 VE-in which we have a collection of memes (meaning capabilities) and are brainstorming for an opportunity in which to use them-we'll want to understand the natural tendencies of the memes and the logical directions in which they will go. This is different than evaluating core competencies of the VE in a static manner and evaluating opportunities; instead, you track the meme equivalent. The result is as if you were seeing how the existing core competencies naturally want to evolve and were looking for an opportunity somewhere on that track.

For brainstorming a Type 1 VE-in which we create a hypothetical opportunity and then brainstorm for the correct recipe for a VE which will suit that opportunity-the situation becomes more complex than the above (Type 2 VE) example and also more interesting. Here, you want to develop a meme aggregation that has a natural learning path which will take it in the direction of the opportunity and that has the natural ability to adjust as the opportunity adjusts. The situation is complex because, unlike genetic evolution, you have some freedom to choose partners and engineer the dynamic coupling among them.

A few characteristics of memes are important to AVE brainstorming: Memes can be categorized according to their domain of influence, and that decomposition is the same as the infrastructure categorization we've developed. Memes associated with contract law have effect only in that domain; similarly in business practices. We have an example of each,

A Key Difference: The Engineering Paradigm and The Role of Common Law.

Those examples show another characteristic as well, that memes can be grouped into higher classes. The French engineering meme concerned with business practices and the Code meme which deals with legal issues both are part of a higher level class of memes, at some level a meme-class. In this French engineering/code case, it's a class that forces a centralized, top-down control of processes.

The more robust memes settle into a system by creating a balance with a complementary meme, creating (or following) a simple symmetry. In our example, the English engineering meme was balanced by the French one; the common law meme opposed by the code law. This is even more evident at the meme-class level. So centralized control systems are opposed in a system of, say, human society, by a decentralized control system, with special cases appearing in law and engineering.

We submit that all memes (of interest to the AVE) fall under a few basic categories of meme-class. We'll present the four basic sets of meme-classes in the section below.

Identification of memes in a system is an art, is highly subjective, and often produces results that are trivial or worse-complex but without providing insight. It's much easier to start with a meme-class and develop or discover a specific meme than the other way around.

To review: the idea behind our metrics is that by understanding the processes in AVEs, we can measure their ability to adapt. In order to have productive brainstorming, you need to have a similar understanding of processes, how they adapt and learn in your system. We find that in complex systems a useful way of understanding how processes or agents adapt is by understanding memes. Furthermore, memes themselves have underlying principles, which we've called meme-classes. Useful brainstorming can leverage these relationships by understanding the principles underlying the memes and especially the fact that each meme-class has a dual, an opposite or complement.

Basic Underlying Principles/Controversies

Here are the four complementary pairs of meme-classes which we think form a complete basis for memes in AVEs and which can be used to create a structured controversy method for brainstorming.

A disclaimer: work presented in other chapters has a more thorough formal underpinning than here. In particular, the work on abstraction that is novel to the metrics' quantification has a formal basis in category theory; the social modeling relies on situation theory; and the infrastructure and communicative act parsing relies on long established organizational and information theories.

The work of this section is drawn from the history of philosophy and from some working ideas that have been refined through empirical work by the AVE Focus Group. It has not, so far, been thoroughly investigated. But we commend it here because it seems right, it works, and it fills a need in agile planning for which there is no comparably apt technique.

The four basic controversies are:

1. Realism versus Phenomenalism

2. Intrinsic versus Random Order

3. Evolution versus Revolution

4. Centralized versus Distributed Control

These four are fully symmetrical (meaning only that each constitutes a complementary pair, and that any grouping of two complement the remaining two). One would strongly expect to find such symmetry in any sufficiently clean basic disposition of principles. The figure shows the symmetries; see the Soft Mathematics Case Study for a French/English Breakdown and Part 4 for relevance to the current problem in the defense world.

The first two pairs of meme-classes are, relatively speaking, more fundamental than the others. They deal with belief systems, with how people believe the world works, and with the characteristics of the fundamental laws of nature. The second two are underlying ways in which the world is instanced, how it exists, using those laws.

The first and the third relate to the dynamics of entities, what constitutes being, and the second and fourth reflect controversies about the dynamics of being, how systems are formed and interact.

Each of the four, of course, denotes a primitive controversy, that is, two classes of opposing principles around which many types of ideas can be formed and self-sustaining as memes. The first listed perspective in of each of the four controversies tend to go together; it is more consistent, for instance, for someone who holds realism as a belief to believe also in intrinsic order, evolution, and centralized control, than to take any of the opposite choices along with any of these.

Each of the four are described below. It may be of interest to observe that the institutional default familiar to most businesses within in Western civilization (as well, not just incidentally, as the Abrahamic religions, Judaism, Christianity, and Islam) is based on the first choice in each of these meme-classes (realism, intrinsic order, and so on).

Realism versus Phenomenalism

Realism is illustrated in the belief that a tree does make a sound in the woods even if unheard, and, more generally, that the woods exist independent of any need for human existence. The world is real, and humans experience it. Some humans, because of individual differences, may experience it differently, but reality is out there, immutable. The notion of absolute truth follows nicely from this. Truth reflects reality. A scientist who works under this belief believes that it is the role of a scientist to discover and understand the truths of this reality.

Phenomenalism is the group of beliefs that the world we encounter is a collection of ordered experiences. Among diverse observers, we may share some consensus about those experiences as truth, and diverge on the actual nature of the world in other ways. Because we can never encounter the real world, only sensory phenomena, reality as an independent state is either moot or unknowable. A scientist working under this system strives toward inventing new ways of explaining phenomena to expand the class of shared consensus. How the scientist communicates and explains is in a sense more important here than what is explained.

Each of these views contain a great many ideas, some of which have a very long tradition. Many memes are spawned and take on their own lives, strengthened in an important way by its complement. Essentially all philosophical discourse uses memes from this class.

As a cogent example, suppose you are brainstorming early in the game of our missile AVE example. The problem you are examining is

"What kind of product strategy is best for us?"

We've picked a mundane example because of its illustrative value and because of the leverage we get by keying off of the tactical example. After addressing this question, the missile prime might brainstorm on how best to support that strategy through an AVE. A simulation would validate/modify the approach, resulting in the input for the metrics.

Realists might hold that the way that the world works is set, that there is a real, strong idea of a better missile design and a less good one. Their brainstorming could be limited to the context of that reality, trying to understand and negotiate it.

Phenomenalists, on the other hand, could believe that a missile is a product, that its worth is in meeting customer requirements, and that there is no basic good design outside of that context. Since the context is one of how the customer sees the world, they would try to analyze not the world of design, but the world of the customer's predilections, foibles, concerns, and inner drives. They might also assume that the customer's world view differs fundamentally from the prime's.

Intrinsic versus Random Order

Some hold a belief that the world is a clockworks; moreover, that it is a single clockwork system. It all makes sense at some basic level, and laws at work in one area are part of the same script as laws that work in another. This is the class of views that are under the belief systems of intrinsic order.

A contrasting view asserts that the fact that we have scientific laws which work does not imply that the world has a single, inner logic. Much of the world is a result of accidents-many of the physical constants for instance. So while we've been inventive enough to determine some local order in the cosmos, underlying it all is a higher order of chaotic system. (We've used the term random order here for historical reasons; the more formal term would be chaotic order, which is an order. Randomness is the lack of order.)

Our missile brainstormers, if they are taking the former position, might have views about economics, politics, and society that presume that the laws of design and the laws of business and the laws of good intention are all congruent. What's good for Hughes is good for the U.S. and the world. Well motivated engineers will make better products, and that's good for business, good for the prime. In the large, the system works.

Their contrarians, taking the other option (favoring random order), would hold that what's good for the prime may not be directly and completely good for the customer (the U.S., or part of it anyway); also that there are many areas where competing forces are at work between, say, best design and corporate goals. These brainstormers will be struggling with finding the correct balance of these competing forces-a game of compromise.

It is easy to see how the two groups would gravitate to radically differing product strategies.

Evolution versus Revolution

The root of the ideas that we class as evolutionary is intuitively understandable. Adherents to this collection of ideas would hold that the law of cause and effect is apparent in every process. Sometimes a small modification in cause produces wide modification in effect, but the function is still apparent. The key discriminator of this class is the belief that any outcome can be predicted if enough causal factors are known.

The opposite hand holds the belief that most every important process in the world has characteristic events where some instability is reached and basic causal conditions are changed. The important, defining element of the process is not the periods of relative stability between punctuations where change is predictable and gradual. Instead, what matters are the periods where many important things change.

In our imaginary brainstorming session, the evolutionists might concern themselves with, say, a product that will provide (a considerable) advantage over the competition, by better understanding and exploiting the currently applicable rules. The revolutionaries might focus on how to create a product that changes the rules to their advantage.

Centralized versus Distributed Control

The easiest to grasp controversy of all pits the paradigm of central control against that of decentralized or distributed control. This (fourth) class and the previous (third) class are not controversies about how the world works, but about preferred ways that the world instances. For both of these, this and the prior one, there are many examples in nature.

In animals, a centralized control system is a mammal with a brain and nervous system. A decentralized example is a bee colony. In U. S. government, central laws and taxes come from Washington, decentralized laws and taxes from innumerable state capitals and town/city halls.

The representative of central thinking will look to a product that leverages the strength of the prime as a prime, so that the prime as a prime is made stronger. The decentralized-minded brainstormer may seize upon a more peer-to-peer AVE based strategy in assembling the portfolio of product options.

Role Playing

To the insights noted above, we'll add two techniques. The first is tried and true, that of role playing. The second is a way to identify process tendencies by playing with Dooley Graphs.

To perform your structured controversy brainstorming, you'll want to assemble a small group of people, less than ten. Ideally, these will be motivated, alert persons from diverse backgrounds. No prior skills are required beyond minimal group skills, but it's obviously necessary that they be familiar with whatever domain you are brainstorming.

You'll divide these people into two groups in different combinations throughout the exercise. You should have a person present who is in neither group and acts as recorder, since the ideas will come hot and heavy; the best ideas will not be apparent at the time and will have to be identified on reflection. How the group is divided is important, as you'll see.

You should also have a facilitator who is familiar with the method and who can set up the problem, This person will also guide the discussion and insure that players stay in character. The session begins by the facilitator explaining (in about 30 minutes) three things:

What the method is all about, at the highest level. The discussion on memes can be omitted, rather focusing on the idea of getting new perspectives by looking at situations by taking a new basic slant in a structured controversy.

What the four basic religious principles are and what each controversy is in each of the principles. (Since the basic positions of course sometimes do indeed reflect various participants' accustomed religious positions, and since traditional arguments about politics and religion, per se, are not the matters of most interest in these brainstorming sessions, we suggest that the facilitator place no moral emphasis on such terms as religion or God or atheism. Instead of relying on religious and political terms, the facilitator may prefer more neutral labels such as world views and presuppositions.)

What will be the problem of the day be around which the group will be brainstorming. Examples:

"What can we do with this upcoming bid to stay strong in the missile business?"

"What new markets should we move into or generate that leverage our strengths?"

"What kind of organization should we become (given a set scenario)?"

"Oops, something (particular) happened, what should our response be?"

Once the problem is understood, you announce that we'll brainstorm on one of the world view (religious) issues. Go around the room and ask each person to identify their default position on that issue. You'll probably start with the simplest issue, centralized versus decentralized control. The process of declaring players' underlying beliefs is a little tricky; since many folks haven't considered their positions before, they may claim to be naturally of one bent while it is historically clear that they are in the other camp. That's okay. Don't get bogged down in this. Take each player at their word. Now, ask each person to take the opposite position for the game,

Divide the group up as evenly as possible into teams. Ideally, the facilitator knows the personalities involved, and is able to level the teams in terms of argumentative ability. The effort may be wasted if a balanced dialog cannot be maintained and unless each person (as well as each team) gets their say.

The game proceeds with few rules, as an informal debate. Each side projects the stated problem according to the basic assumptions they have temporarily adopted. The nature of advocating each position should include arguing against the position of the other side. What creates the richest possibilities are three things:

That bright people are using perspectives which are contrary to those which they otherwise use (outside of the brainstorming exercises)

That they actively advocate those positions (and present new approaches to products, markets, business strategies...)

And that the basic issues (on which the positions are based) are defined in a way that mirrors fundamental controversies both with their natural divisions, and in ways that spawn or leverage robust memes.

In conducting the exercise, normal group management skills apply; keep the focus, watch for dominating speeches, repetitive exchanges, and dwelling too long on a topic. You should arrange to have an out-of-character signal, so that a person can stop to ask questions if needed, but otherwise save comments and complaints for the time between rounds. Unless the signal is invoked, all players are presumed to be in character.

Each round should only last 30 minutes or so. During this time, it's often the case that players are dying to take the other side's position. So it's a good idea to give the option of swapping sides for the subsequent section. Don't worry about people stretching their assumptions if they switch back to their natural beliefs. By this time things will usually have transcended preconceived constraints.

Select another set of basic issues, choose sides again, and have another round, still working on the same problem. It is up to the facilitator to select a basic controversy that is apt for the problem and also takes advantage of some of the issues raised in prior rounds. Players tend to carry over gems from one round to the other, and it's good to provide a different medium for forming the thought.

Only spend two or three rounds per problem statement. You'll quickly exhaust the pool of spontaneous insights. Move on to another problem statement. If you have an important problem that you really need addressed, make it the second problem, having the first as a warm-up. But don't make that first one trivial.

The whole session should occupy no more than a morning or an afternoon. The facilitator should not be any player's supervisor, and the session should not be tape recorded. All these are impediments.

In many cases, the payoff comes in a second session in a few days or a week with the same players (no more than two substitutes). This time, everyone understands the idea of playing a role, and you'll have some results to report from the first round. If the analysis between sessions is competent, and the group is typical, they'll be surprised at the gems that can be found; they zip by so fast when the discussion is underway, they're not noticed.

This general method was created, refined, and validated by the AVE Focus Group and has been used elsewhere with positive results. In all cases, the topics were not narrowly constrained to engineering an AVE.

Future Work: the Dooley Graph Link

We envision adding another dimension of the structured controversy method: fleshing out the structure of the basic controversies which have been defined (soft modeling) and focusing in on engineering of agile systems. In particular, we believe that certain configurations in an organization are inherently agile (given certain contexts), and that those configurations create recognizable patterns in the resulting Dooley Graphs. This insight extends the brainstorming method from a role-playing game to utility by an analyst or analytical team using a modeling workbench.

This is work to be done by others, and is being looked at in the agility research of the Work and Technology Institute.

Some Results

You might be interested in some results of this process. A group brainstormed on general threats to the corporate economy with the intent of advising some research. The severe threats that were identified were the usual ones: pandemics, terrorism (especially certain forms of information terrorism), and the like. The only really interesting threat, one examined in detail was the unexpected collapse of Moore's Law. This is the principle (to which the world's economy is apparently addicted) that processing/storage capability doubles every 18 months at the same price. Perturbations in this phenomenon will have profound consequences.

Interesting. But perhaps more useful will be the following eight trends that could emerge. These guided the planning outlined on page -49.

Brand Loyalty Plus: People will much more relate their brand of products to their identification of their life-style (and ethnic/group identity).
MegaWealth Generators: As a result of this (and other factors) the "brand name" manufacturers will become few powerful wealth collectors. But they will do less and less of the actual manufacturing; that will be pushed down the supplier chain, perhaps into AVEs.
A Third, Dynamic Class: Markets and innovation will be highly dynamic. The Have-nots will grow in size. There will be a new class: the Used-to-haves.
Wealth by Ability to Change: Wealth will be more granted by analyses of the investor community. Immediate wealth (profitability) will be eclipsed by estimates of how profitable you are likely to be in the future. Your wealth will be determined by impressions of your ability to stay ahead of the power curve (agility?).
Products as Strategic Weapons: Delighting the customer will take a back seat to the product as a competitive weapon against the competition. (The Microsoft model.)
Lifetime Product Marketing: The social identification of the product will be leveraged for after-sale sales: example, lifetime improvements in autos.
New Social Roles for Commerce: More social services will move from government to commercial aegis, under the megaprimes who will be less nationally concerned: and be tied in with brand identification. (e. g. insurance company reinvention.) This will redefine civil responsibility (and the role of haves in the supply chain).
Integration as the Goal of Research: Massive payoff will be in the primes who can integrate processes against the soft market context, leaving the innovation of processes to supplier "partners."