by admin on July 17, 2006 06:00pm
Hank just blogged about critical thinking. If I had to state my own concise definition of what lies at the heart of critical thinking, it would be a personal commitment to finding the right solution to any problem, regardless of whether or not figuring it out and the subsequent implications are easy or comfortable (in practice, this usually means being the resident skeptic right at the point everyone else is getting excited.) This is not necessarily a comfortable thing to do: we tend to have a psychological affinity for propositions that confirm or reinforce, rather than challenge, our existing beliefs. And there are many sources of social and institutional pressure that militate against “naysayers” and “people who make things more complicated than they need to be.” Moreover, no matter how many tools the hardcore skeptic has in his or her toolbox—years of accumulated experience and wisdom, technical savvy, statistics and operations research skills, or sharp psychological intuition—the context for bringing those tools to bear is seldom ideal.
As an example, consider a team trying to lock on a project plan under some challenging time constraints: success means leaving the room with (1) agreement on a plan; (2) shared conviction that the plan can achieve the team’s goals, if every contributor executes against their commitments; and (3) an accurate understanding of the probability of achieving the goals. From experience, you probably know (1) and (2) are difficult enough—now think about being the one person in the room who says “you know, if you bracket our point estimates for milestones with sensitivity limits and run a Monte Carlo, I think our point-based-looks-like-99%-chance-of-success looks a lot more like 20% chance of success…”. (I use this example deliberately, thinking of the Carnegie-Mellon Software Engineering Institute (SEI)’s Capability Maturity Model Index (CMMI) —“level 2” project management entails items (1) and (2) from my example—the kind of quantitative management becomes institutionalized at “level 4.” Whether or not you are a CMMI fan, it does provide a benchmark indicating how few organizations would find our data-driven critical thinker’s suggestion easy, comfortable, or routine.)
On Port 25 we’ve had discussion of some hard problems: about the tough choice between MS building product capabilities versus partners and ISVs and about licensing and shared source project requirements. A big reason Port25 exists is because our point-of-view in the lab is that the best possible answers to these and other questions are unlikely to be easy or comfortable . (To use some extreme touchpoints, in my view the position that MS could answer these questions optimally by discounting the phenomenon of open source development and its history is just as almost certainly incorrect as the position that MS should answer these questions optimally by discounting the phenomenon of commercialized software development and its history in favor of “opening everything.”) If this is true, a dialogue among diverse perspectives is essential to continuously push the thinking of everyone involved away from the personally easy or comfortable—hence, the Port25 dialogue. There’s a new empirical study that I hope drives this point home and offers you the same motivation to continue to read and post to Port25 that it offers me.
Kevin Boudreau at MIT’s Sloan School of Management took an empirical approach to the question “Does Opening a Platform Generate More Innovation?” Cleverly, he looks at handheld computers (PDAs)—an area with multiple software and hardware players—and does a deep analysis of innovation measures in relation to “openness” measures. (What’s also clever, IMO, is that he includes different types of openness (ranging from licensing to SDK’s and documentation)—and innovation (differentiating between lots of incremental deltas and big breakthroughs)). What he finds overall is:
… setting an optimal open strategy may be a far more nuanced problem when trying to promote innovation, in contrast to managing the traditional trade-off between promoting adoption (by opening) versus retaining appropriability (by closing). In this traditional perspective, intermediate levels of openness might have been understood as a means of achieving some middle ground between two relatively simply opposing interests. In the case of opening to promote innovation, an intermediate level of openness might in fact be in the best interest of promoting innovation; there may not be so severe a trade-off with maintaining appropriability. However, the decision to open a little or a lot, and precisely how, will also likely involve trade-offs across multiple dimensions of innovation. (p. 25)
“Openness” was non-monotonic in relation to innovation, meaning (depending on the particular constructs used in different analyses) the curve peaked and then declined at some point. And the type of openness appeared to promote different types of innovation—lots of incremental or imitative innovations as opposed to a few breakthroughs:
Openness therefore should not only affect the rate of technical change, but also its direction. Therefore, these findings offer another explanation of why we observe so few “perfectly” open strategies in practice and why there might plausibly be place for heterogeneity of open strategies, insofar as there is space for heterogenous innovations and differentiation in a product market.” (p. 26)
Why is this exciting? Because it provides compelling empirical data from one technology domain that the question of optimal “openness” is an empirical question, not an ideological one. And that “openness” in relation to a traditional software business model is not a zero-sum, oppositional game. That means not only empirical research but also the Port25 dialogue are essential—even determinative. The bottom line: what we discuss here matters. If you download the paper, let me know your reaction. – Bryan