Dion on "the edge"

In his latest ZDNet blog posting, Dion Hinchcliffe discusses "edge businesses."

I like the reCAPTCHA example and his discussion of how external interfaces are beginning to blur into a fractal mixture of business, customers, and suppliers.

However, the focus on business structure makes me a bit uneasy. It's true that changes in transportation, communication, and information technologies open up disruptive possibilities...the past 150 years provide many examples. However, there's also a risk that pursuing structural opportunities will begin to crowd out business fundamentals. The explosion in financial engineering over the past 40 years is perhaps the most prominent recent example.

Bottom line: As a long-time reader of Dion's blogs, I like the conversation and think it's worthwhile. But, discussions of business fundamentals that don't change (e.g., deep domain expertise) should not be negelected. For a good recent discussion see The Puritan Gift.

Closing the Learning Loop

I tend to think of learning in terms of sensemaking (e.g., Klein's Data-Frame model), rather than, say direct instruction or constructivism.

However, an article I was reading recently (not on the web) triggered some thoughts about another view that I've seen expressed in a various disciplines...a view which focuses on how we internalize then communicate experience.

The key steps are:

1. Experience - a lived narrative. Raw sense data is stored in a loose structure that includes time, past narrative fragments, emotional intensity, etc.

2. Retrospective structuring - an ongoing process where the experience is relived, sensory data is structured, inferences are drawn, and individual data items are emphasized/de-emphasized. As the raw data is recalled and abstracted, a propositionally coherent structure is imposed on the data. Karl Weick sees this as a key aspect of sensemaking...hence his famous observation "how can I know what I think until I see what I say?" However, the emphasis here remains on the narrative...within an inferred structure that is the ground to the narrative figure.

3. Communication to self/others - this is the culmination of the previous step where the key aspects of the experience are verbalized.

This last step is where two approaches to communication become apparent. Depending on the context, either a narrative or a propositional approach may be taken. If the purpose is to entertain, narrative is likely. If the purpose is to persuade, propositions are often favored, especially in technical contexts.

None of this is especially noteworthy...except for technically-oriented professionals. We tend to be proposition-heavy and narrative-light in our communication.

Why is this a problem? It may not be. If a rich shared narrative exists, a proposition-heavy communication may be acceptable. However, if no such narrative exists, our audience is likely to struggle...if we don't wrap the propositions in a memorable narrative. Or, in a collaborative context, we may struggle to get beyond "talking past each other's propositions."

So, try to ensure that the learning loop is closed when communicating....from personal narrative to generalized propositions, then back to contextually-framed shared narrative.

Creating such a narrative for this set of propositions is left as an exercise to the reader...as is a discussion of the risks of over-emphasizing narrative (e.g., being perceived as superficial and manipulative)... :-)

Forgetting and Learning

While watching a PBS program on the "Botany of Desire" today, I was reminded of how important forgetting is in learning.

The program discusses THC (the psycho-active chemical in cannabis), a neurotransmitter with a similar shape (anandamide), and the importance of being able to forget (one effect of both chemicals) to avoid being paralyzed by fear or overwhelmed with contextually irrelevant data. A summary is available at "Cannabis, Learning, and The Botany of Desire" (pp. 14-ff).

Karl Weick, in his classic "The Social Psychology of Organizing", discusses the same issue as it relates to organizations. One section addresses the importance (and difficulty) of organizational "forgetting." In it he quotes a Journal of Applied Behavioral Science article discussing Albert Speer's (Hitler's minister of armaments) use of Allied bombing raids to enable organizational forgetting:

"These raids were 'helpful,' according to Speer, because they destroyed the filing facilities, those containers of paper which enable organizations to establish traditions, procedures, and so on, which are mainstays of bureaucracy. Speer was so enamored with the results of these bombing raids that, upon learning of the destruction of his ministry in the Allied air raid of November 22, 1943, he commented: 'Although we have been fortunate in that large parts of the current files of the Ministry have been burned and so relieved us for a time of useless ballast, we cannot really expect that such events will continually introduce the necessary fresh air into our work' (Singer and Wooten 1976, pp. 86-86).”

Organizational learning currently focuses on acquiring, structuring, storing, and retrieving information. As the volume of information explodes and hyperconnectivity moves many decision contexts into the complex domain, forgetting may need to be added as a core learning competency.

Contextualizing the Web

The techno-centric approach to adding context to the Web is the Semantic Web. Two recent articles make me wonder if the sensemaking approach to adding context will be the Mobile/Social Web:

• Techcrunch has an interesting discussion of a presentation by Morgan Stanley analyst Mary Meeker highlighting the astonishing growth rates associated with the Mobile Web and the iPhone/iTouch. The entire presentation is available and worth reviewing.
• Dion Hinchcliffe has a much longer discussion of the shift from traditional Web sites to the Mobile/Social Web.

Neither really address what may be the truly revolutionary aspect of this shift: dramatically improved sensemaking driven by the contextual aspect of mobile/social. Instead of the user being almost totally responsible for framing and crafting IT/data to fit a use context, mobile/social IT can automatically sense key aspects of the use context and perform some of the framing/crafting work. The result: a potentially dramatic increase in the sophistication/depth of data/services that can be incorporated into a use context.

I suppose I should note that it's not either Semantic or Mobile/Social...it's both.

Limiting Innovation

Peter Merholz (Adaptive Path) talks about the "Highlander Principle" in this HBR blog post. He asserts that "there can only be one" true innovation in a new product if it is to succeed.

This is another area where cognitive & social limitations are often obvious only in retrospect. A use context of any complexity is usually tightly coupled to a larger ecosystem in ways that are difficult for both novices and experts to see. If an innovation is not almost "friction-free", the odds of it successfully carving out a niche (which usually means displacing existing capabilities and/or re-allocating/configuring existing resources) are low.

MIT's Sloan Management Review has a related discussion: small-fast use context experiments are a significant new tactic for innovation...at least for IT/data-intensive ecosystems.

DoD SOA Acquisition

The SOA craze has generated massive volumes of techno-talk, moderate volumes of biz-talk, and little discussion of how to move from acquiring systems to acquiring services.

Thanks to a colleague, I just ran across a good discussion of this issue: the AFEI's "Industry Recommendations for DoD Acquisition of Information Services and SOA Systems (7 July 2008)."

What I liked in the paper:

• Mission Logic and Mission Data receive equal attention
• The SOA stack shows dependencies up and down (i.e., SLAs & Reverse SLAs)
• The need for a different set of roles/rights/responsibilities is addressed. Table 1 seems to imply that the "role center of gravity" is shifting from technology to mission...which seems appropriate.
• Appendix A is a thoughtful proposal on how to incrementally employ SOA concepts

Things I wish had been discussed more:

• How organizational structures and processes need to change to cohere with the changes in roles.
• How agile is implemented at the mission level

Figure 4 ("Evolution of Mission Capabilities via SOA") is interesting...it traces the structure of the stack from the 1960's mainframe to today's virtualized service-based open system. It shows how standardization/interoperability has moved up the stack, resulting in finer-grained plug-n-play functions that allow solutions that are more agile/adaptable. But, it seems to be mostly techno-centric...there's no discussion of how this affects the way IT, data, decision making, decision makers, and groups are woven together to complete a task or mission.

Incremental Transformation?

This statement from Chapter 1 of "The Great Transformation" (Scott D. Anthony) caught my eye last week:

"Transformation comes from entering new markets and leaving old ones. Companies rarely transform themselves through cost cutting or improved operational effectiveness....in almost all cases, operational effectiveness is insufficient to stave off disruption and drive long-term competitive advantage."

In a footnote Anthony references a 2008 book by Steven Spear ("Chasing the Rabbit") that looks at some exceptions (Toyota, Southwest Airlines, Alcoa) that are characterized by a "learning culture."

I don't quite know what to think about the exceptions...is their synthesis of Exploitation and Exploration possible (at least in part), for example, because their business/capital/knowledge/etc. ecosystems are at the lower end of the complexity and/or rate-of-change spectrum? Or is this kind of incremental longevity possible in any ecosystem? Speculation is probably all that's possible...a large range of intertwined factors, many of which are impossible to analyze with any rigor, are likely responsible. About the only observation I feel comfortable making is that recipes are a sure path to failure.

Constraining Disruption

Disruptive innovation is often thought of as an unexpected and significant shift in how an offering addresses a job need. By that definition, it's unclear whether the MIT picture taking example John Sviokla discusses in "Getting Started in Disruptive Business Design" is really disruptive...it's not clear exactly what job/market is being served by a $150 high altititude picture-taking capability.

However, each of the four ideas he discusses shows how important constraints are in sparking disruptive innovation. One key constraint is the importance of focusing on a radical improvement in only *one* area (e.g., cost in the MIT example).

The MIT example also, perhaps inadvertently, illustrates two other aspects of disruptive innovation...one positive and one negative.

First, very expensive IT-intensive infrastructure is becoming cheap, pervasive, and interoperable. The MIT students leveraged the cell phone and GPS systems to quickly and inexpensively cobble together a basic capability.

And, as one of the students admits (in the Boston Globe article), most of their engineering classmates are unimpressed...what they did was just not complex enough. This may be the most significant barrier to disruptive innovation---engineers who prefer "cool engineering" over "cool capabilities."

Sensemaking - The Toyota Way

The Toyota Way and Toyota Production System are often thought of in terms of process-orientation and continuous problem solving. A manager in the West tends to translate this into a system that is more about analysis than sensemaking. This article in the Financial Post highlights how just-in-time distributed decision making in the system is a complex mix of analysis and sensemaking...which may explain why it's been more successful in the East.

Matching Tools to Needs

We're in love with "magic incantations." Over the past hundred years or so, a popular version of the scientific method has been slapped on top of almost every area of knowledge. In this version, someone creates a model, gathers data to "prove" the model, and then promotes the model as a magic incantation that will solve all (or at least most) of your problems.

In the management domain especially, the fads come thick and fast. Usually, they have a valid use, but they rarely come with specific instructions of when/where to use them (and, maybe more importantly, where/when *not* to use them).

This is equally true with IT...SOA, wikis, Enterprise 2.0, etc., etc., etc. I've discussed this previously, but didn't have any suggestions beyond noting that a new tool will probably involve some exploring to determine where it does (and doesn't) fit. So, it's nice to see Ken Oestreich propose a framework for evaluating whether an application belongs in a cloud.

CapGemini's Take on Exploration

CapGemini is promoting a framework that emphasizes how IT is enabling exploration.

I like the direction this points to, but it feels a bit on the fluffy side. However, it may provide some structure in a conversation with folks who are unfamiliar with recent developments in IT.

Streaming Sensemaking

John Jonas has about the only techno-centric/analytical take on Knowledge Management I've seen that seems to avoid the traps of (a) thick/clunky semantic-trending-toward-AI technology, and (b) large complex models. Not surprisingly, it focuses on (a) context, and (b) streaming processing of event data.

This presentation needs the voice track to fully communicate (especially his description of how these concepts informed the fraud detection systems he built for casinos), but it clearly shows the challenge posed by the hyperconnectivity-driven growth in data and decisions. Regardless of where this particular technology goes, Jonas has some of the key nouns right (context, sensemaking, etc.).

Complex Enterprise Architecture

Formal enterprise architecting frameworks tend to emerge in a systems engineering context. As a result, they have an analytical bent that is often ineffective at the enterprise scale, especially as IT continues to dis-integrate into finer-grained composable chunks (e.g., services).

This post by Dion Hinchcliffe reflects a growing realization that enterprise-level architecting must be capable of operating in a Complex (Cynefin) domain....where the key issues are better addressed by orchestrating boundaries and attractors than by sophisticated analysis. Key issues mentioned include COIs, adaptable policy/governance frameworks, and edge-driven capabilities.

Exactly where this is headed is unclear. Experienced architects understand that the enterprise is a mixture of the Complex, Complicated, and Simple, and that our tools for effectively engaging the Complex are immature and few. And, as the boundary between IT and decisions becomes more fractal, the fundamental impedance mismatch between the analog world of human decision making and the binary world of IT becomes increasingly problematic.

Exploration as a Core Competency

More on the shift toward IT-catalyzed experimentation from the Sloan Management Review. It's mostly an up-beat take, with hints of COIs, governance, security, and the challenge of integrating the enterprise's Executing Core with its Exploring Edge.

When Is "Good Enough" Good Enough?

Wired has a nice article ("The Good Enough Revolution") discussing various technologies that are technically a step backward, but functionally a step forward (e.g., MP3, the Flip video camera, the Predator UAV).

As the article points out, the shared key success factor is accessibility. The traditional engineering definition of a better mousetrap tends to focus on how well it catches mice, not on how easy/fun it is to configure, deploy, and empty.

For IT-intensive technologies, Moore's Law and standardization is allowing engineers to shift the focus of the design effort away from technical considerations and toward usability. Whether engineers are equipped to exploit this opportunity is unclear...most have little training (and often little interest) in the much softer challenge of understanding the social/cognitive use context.

So, when is "good enough" good enough? There's probably no general answer/heuristic...but the question may help engineers be more aggressive in pursuing minimalist/bare-bones solutions that revolutionize a use context with commodity technologies, instead of pursuing revolutionary technologies for commodity use contexts.

NCW Needs Cynefin

In the Summer 2009 issue of "Parameters" (U.S. Army War College quarterly), Gautam Mukunda and William Troy draw some interesting parallels between the recent worldwide financial crisis and Network Centric Warfare ("Caught in the Net: Lessons from the Financial Crisis for a Networked Future").

I blogged about a similar link between the financial crisis and SOA last fall (Distrust 2.0), focusing on the lack of transparency and the difficulty of maintaining trust when entities are built to be combined/linked in unanticipated ways.

Though the authors don't quite put it this way, much of the article seems to highlight the risks of using Analytical (i.e., Complicated/Knowable domain of Cynefin) tools to cope with a Complex domain. However, the authors lack the concepts/language to do more than make some helpful observations. Putting those observations into a Complex-Complicated framework would help highlight why NCW is more about sensemaking and "managing attractors within boundaries" (Snowden) than it is about analysis, model building, and simulation.

To be fair, the proponents of NCW who the authors criticize are the ones who are dangerously confused in not clearly distinguishing the Complex from Complicated. However,the authors' critique would be more rigorous if it was grounded in a framework like Cynefin.

Organizational Shifting

As an engineer, I'm familiar with the fascination the IT world has with emerging decentralizing technologies like SOA and cloud computing.

And, I know how foreign the world of organizational theory, structures, and behavior is to most IT folks...but IT leaders are going to have to become literate in this area if they're going to be effective in matching the new IT to business needs.

The basic reason is that all network technologies trigger massive non-technological shifts. This was true for railroads, telecommunications, and electricity.

John Hagel, John Seely Brown, and Lang Davison recently began a blog entitled "The Big Shift" to discuss how IT is triggering a fundamental shift in organizations from push-oriented to pull-oriented. They've written about the topic before (e.g., "From Push to Pull" in 2005), but the blog entries are perhaps more accessible.

Is push-pull a more catalyzing contrast than exploit-explore? than complex-complicated?

That's probably the wrong question...more relevant for systems engineers and architects is do you understand how these concepts highlight key aspects of this shift?

The Power of Asymmetric Knowledge

This recent post by Joshua Porter at Bokardo comparing a symmetric social network (Facebook) with one that enables asymmetric social networks (Twitter) is worth reading.

Though the authors don't say it quite this way, my snap reaction was:

Asymmetric relationships are more powerful because they enable much more agile and adaptable in matching information to a decision context.

Symmetric relationships require a much richer shared context...which allows for deeper exploration of a specific context, but is much less agile/adaptable.

Porter links to James Governor's Monkchips blog on the topic which states "Asymmetric Follow is a core pattern of Web 2.0"...which sounds about right.

Why Services Are Hard...

...or at least one reason why... :-)

I ran across a quote from "Who Says Elephants Can't Dance?", Lou Gerstner's book about how he changed IBM:

“I came to see, in my time at IBM, that culture isn't just one aspect of the game—it is the game. In the end, an organization is nothing more than the collective capacity of its people to create value....I have a theory about how culture emerges and evolves in large institutions: Successful institutions almost always develop strong cultures that reinforce those elements that make the institution great. They reflect the environment from which they emerged. When that environment shifts, it is very hard for the culture to change. In fact, it becomes an enormous impediment to the institution’s ability to adapt.”

I suppose this has always been true, but was less noticeable when organizations were smaller and the world changed more slowly.

Today, most large organizations are fundamentally complex transaction-oriented information ecosystems, or they specialize in the design, development, or maintenance of systems within such ecosystems.

Service orientation (not technology per se) transforms a transaction-oriented information ecosystem into a hybrid that both (a) maintains its transaction processing capabilities and characteristics, and (b) exposes services, enables the relatively unstructured bottom-up/middle-out wiring together of services, and creates innovative new information transformation capabilities (along with new management capabilities).

The culture that produces transaction-processing excellence tends to focus on such qualities as predictability, reliability, and lack of deviance from a standard, and tends toward a bureaucratic organization with well-defined roles, rights, responsibilities, and processes to create and maintain that excellence.

Such a culture is in most ways poorly suited to thrive in a service-oriented ecosystem with its dynamic, emergent, and relatively unstructured exploration of potential information transformation capabilities...many of which are also potentially disruptive.

If you work in a large organization, you may have seen a "deer in the headlights" response to this new "dance" challenge, which seems to be nearly universal these days for elephants since large organizations today are usually information-intensive.

If it's hard for the elephants, it seems even harder for the elephant keepers...who don't know how to do anything but design, develop, and maintain elephants.

And, what makes this especially difficult is that it seems to be as much about turning elephants into composable "ropes, spears, fans, etc." as it is about getting the elephant to move learn new steps.

Seems like more than a minor identity crisis is emerging...

Self-Creation and Knowledge

For the past couple of decades, the term "autopoiesis" has been popular in certain scientific and organizational circles. Its literal meaning ("self-creation") hints at the possibility of some sort of emergent behavior that results in "information for free." This diagram from emergentfool.com is typical.

From a philosophical perspective the term is potentially problematic since it seems to imply that something can emerge from nothing...that there are uncaused effects (or that the assertion that all effects have causes is false). A part of the brilliance of Cynefin is that it largely dodges the ontological issue and focuses on the epistemological one.

Regardless, autopoiesis has been used as a concept to explore how organizations know and how they translate knowledge into action. "Exploring the Foundations of Organisational Knowledge" (Vines, Hall,and Naismith) is a recent paper that takes this approach.

They have a nice discussion of Karl Popper's "three worlds", and apply evolutionary theory to KM in a metaphorical fashion...though I was a bit confused by the conflation of evolution's "random mutation" with an organization's individual intelligent agents engaged in purposeful activity.

As a rough metaphor, there's a lot I like about highlighting the emergent aspect of knowledge. However, I'm not sure how useful it is when it comes to specific decisions and actions. Having slogged through Stuart Kauffman's "Origins of Order" a decade ago, I understand the attraction. And, I'm intrigued by ongoing laboratory efforts reproduce the origin of life and the emergence of complex subsystems in living organisms.

Agent-based approaches that decentralize attractor/constraint-oriented behavior clearly have value. But I wonder whether we've hit "peak emergence" (ala "peak oil") in our understanding of this phenomena...we don't seem to be much closer to a good understanding of the underlying causes of emergent information and behavior than we were before the topic became a staple of popular science 20+ years ago.

Whether we have or not, it seems inevitable that the rate of new emergent behavior will continue to accelerate for the foreseeable future as connectivity and interoperability continue to increase at warp speed.

Side Trails from Empirical Knowing

A couple of references from "Imperial Secrets" I found interesting:

• Persistent Surveillance and Its Implications for the Common Operating Picture" - Kelley's critique in "Imperial Secrets" was of such statements as "Once achieved, persistent ISR coverage will, in theory, deny the adversary sanctuary, enabling coherent decision making and action with reduced risk." I'm reminded of Dave Snowden's highlighting the fact that we filter ~95% of the raw data that hits our senses, and that the sense we make of what gets through our filters is shaped by a variety of factors, including which narrative fragments have been recently activated. For a nice summary of the limits/biases of individual sensemaking, see "The Psychology of Intelligence Analysis." A Panopticon-style capability would dramatically improve decision making in some ways...but it would also shift attention to the limits imposed by basic cognitive and social constraints on translating data into action, and, perhaps more important, would have significant and unpredictable effects on the activities of those under surveillance. Overall, it would seem that the net result is a more Complex space where prediction becomes more, not less, difficult.


• Critical Thinking and Intelligence Analysis - Another NDIC publication that thoughtfully considers the art and science of transforming data into decisions. I especially liked Table 7 (Analysis: Past, Present, and Future), which highlights various aspects of the shift from in-depth slice/dice analysis to agile sensemaking.

The challenge of knowing irregular, ambiguous, and rapidly changing threats has resulted in an emphasis on capabilities that are more exploratory/agile than exploitative/structured. A couple of articles in Joint Forces Quarterly (Q1 2009) highlight two aspects to this trend. Both reflect a maturing understanding of the limits of exploratory and exploitative capabilities.

• "Hybrid Warfare and Challenges" (Frank Hoffman, pp. 34-39) - "hybrid warfare"is the term associated with what seems to be the next phase of an ongoing transformation to make the U.S. military more agile. It emphasizes the need to operate in multiple "modes" simultaneously to effectively engage an opponent using conventional, irregular, and terrorist tactics in a coordinated fashion. Hoffman acknowledges the difficulty of such an endeavor...one that is similar in some ways to that discussed by Cash, Earl, and Morrison in the Nov 2008 HBR, which I discussed here.
• "Systems versus Classical Approach to Warfare" (Milan Vego, pp. 40-48) is a good summary of the concerns that have been raised about NCW, EBO, and SOD. Attempts to apply Complicated (expert, analytical) tools to a Complicated domain have, unsurprisingly, provoked reactions ranging from skepticism to hostility among many students of warfare. If you're unfamiliar with recent critiques of Systems of Systems Analysis, Systems Thinking, and other mechanistic approaches to warfare, you might find this article interesting. The bottom line is "Uncertainty in war is not only a result of a lack of information, but is also often caused by what one does not comprehend in a given situation." Analytical approaches (crafted for Complicated situations) may actually degrade the ability to comprehend in Complex situations.

Finally, John Arquilla has been writing about network/swarming tactics for years. "The Coming Swarm" is a recent summary of the current situation...it reminded me of Israel's approach to dealing with swarm terrorism.

Empirical Knowing

A recent book by Patrick Kelley explores how empires know. Entitled "Imperial Secrets: Remapping The Mind of Empire", it's a fascinating and non-traditional exploration of the challenges and limitations empires face in translating data into action.

Although most organizations are not empires, many dominate a specific context in ways that create the same kind of challenges in knowing. So, it's worth at least a look if you're a student of organizational behavior.

Kelley focuses on three empires facing challenges in knowing: Rome in the first and second century AD, Ottoman Turkey, and Britain in India. He draws lessons from specific narratives, including:

• As an empire comes to dominate its time and place, it struggles to detect signals that are weak or don't fit the frames that made the empire successful. An additional complication is that an empire's dominant presence warps the information space it inhabits.


• Informal networks are often critical and unrecognized in filling in an empire's blind spots.


• Empires are occasionally surprised when actions it perceives as being ordinary provoke an extraordinary response.

Although much of this will be familiar to students of epistemological approaches to organizational sensemaking (e.g., Karl Weick), the narratives provide a framework for a nuanced explication of the concepts. And, Kelley is a thoughtful analyst. Here's a few excerpts I liked:

• "I wish to contest the unidirectional read of how 'knowing' works; i.e., that an observing subject gains ever more knowledge of some given object and consequently, power over the latter accrues to the former."
• "I propose that the Panopticon actually operates in reverse in the imperial context....empires will nearly always operate at an information deficit in relation to their subjects....The overseer is there for everyone to observe—what the empire believes, does, wants, and will do is laid out in imperial media, legal codes and judicial decisions, the conduct of its agents, and the architectural and scientific 'performances' of power—all in contrast to the 'inscrutable oriental' who resists observation physically, linguistically, and epistemologically in his capacity rather than inability to 'imagine that....' "
• "...imperial policy makers faced the problems of 'slippery' knowledge—data points collected, analyzed and presented in a context different from where they originally resided by virtue of imperial process."
• "Empire consists of a certain set of patterns that characterize the meaning and function of physical reality—Empire is information, which may almost sound like a definition, but my intent is rhetorical emphasis. Consequently, the management of information—disseminating the preferred patterns, and identifying competing patterns for elimination—becomes the core function of imperial administration."
• "The risk of drawing virtual lines and imposing order on the grand confusion of reality is not simply that our lines may be imperfect, our approximation inexact. Rather, these virtual creations can become so comfortable and accessible that we come to inhabit them as real, producing 'knowledge' about things and places that don’t actually exist. Reality as such then enters into the 'species of alterity'; it literally becomes something foreign and other—opaque, if not irrelevant, to our attempts to understand it."
• "... the strategic information most fundamental to a successful imperial order—information about beliefs, identity, authority and allegiance—does not appear as some 'golden nugget' at the end of an operational or tactical tasking order. Rather, it circulates along networks of exchange, akin to an economy in which information serves as currency."
• "The unique intelligence problem of empire, in contrast, is precisely that such distinctions [foreign-domestic] do not exist—or rather that they exist in over-abundance, with inside/outside fractures splintering and overlapping ad infinitum within and between spaces, communities and individuals under the imperial umbrella."
• "If absolute power corrupts absolutely, it also tends to isolate completely—twin tendencies any executive authority risks as it ascends to the heights of imperial power. Bureaucracies rise in tandem with that isolation, providing the intellectual equivalent of walls and gates; but subverting that intellectual structure by act of will can prove nearly as impossible as escaping from the physical walls for reasons of either status or security. "
• "Modern strategic discourse is full of discussions about decision cycles and getting inside an opponent’s OODA loop (Orient, Observe, Decide, Act). Less well explored is how to get inside or engage another’s time, as such, which may be moving at a different rate and in a different direction."
• "A key aspect of the British performance, and an enduring problem of imperial intelligence, I suggest, is not to correctly 'know' foreign minds, but to read backward, and accurately perceive how those foreign minds come to understand the apparently familiar and domestic."
• "As I have argued throughout this text, imperial intelligence is less a problem of determining the truth or falsehood of specific facts, and more an issue of negotiating how truth is constituted and what 'knowing' means. These issues leap to the fore in the contest over education, which is fundamentally a fight to frame how meaning can and will be constituted in both the past and future."
• "None of these [narrative] products would suitably respond to any conceivable state information request, and certainly would not fit into the almost infinitely replicated 'intelligence cycle' model, but they might well answer information requirements. The rhizomatic nature of information, particularly in an imperial environment, indicates that there will nearly always be connections and contexts not immediately evident to any decision-maker posing specific questions....Better still, they are quite obvious in what they leave out, and in this are less deceptive than traditional maps."
• "Th e language and image of empire is universally visible and available to its nominal subjects, while the reverse is not usually the case."
• "...the rise of imperial power creates the demand for a whole host of new government functions and institutions, which cannot be created, except with great difficulty and foresight, ex nihilo. Far easier, and more common, is to simply adapt existing institutions to new roles..."
• "Modern Americans have come to believe that the norms and values encapsulated in their form of government and their ways of conducting foreign relations are the birthright and open options for men everywhere. In accordance with this persuasion there simply can be no 'others.' " [Adda Bozeman quote]
• "...new information—in order to make sense—must fit into a narrative, a history, and this history amplifies the problem of the other/same dichotomy. 'As distinct from the present, the past is alien, exotic, or strange; as continuous with it, this past is familiar, recognizable and potentially fully knowable.' [Hayden White quote] "
• "Empires thrive despite, rather than because of, their information institutions."
• "In early days, say just after you have sacked Carthage or defeated the Nawab of Bengal, you might know very little, but you also know what you don’t know, i.e. basically everything. This makes the 'known unknowns' a very large category, but also makes the more dangerous 'unknown unknowns' very small. As imperial experience progresses, as administrators learn the languages and jurists incorporate local legal principles, the 'known known' flourishes in a rhizomatic way...For as 'known unknown' shrinks with every census and cartographic survey, the 'unknown unknown' consequently expands proportionately. While the analogy may be inexact, imperial power becomes more expert over time, and consequently more subject to the patterning and heuristic biases associated with the 'Paradox of Expertise.' "
• "My general assessment has been that empires are always at an information deficit—telling more than they hear—and the deficit over time becomes associated with a lost capacity to listen."
• "A perspective which presumes the role of an objective machine loses twice over by deliberately abjuring to assimilate the specifically human factors that only humans can collect, while making inevitable bias more difficult to detect in bureaucratically neutral discourse."

As someone who is more critical realist than postmodern, I think Kelley occasionally overstates his case. But, given the growing importance of Complex contexts and the growing number of "virtual empires", he may be right to do so.

Complex Knowing

A few miscellaneous items from the past couple of weeks that highlight the limits to and evolution of analytical approaches to decision making and decision support:

• "When Knowledge Management Hurts" - This short article by London Business School professor Freek Vermeulen discusses a study asserting that incorporating formal knowledge into decision making can actually result in worse decisions. I don't have time to discuss the actual study (link in article), but it's worth looking at if you are a KM practitioner.
• Unintended consequences are inevitable when services are made mashable. This slashdot posting discusses Google having to shut down free SMS message after a popular iPhone app overwhelmed Google's servers.
• "The Rise of The Social Network System" - a short article on social hyperconnectivity...it's about linking contexts. I'm always slightly surprised by articles like this in light of how much science fiction has explored this area in the past few years. Also, Tim O'Reilly's discussion of the article.
• Lots has been written recently about the decline of newsprint. Clay Shirky has a nice overview of the rise and decline of this particular asynchronous hardcopy Few-to-Many media channel. It's nice to see some historical perspective in something as ephemeral as a blog post.
• Finally, Deloitte has posted a study of mining safety in South Africa that was done using Cynefin. If you've found this framework a bit abstract, this study might help you better understand one way it can be used.

Those Who Can, Do...

Those who can't, scribble?

That was my initial reaction to some recent discussions of 90-9-1 as it applies to Jive. Whether my reaction makes much sense is unclear, since "doing" and "scribbling/teaching/creating/contributing" are not either-or categories. And, I'm struck at how fundamental language/conversation/writing is to the creative process. So, I'm not implying that speaking/writing is necessarily wasteful...it's just that I get the feeling that separating signal from noise is getting exponentially more difficult.

Regardless, much of the discussion of 90-9-1 is about increasing the percentage of those who contribute (vs. lurk), and not much is said about increasing the signal-to-noise ratio.

In the context of enterprise social media this distribution makes me wonder...when should we start worrying a little more about quality and a little less about quantity?...do our metrics emphasize quantity simply that's what's easy to measure?...or perhaps the emphasis on participation is simply because enterprise social media is in an orientation phase where the key challenge is getting folks to participate because they'll never "get it" until they do?

Is this the sort of space where you have to increase the noise to a tipping point before you can begin to detect really valuable signal?....or before you get emergent value?...I'm not sure anyone really knows.

Decision Contexts and Flow

This TED video is a nice example of how the "intersection of decisions and technology" is becoming ever more fractal...and why sensemaking becoming a central concern in the creation of innovative information technology.

Whether you design IT, or are simply a user, you're increasingly going to be struck by how truly innovative IT morphs itself to a decision context in a way that fits the user's decision flow...and you'll begin to notice just how much traditional IT interrupts that flow.

Overloading CASE

In the engineering arena, we think of software when we see the acronym CASE.

Barry E. White (MITRE) is using it as a combination of Complex Adaptive Systems (CAS) and Systems Engineering...Complex Adaptive Systems Engineering. I had not seen any of his work when I last visited this topic.

This presentation and paper are a recent discussion of the topic. In the presentation, I like White's highlighting of the importance of non-technical factors, and I especially like the Enterprise Systems Engineering and Systems Engineering profilers (slides 20, 21). I'm a bit puzzled by the inclusion of an Enneagram Web...its origin in what appears to be mysticism seems to clash with the rest of the presentation. The paper strikes a more serious tone...its discussion of the limits of control and the importance of context are applicable to any large scale architecting/design effort.

Regardless, it's nice to see a CAS-oriented discussion of systems engineering.

Backlash Against Process Standardization?

Joseph Hall and Eric Johnson have an article entitled "When Should a Process be Art, Not Science?" in the March 2009 Harvard Business Review. In it they discuss the fact that some contexts are too variable to allow standardized processes to be effective.

Although there are echoes of Tushman's Exploitation-Exploration contrast and Cynefin's Complicated-Complex contrast here, Hall and Johnson emphasize the need for "artistic processes" when the following condition exists:

• A variable environment (inputs, process, and/or desired outputs) - at some point, the costs of standardized processes outweigh the benefits. Costs may be real (i.e., a very complex "scientific" process that's expensive to design and maintain) and/or opportunity (i.e., the "scientific" process is either too slow and inflexible to address key customer needs).
  And, the authors rightly emphasize the fact that most real world contexts have artistic aspects and scientific aspects. Where analysis and best practices are effective they should be used (science); where they are not, a probing approach that uses qualitative metrics should be used (art).

Since the article's purpose is not to examine the sources of a focus on standardization, the authors don't have much to say about why the need to balance science and art is becoming more pressing. Here's a few snap reactions I had...I suspect you can think of a few more if you're interested in systems and their structure. These are reasons why there tends to be a bias toward "scientific processes."

• We like predictability...the more the better (up to a point, see "Flow: The Psychology of Optimal Experience" for more detail). In my more cynical moments, I wonder if American management writing might be best summarized as a continual search for the Holy Grail of "The Magical Incantation" that allows any unthinking human to mindlessly create unending value. I'm sure the authors generally don't feel that way (after all, they are innovative thinkers almost by definition), but the mass readership's appetite for the latest concept make me wonder.
• As the lower layers of various capabilities become standardized (everything from double entry accounting to gun parts to automobile designs to the IP stack), dramatically more complex capabilities are enabled...which (at least initially) have a large amount of variability as users probe the new solution space that such standardization opens up. Eventually, aspects of that solution space become standardized (sometimes de facto, sometimes de jure), and the cycle starts over. For example, in the IT arena, the emerging emphasis in services is due in large part to the increasing standardization of technologies that used to be largely artistic (trends like Moore's Law are significant drivers in enabling this standardization). And, as the standardization-complex-standardization cycle time compresses, the importance of balancing the two increases.
• Since processes attempt to interact with the world in a predictable way to achieve a desired goal, they instantiate/reflect a model of that world. Although traditional modeling may contain non-deterministic elements (e.g., events may have an associated probability distribution, etc.), the modeling mindset is primarily quantitative, deterministic, and "scientific." So, processes tend to be scientific.
• The Scientific Revolution, the subsequent Technological Revolution, and the Quality Revolution have combined, for the first time in world history, to create amazing amounts of excess wealth. Routinization of a process enables first the use of lower skill labor, then automation, of that process, resulting in lower costs for a specific capability. Ford's River Rouge complex is what I always think of. The coming worldwide demographic bust will create even more pressure for moving artistic work into scientific space.\

If you're in an environment that seems relentlessly focused on standardizing processes, this article is a nice summary of why some balance may be due.

User-based Blending

The traditional approach to modeling is for modeling experts to observe, analyze, synthesize, etc. to produce a model of a real world decision space. This model is then used by a non-expert model user to explore that decision space.

Occasionally, a modeling tool emerges that allows non-experts to create fairly sophisticated models without expert intervention....the one I always think of is electronic spreadsheets, which were the initial killer app for the PC.

A potentially interesting development in user-based, non-expert, modeling tools is described in a recent TED presention...Siftables from David Merrill of the MIT Media Lab. The basic idea raises interesting possibilities and questions. For example, each Siftable could be instrumented to monitor how it is used, and log and report that data to a central repository which could then be mined by experts for insight into how to modify the existing Siftables' methods, display, etc. Or, users and experts could use Siftables as a collaborative design artifact to explore how to define a decision space. Or, Siftables could be used as a distributed workflow data collection tool by distributing them to individuals in a value chain/net with instructions on how to model decisions/workflows in that chain/net....sort of a physical wikipedia entry creation activity.

This sort of thing has been enabled in a virtual way with multi-touch capabilities like Microsoft's Surface. However, it may be that Surface is too configurable to efficiently address some modeling needs...the physical constraints imposed by Siftables may actually help catalyze more effective exploration in some situations.

Maybe this goes nowhere, but it seems like an innovative new way to explore a decision space by blending the physical and information domains...it's that distinctive blending of the two that I find most fascinating.

Shifting Identity

In a world where ideas, values, and beliefs constantly swirl, individual identity arguably is more fluid than it was, say, 100 years ago. Whether the same is true of organizations is unclear; they often struggle with establishing an identity, much less working to establish a new one.

A new HBS Working Paper ("Technology , Identity, and Inertia through the lens of 'The Digital Photography Company'", Mary Tripsas, August 2008) sheds some helpful light on this topic by discussing the place of identity in organizational sensemaking in times of transition.

When it comes to defining a company's strategic direction, those of us with a business background tend to think of frameworks that are more ontologically oriented (e.g., SWOT) than those that are epistemologically-oriented (e.g., Cynefin). And, unless we have an interest in sociology or anthropology, we probably never think of identity. The closest we get is terms like "culture", but they tend to lack the sense of purpose that comes with identity.

Here's a few of the paper's key points that are consistent with what I've observed in the workplace over the past few years.

• Tripsas notes that identity is both internal and external. It's not just how individuals and groups see themselves within the company, but how customers, suppliers, competitors, and financial analysts see the company.
• External identity often serves as a "forcing function". One aspect of this is that organizational identity is coupled to its business context...if key aspect(s) change enough, it forces an identity crisis. And, external identity helps solidify an emerging internal identity.
• When identity must change, a period of ambiguity is important in establishing a new identity. This is clearly Complex space.
• "Established firms have particular difficulty adapting to ... change that requires the acquisition of fundamentally new knowledge and routines. .... Some routines such as TQM that are meant to improve a firm's efficiency in a core business crowd out more exploratory initiatives, limiting the firm's ability to take advantage of novel opportunities." I suspect this is especially true in industries (e.g., automotive, defense) that have placed a lot of emphasis on improving exploitation.
• "As the core essence of the organization, identity directs and constrains action. The routines, procedures, information filters, capabilities, knowledge base, and beliefs of an organization all reflect its identity." This is perhaps the key point...since identity tends to be perceived as ground (not figure), it tends to shape more than be shaped.

As I've said before, the topic of organizational identity is probably underappreciated. This was perhaps understandable 30-40 years ago. It's much less so today as the effects of hyperconnectivity and hypercollaboration begin to fundamentally restructure the business landscape.

For a more conceptual discussion of organizational identity, see Kurtz & Snowden's "Bramble Bushes in a Thicket" at http://www.cognitive-edge.com/ceresources/articles/52_Bramble_Bushes_in_a_Thicket.pdf

Business Case ROI

In this article, Susan Case asked if business cases are a waste of time. Her answer was "yes", but do them anyway.

Although she didn't quite put it this way, what I heard her saying was that opportunities are often a mixture of Simple, Complicated, Complex and Chaotic (Cynefin taxonomy). We tend to limit business cases to the analytic (i.e., Complicated) domain.

However, most opportunities will have Complex aspects that must be addressed with techniques that are more exploratory than exploitative. Susan clearly understands this and helps highlight the limits of analysis in achieving a positive ROI when pursuing an opportunity.

"Complex" Maturity Models?

In the systems engineering domain, I'm seeing increasing interest in complexity. There seems to be a growing realization that IT (e.g., SOA) is pushing all technology toward a greater intertwining of individuals/groups and technology.

This is a disruptive shift. When tools are largely decoupled from exploration/sensemaking, the designers, implementers, and maintainers of those tools can largely ignore how individuals and groups explore a context to make sense of it. When tools begin to be woven into individual/group exploratory work, they can no longer ignore sensemaking concepts.

It's unclear whether there's much recognition of this need in the mainline systems engineering community. A few recent observations:

• "Modeling and Analysis of Interoperability Risk in Systems of Systems Environments" - This article in the Nov 2008 issue of CrossTalk is one of the few I've seen that seems to understand that interoperability has a significant cognitive component. The basic representations of interoperability reminded me of DSMs...not exactly the same, but a similar concept.
• "Systems Thinking Comes of Age" - This was the editorial by Pat Hale in the Dec 2008 INCOSE Insight magazine. Since you have to be a member INCOSE to read this, I'll refer you to this presention by Pat Hale at a recent conference.
• "SOA Contract Maturity Model" - This article by Kjell-Sverre Jerijærvi provides an approach that seems a bit more appropriate for Complex contexts than the CMMI-centric SOA maturity models that seem to be more at home in the Simple & Complicated domains.
• "Cloud Maturity Models Don't Make Sense" - This post by Roger Smith is one of the few I've seen questioning the limits of maturity models. Since most engineers have little understanding of the Complex domain, I suppose it's understandable that few consider the limits of analytical approaches.

A final note: I'm seeing more interest in Systems Thinking, System Dynamics, etc. ala Senge, Forrester, Beer, Ashby, Weiner, et.al. All engineers should have a good understanding of the tools in this area, since they're part of a good foundation for modeling and simulation. However, my impression is that these tools are more Complicated than Complex. For the truly Complex, you're probably going to need a different approach (e..g, Cynefin).

For those who love taxonomies, here's my favorite diagram of where Systems Thinking fits in the overall scheme of thought...from the International Institute for General Systems Studies. I discussed this topic previously on this blog here. I suspect it won't be the last time.

Against Collaboration?

I suppose there are several reasons why collaboration has become more popular over the past decade.

First, there's a long term cultural/philosophical shift toward radical egalitarianism that is uncomfortable with any distinction among individuals (a topic which is beyond the scope of this blog). Many members of my generation find it a bit disorienting to encounter young people who were raised in a sub-culture where everyone always got a trophy. This perspective emphasizes collaboration over individual initiative (not that the two are necessarily in conflict, but they're not exactly orthogonal).

Associated with this trend is a backlash against the "lone gun" hero stereotype who has a firm grasp on Moral Truth and The One Right Way. And, there's the simple fact that we're creating larger and more complex capabilities...which is only possible when large numbers of individuals sacrifice their interests and desires (to some degree) to achieve a group goal.

In the organizational theory domain, we've seen the rise of Wegner's CoPs and DoD's CoIs, along with a slew of other "communities of" TLAs that attempt to increase the effectiveness of organizational collaboration.

Perhaps the primary reason for the explosion of interest in collaboration recently is the rapid growth of social media capabilities on the Internet. This growth has highlighted one of the key challenges of group action: the formulation and governance of a web/hierarchy of shared purpose.

Since group agility slows exponentially as group size increases, collaboration is a two-edged sword in Complex contexts. The diversity of understandings (of causes, effects, and shared purpose) provided by a group will, on average, probably make sense of a Complex context faster than an individual. On the other hand, if the situation requires a lot of fast exploration of intertwined alternatives, it's not clear that a group will necessarily outperform an individual, especially when purpose, cause, and effect form tangled hierarchies.

Those of us who've worked in group settings recognize there's also a danger of synthesizing several understandings into a mediocre whole. And, the classic danger is the group being hijacked by a strong personality.

As with all such swings, the swing toward collaboration runs the risk of going to an extreme where highly capable individuals are not effectively used.

These observations were triggered by a recent post by Oliver Marks that mentions a 2006 article by David Freedman entitled "What's Next: The Idiocy of Crowds."

As our ability to collaborate increases, we may bump into basic cognitive and social limitations more frequently and violently...it's not clear that all such limitations necessarily point toward more Complicated collaboration tools, processes, or frameworks...they may point instead toward Complex tools, processes, and frameworks.

Innovation and CAS-Driven Organizations

The increasingly Complex (even Chaotic) nature of information-intensive contexts is highlighting the relevance of Complex Adaptive Systems (CAS) theory for organizations. As I've said repeatedly, Cynefin is the best framework I've run across for thinking about the implications of this change.

Most technical and business people have little understanding of CAS, so papers that describe how CAS theory applies to organizations can provide a quick introduction. "Complexity-Based Agile Enterprises", by Dyer & Eriksen, has just been published by Cornell's Center for Advanced Human Resource Studies. If you're looking for an introduction to CAS from a traditional management perspective, you might find this worth reading.

I have a few misgivings. The paper seems to assume that a context is either ordered (Complicated/Simple) or Complex. Although a specific context may be dominated by the Complex (or Complicated, etc.), most contexts have aspects of multiple domains (Chaotic, Complex, Complicated, Simple is the Cynefin taxonomy) that allow the decision maker to move the context from one domain to another. That's one of the strengths of Cynefin...it defines strategies and tactics for doing this. This paper seems to emphasize a linear movement (Explore, Exploit, Adapt, Exit), where Cynefin emphasizes deliberate movement among domains in a direction dictated by current needs.

The paper does a nice job of describing some of the ways in which a primarily Complex context has different staffing and organizational needs. And, they emphasize our knowledge in this area is relatively immature.

Anyway, if you're academically-oriented and interested in how CAS theory applies to organizing, check it out.

Can Collaboration be Monetized?

Lots of very smart people have thrown millions (billions?) of dollars at social media with very disappointing financial results. I have no idea whether/how social media can discover a distincitive business model that breaks new ground and lays a robust financial foundation.

This post is an interesting perspective on the question.

Cross-Silo Exploitation/Exploration

IT-driven hyperconnectivity continues to increase the need for and challenge of creating cross-silo capabilities. James Cash, Michael Earl, and Robert Morrison present an interesting proposal in their November 2008 Harvard Business Review article entitled "Teaming Up to Crack Innovation & Enterprise Integration."

They propose two small/agile enterprise-wide groups engage in cross-silo Exploration and Exploitation; called the Distributed Innovation Group (DIG) and the Enterprise Integration Group (EIG) respectively. Each group is more of a catalyst than a traditional matrix. The authors note that "businesses are better at stifling innovation than capitalizing on it," and "better at optimizing local operations than integrating them for the good of the enterprise and its customers."

Since innovation is distributed across the business, the DIG focuses on "fostering and channeling" cross-silo innovation rather than being responsible for it. And, the EIG prioritizes and provides resources for cross-silo horizontal integration projects.

The DIG:
- "Scouts for new ideas and untapped potential in current technologies"
- "Scans…for emerging technologies"
- "Facilitates participation in idea forums"
- "Acts as a center for innovation expertise"
- "Publicizes [and incubates] promising innovations"

The EIG:
- "Manages the corporate portfolio of integration initiatives"
- "Serves as the corporate center of excellence in process management and improvement"
- "Provides staff to major business integration initiatives"
- "Is responsible for enterprise architecture"
- "Anticipates how operations might work in a more integrated fashion" \

The article also discusses what the groups are not, and explores in some detail what capabilities are needed to form and deploy them. Integration work is discussed in more detail than Innovation…perhaps because it is more amenable to traditional management techniques.

Six sets of skills are identified as central to Integration:
- Familiar with business process design/improvement
- Experience with cross-functional systems implementation
- Competence in architecture analysis
- Expertise in information management
- Experience in program management
- A talent for relationship management

They note that the last skill is rare, but essential to "open people’s eyes to the possibilities and benefits of horizontal integration and to enlist their commitment to making integration happen."
Finally, they note that the DIG and EIG groups are similar:
- "Each is a collection of catalysts"
- "Neither offer direct solutions"
- "Both lead through relationships, communication, and targeted expertise"
- "Networking is an essential activity for both"
- "Both focus on adding customer value"
- "Most members must be trilingual…in the languages of business, IT, and sociability"
- "They must know the organization"

The details proposed in this article seem reasonable and implementable, though I suspect that more innovation may arise from the EIG than the authors seem to anticipate.

My guess is that the most difficult challenge will be getting these groups institutionalized as effectively carrying out their duties. Strong top management support and staffing by top-notch personnel who are already carrying out these functions would seem to be essential if they are to have any chance of success. I suspect many large organizations will find it difficult to pay the opportunity cost required to make these groups successful.