Innovation: The Paradox of Constraints

If you've worked in a group on a creative task, you may have noticed how constraints can spark innovation.

This article from the WSJ (Chains That Set Us Free) discusses some of the experimental evidence.

And, this article on groupthink from The New Yorker is a nice summary of the history of the technique of brainstorming (thanks for John Ganter for bringing it to my attention), along with a refreshing discussion of why the lack of constraints in brainstorming seem to make it a relatively ineffective approach to innovation.

There seems to be a bit of a paradox here: many formal innovation activities that focus on the removal of boundaries to get folks "outside of their box" result only in shallow thinking.

This is consistent with my experience where such exercises usually result in bland lowest-common-denominator ideas with little new value.  What seems to help is the creation of new "boxes" that help highlight and break existing (often unseen) boxes.  Here's a few boundaries I've seen that helped unearth creative ideas:

• Identify key resource constraints in the domain under discourse and play with scenarios where they become either 10x more constrained or 10x less constrained.  This highlights the constraint structure of the problem/solution landscape in a way that can help identify radically different solutions.  IT-intensive domains are especially susceptible to this approach; even though we understand Moore's Law at a cognitive level, it's usually very unclear how that affects a solution space.
• Keep the group size to five or less.  If you don't, the more assertive members will tend to define and structure a discourse landscape that's largely inside the boxes they know.  If you keep the group small with diverse subject matter experts (see the next bullet), "ritual dissent" tends to emerge organically.
• Ensure that group members are subject matter experts in a different, but related, areas ... the goal is to force discussions about non-obvious, but key, underlying assumptions.  The tendency toward lowest-common-denominator solutions means that each member should be at or above the level of innovation you're trying to achieve.
• Don't use a formal innovation process ... recipes generally produce minor variations of existing ideas instead of something new.  Using Complicated approaches in a largely Complex domain is a common mistake.
• Incorporate multiple cycles of focused discussions, unfocused discussions, focused thinking alone, unfocused thinking alone.  Ideally, scatter these over a period of weeks to months.  Ideas need time to percolate. 

Hope that's not too much of a recipe ... :-)   Dave Snowden's "management of the emergence of coherence within attractors within boundaries" remains for me the best succinct description of any type of exploratory activity in a largely Complex context.

Social Media, Emergence, and Telos

MIT's Sloan Management Review recently published an interview with Anthony Bradley and Mark McDonald about their new book "The Social Organization: How to Use Social Media to Tap the Collective Genius of Your Customers and Employees."

I was encouraged to see them emphasize the need for Purpose in the creation of effective social media-based communities.  There's a lot of money being thrown at corporate social media with a "build it and they will come" / "emergence magic" mindset.

Social media with no coherence, boundaries, and attractors is Chaos.  Internet-scale chaos works for socialization purposes where individuals form their own informal communities.  But, even the largest companies will probably find the intersection of "bottom-up" communities and the company's key exploratory activities relatively small.

Within a company, purposefully working to "manage the emergence of beneficial coherence within boundaries, within attractors" would seem to be an essential partial constraint on social media.

Seems like we prefer mindless solutions: an algorithm at one extreme or "emergence magic" at the other.  Managing complexity (and complicatedness) requires intention and mindfulness ... drones need not apply.

Identity as a Learning Disability

Two recent articles on identity's effect on learning caught my eye.  Both seem to lean toward the possibility that Identity inhibits learning when it focuses more on what I am (we are) than on what I do (we do) ... ontology vs. action/telos.

• "Why Do Some People Learn Faster?" - Wired blog post discussing a new study showing that academic accomplishment is increased by praising effort and inhibited by praising ability.  My reaction: seems like more fallout from a century-old societal shift  from Christian skepticism about fallen humanity to secular optimism about human perfectability.  For more, see Freud , John Dewey, the birth of modern progressivism.  The therapeutic mindset, so strange barely a 100 years ago, is now so pervasive as to be invisible.
• "Can Everyone Be Smart At Everything?" - KQED covering much of the same ground.

This is one of those topics that quickly becomes a lightening rod for all kinds of noise. I suspect it's  not productive to allow culture war issues associated with identity to detract from the task at hand.  As someone who loves to create stories (and despite my antipathy towards pragmatism as a worldview), I liked this quote from a recent Peggy Noonan column:

"Here's the problem: There is no story.  At the end of the day, there is only reality.  Things work or they don't.  When they work, people notice, and say it."

Data Priori Computing

A year ago I discussed the possibility of an emerging frontier of computing ... "smart machine, smart human".  Instead of me sitting in/on a machine's hard-coded data processing loop, the machine uses a mesh of sensors and processors to look over my shoulder and anticipate what information I need to maker better decisions.

It's a "machine on my loop" instead of "me in/on a machine's loop".

Walt Mossberg's WSJ review of a new browser add-on called Digital Folio made me wonder whether smart computing will be more about data than function.  What makes Digital Folio compelling is that it's tightly focused on one type of data (price) within a specific context (shopping), and that the add-on has been designed to smartly and unobtrusively weave "just enough" additional data into the user's sensemaking work.

At a more general level, I'm beginning to wonder if the primary "smart machine" design challenge is to make it fun for the novice user to "play" with them.  And, to make it unobtrusively slip into the user's normal sensemaking flow.
If that's the case, then what makes a smart machine fun/playful?  Here's some snap reactions on a design approach:

• Identify the one/few data items that are the "pivot points" of a decision/sensemaking context.  For Digital Folio, these include the product and its price, along with the context of shopping for electronics equipment.
• Identify what makes these data item(s) most intriguing within the context ... specifically, determine how the core data item values/relationships shift between ambiguity and clarity.  For Digital Folio, price is the key data item.
• Finally, create a compelling "in-the-flow" user interface so that a novice user can "play with / probe" the data and its connections within a specific sensemaking context.

The design work order ....

• Identify key data within a specific sensemaking context
• Design a UI that  (a) makes key data/connections compelling, (b) allows the user to immediately see interesting possibilities in the data/connections, and (c) enables the user to easily explore the data/connections
• Define the required functionality

... reverses the typical machine design process:

• Define function(s) that automate one or more tasks
• Create a user interface for a specific user community-need
• Design the data schema

Since the user's key sensemaking/decision data precedes everything, I thought the term Data Priori Computing captured the overall concept which:

• focuses on sensemaking-centric data (traditional computing focuses on machine-function-centric data)
• leverages the user's contextual flow of sensemaking data to create a framework for the emergence of presentation and function, possibly guided by the user's interaction with the data
• is a context-centric bundle of data-presentation-function where the data is a tightly focused nexus around which all sensemaking and decision making revolves (it is not BI or data mining)

BTW, when I googled "Data Priori" the only place I saw this term used in a similar way was "The Global Data Palette: Massive Databases and the Reformation of Content Creation in Film/Video and Music/Sound Art Practice" ... data-priori movie-making.

Deloitte Report on Social SW in Business

Deloitte has released an interesting report on the ongoing effort to discover how social SW can create real value in an enterprise.

It's a helpful reminder that social SW in business requires much more than "build it and they will come", since a business's stakeholder community has neither the size nor the scope of the Internet.

And, it emphasizes the need to focus on outputs (eg, value-added) instead of inputs (eg, adoption rates).

There are several challenges discussed:

1. Employees must gain some basic familiarity with the tool. While some Internet-based knowledge will transfer, business-specific tools will have their own learning curve, even if they're based on common open source SW.

2. Employees must discover how the tool can help them do their job better. There's no recipe for this, and I suspect that existing roles & processes will have to remain largely intact during the initial adoption phase. One employee at Alcoa is quoted as saying "One day it just clicked for me. ... That's when I started seeing Traction help me do my job better."

3. There's a bit of an ontological mismatch between the formal teleological structures of a workplace (processes, roles, organizations, etc) and the unstructured (mostly) telos-free landscape of social SW. A "deer in the headlights" response is entirely appropriate. Add in employee workload and a relative low level of employee engagement, and you're looking at a daunting task to (a) find potential high-value, low-cost bridges between "as-is" and "to-be", and (b) build critical mass once promising bridges are identified. This mismatch seems to grow exponentially as the size & bureaucracy of the business grows.

None of this is Simple or Complicated; it's definitely "managing the emergence of beneficial coherence within attractors within boundaries" (Snowden).

Deloitte asserts that the need for these tools is becoming critical as work is increasingly characterized by activities with significant Complex (exploratory) aspects (page 7).

Figure 3 lists five "unique" capabilities of social SW:
1. Identify expertise
2. Facilitate cross-boundary communication & conversation
3. Preserve institutional memory
4. Harness distributed knowledge
5. Discover emerging opportunities

Unfortunately, there seems to be little, if any, awareness that these tools are unlikely to be used effectively for these purposes unless there's good social connectivity across the business. There's a "chicken-egg" problem here that folks like Dave Snowden have addressed with such activities as Social Network Stimulation. I'm not sure I see much awareness in this report of how critical this issue is.

The discussion of metrics (pp. 9-11) made me slightly uneasy (eg. "Recipe for Success"); an understanding that there's a risk that a metric will become a goal (thereby ceasing to be a metric) seems to be missing. A boundaries/attractors approach would seem to be more effective; maybe this was done for the OSIsoft example, but if so, it's not clear. It seems more of a "build it and they will come", though that may just be retrospective coherence.

Figure 10, a "Social software capability and tool heatmap", may be the most interesting part of the report. My initial reaction was that the ratings (eg, a Prediction Market is rated as Low for "Preserve institutional memory") might be a bit simplistic, but they're still interesting. I suppose you could try to envision scenarios where a rating is completely wrong if you find the table offensive.

Bottom line: the overall approach feels more Complicated than Complex; I kept thinking that a sensemaking approach (eg, construction of a Cynefin map) would be much more effective in identifying and exploring how to best use social SW in a business context. Although there are hints that the authors understand this (pp. 16-17), I would have thought that they (John Hagel and John Seely Brown) would have spent more time discussing Complex needs and approaches.

A Flat World is Noisy

Much (negative and positive) has been written about Thomas Friedman's 2005 discussion of globalization ("The World Is Flat"). A key enabler of globalization has been transportation, communications, and information technologies.

The current issue of the Sloan Management Review has an article discussing leadership in this context ("Flat World, Hard Boundaries - How To Lead Across Them", Ernst & Chrobot-Mason).

A few snap reactions (which are equally applicable to the use of social media inside an organization):

1. A flat world is noisy. Although the opportunity for amazing new "signal" emerges in a flat world, the risk of increased noise drowning it out is real. Brokers of various types used to manage this risk in the physical world. Increasingly connected and transparent individuals and organizations are almost bereft of tools to manage this risk in the virtual world.

2. Focus becomes much more important in a noisy world. At the same time, leaders are much more constrained in their ability to control & command focus. They are left with a largely complex task; in Dave Snowden's memorable phrase "managing the emergence of beneficial coherence within attractors within boundaries." At the same time, they are largely unprepared to either recognize or effectively engage complex contexts.

3. As a result, leaders must simultaneously address at two difficult and novel issues: increased noise, and managing complexity.

The six boundary-spanning practices the authors discuss are not inherently bad. I'd feel better, however, if they emphasized the need for these practices to be introduced in the context of very specific customer-driven tasks and goals. Without that focus and discipline I'm afraid the authors' advice could easily be warped into some sort of "happy-clappy" initiative where generic "do-good" activities do more to breed cynicism than success.

Misc. Sensemaking Items

• Figure 5 of this AF report has sensemaking woven all through it
• A human-intensive approach to defeating IED attacks shows promise
• Our understanding of how to weave the mindful and the mindless (i.e., unordered and ordered) is immature at best; see, for example, experiments in easing traffic-related "boundaries"
• "Self-evident computing"? ... something about "self-evident" grates, but the basic idea reflects the trend toward weaving technology and sensemaking
• Sensemaking always has a teleological aspect ... a good, if rambling reminder, from one social media commentator
• A key driver for sensemaking is the explosion of IT-based solution spaces and the associated need to constrain them enough for useful innovation to occur
• More evidence of just how utopian the dreams of singularity are ... and of the vast gulf between IT and neurological structures/processes
• As someone who has taught Statistics, I liked this discussion of the scientific method and statistical analysis. It highlights a basic fact: getting really good data is the hard part; the numbers are easy. And, despite a few exceptions, it shows that any statistical discussion of a complex system (and all biological systems are complex) is very vulnerable to mis-interpretation.