Learning from Expert Innovators

I recently attended a presentation given by Harvard Business School professor Rob Austin as part of Harvard's Computer Science Colloquium Series.  Professor Austin was reviewing his continued collaboration with Swarthmore College's professor of theater emeritus, Lee Devin, on the processes, principles, and practices of innovation shared by business and artist innovators.  This work is the extension of the work in their co-authored text, Artful Making: What Managers Need to Know About How Artists Work. 

Defining "innovation" broadly as "activity that leads to the creation of valuable novelty", the researchers look for commonalities in the processes, management principles, and practices across a range of fields in which innovation is considered important.  This includes the arts, design, entertainment, media, and scientific research.  As professor Austin summarizes the results: "Our methodology is inductive and case-based. Our objective is to advance general theoretical understanding of work and management approaches that repeatedly generate valuable novelty, and to relate the emerging theory to theoretical constructs in adjacent areas of research, such as business innovation, the psychology of individual creativity, and the social psychology of creativity. We find numerous categories of commonality that transcend fields of application."

He first provided the context of where he feels their study fits in the landscape of adjacent research.  He feels it cuts across the research done by his peers at HBS -- that of innovation and product development (Clay Christensen), organizational learning (Amy Edmondson), knowledge management (Dorothy Leonard), and the social psychology of creativity (Teresa Amabile).  He takes care to state that while the psychology of individual creativity focuses on inherent talent, and the social psychology of creativity focuses on the context (work space, incentive systems, etc.) that might produce creativity, their study is more about the process -- what are the learned/evolved ways of working that lead to more innovative outcomes?

Professor Austin showed video segments edited together to showcase the powerful idea of symmetry in process across a diverse range of practitioners.  The similarity in both language and concept was striking. 

• Agile software developers and artists both shunned the idea of working from any kind of pre-ordained, up-front "spec". 
• The idea of prototyping was prevalent (software prototypes, paper-and-pencil sketches) with a recognition by all parties that it allowed experimentation with a low cost of starting over or changing direction of an idea didn't work.
• While general management axioms belittle working without a proper plan ("If you don't know where you're going, any map will do."), to these innovators the idea of unintended benefits was welcomed -- "I didn't really know where I was going at first, but I'm happy with the result."
• Perhaps sentiment toward prototyping and unintended benefits also fed their attitudes toward failure.  They saw it as a necessary step in the creative process.  One painter remarked that nothing was lost but the paint underneath the finished painting.  And not even time spent on a failed outcome was seen as wasted since it led to insights that will affect future endeavors.
• A lack of compromise was palpable.  "We don't compromise on the beer, we don't compromise on the label" was one artist's claim when recounting how marketing executives tried to get him to change the label he designed for a new brand of beer.
• The idea of endless market research to be sure to design "for the customer" was nonexistent, revealing a process much more complex than "the customer is always right."  The lead designer for Bang & Olufsen (a $600M firm that develops top-end electronic products -- also see them referenced in the blog entry on Simplicity) stated it most eloquently, declaring that they don't do market research at all because they don't believe the customer knows what s/he wants.  If you ask them what they want, they'll tell you to build something they already envision.  The  trick is to "come up with something they didn't know they needed but can't resist."

But merely reviewing these commonalities is not the extent of their research. In Artful Making, Austin and Devin showed that the iterative process and workflow structures in the agile software development movement had much in common with the process of rehearsal inherent in the theater environment.  From this they followed an inductive, case-based research approach to induce generalities from instances.  These common generalities include the following ideas (over-simplified in this recounting) which form the basis for an emerging theoretical structure to their research:

• The "costs of making" include reconfiguration and exploration ("iteration costs") as well as setup costs.  As the ratio of iteration costs to setup costs increases, the "cost of novelty" increases. Simulation and other visualization technologies would be one attempt at reducing the overhead associated with these costs to encourage exploration.
• The "benefits from making" include those derived from differentiation (the novelty of the outcome) as well as those derived from the core functionality of the outcome (regardless of the novelty).  As the ratio of novelty benefits to core benefits increases, the demand for novelty increases.  Are customers willing to pay for a differentiated outcome?

This matrix helps convey their framework:

                                               

	LOW     Cost of Iteration	HIGH     Cost of Iteration
HIGH     Demand for Novelty	Artful Making     (Artists, some Business)	Craft
LOW     Demand for Novelty	Not Stable	Industrial Making     (Not just business, some artists)

The "Industrial Making" quadrant is then those sets of circumstances -- high cost of iteration, low demand for novelty -- that might best warrant development using traditional, sequential workflow processes.  The Industrial Revolution might then be seen as an era where the demand for novelty decreased -- people were willing to accept tradeoffs like less differentiated but cheaper products. 

The Post-Industrial era might present different options --  low cost of iteration, high demand for novelty --  that likewise recommend a more iterative, agile or "artistic" approach to development.  The authors do not claim that one method is necessarily better -- sometimes an interative structure produces better results, whereas sometimes a sequential structure might be preferable.

So why did agile software development processes resemble those of theatrical rehearsals?  On the supply side, the materials involved in making (software, people) are pliable and have a low cost of reconfigurement, whereas the demand side places a high value on the novelty of the outcome (innovation).  But once a play needs to be "operationalized" -- as happens, for instance, when the Phantom of the Opera tours to a foreign country and the company is charged with the faithful execution that honors the quality and brand characteristics of the original -- then perhaps a more Industrial Making or sequential approach is preferred.

This idea that determinants or factors related to the supply/demand of novelty (making costs; willingness to pay for novelty) lead to process/structure decisions (iterative vs. sequential) then can be assessed for the management implications that result: what type of process to use, attitudes toward failure and feedback, knowledge management implications, examination of how motivation affects the process, etc.