Ecologies of Learning II
First Part
Topic 1: Coevolution
What is the essential idea of coevolution?
Organizations and their environments interact and change each other. Coevolution suggests that all interactions between populations of organizations result from direct and indirect feedback. An organization that stimulates the evolution of another organization is, in turn, itself responsive to that evolution, and the response is predictable.
In direct coevolution, two populations evolve in response to each other. In diffuse coevolution, one or more populations evolve in response to several other organizations in a community. In some situations there may be standard evolution as well. Coevolution implies nonlinear feedback among interacting populations.
Why is it important?
The idea suggests that changes in populations of organizations can have unexpected and non-intuitive effects in other populations of organization via feedback loops, especially in complex situations. Variables become co-dependent on each other. It also implies that one must consider the overall system and effects of direct interactions and feedback loops to understand and predict organizational behavior and change.
What would a co-evolution theory of search over a terrain of endogenously specified opportunities look like?
In a coevolution theory all opportunities are interrelated. New alternatives would promote the creation or "founding" of other similar alternatives until there are too many to evaluation when evaluation pressures will reduce the growth of alternatives. The exploitation of one family of alternatives will have an effect on the growth of other alternatives. There are feedback loops present that may cause peculiar interactions between evolving alternatives.
For example, using Technology A may create lots of alternatives but their creation may inhibit the evaluation of Technology B's alternatives.
What difference would it make?
Success might increase one set of alternatives and decrease another set, which may or not be optimal. Focusing on one set of alternative may cause you to ignore another more lucrative set. Reducing search among one set of alternatives may unknowingly reduce search among another set.
What is the link between the idea of coevolution and comparative statics (or functionalist) perspectives on long-run learning outcomes?
Maybe both ideas focus on directionality and outcome rather than try to measure the exact interactions. They don't try to build explicit causality, just try to model the effects of interactions.
Another explanation is that in some situations the long-term effects are stable (in coevolution when the self-effects are higher than the competitive effects). When competitive effects are predominant, the systems can be highly unstable and have no equilibrium.
Topic 2: Competition, armaments races, and the "red queen" effect
In a competitive situation the "environment" of each learner consists (at least in part) of competitors who are also learning, in part by extracting knowledge learned by others. One common observation about such situations is that a sustainable competitive advantage cannot exist. Comment.
In a competitive environment each learner is coevolving. Like the "red queen" effect, mutually learning competitors in an environment suggests that one must sustain a level of continual improvement equal to competitors to sustain a competitive advantage. This implies a careful balance between exploitation and exploration (slack) that would be extremely difficult to maintain. Also the technological richness or maturity of the environment will also affect an organization's ability to stay far ahead of competitors. Absorptive capacity arguments also suggest that the more exploration and learning one does the more spillover and thus learning that occurs by other organizations.
A sustainable competitive advantage can exist if other constraints (e.g., legal monopolies, etc.) are in place or if one company has a sizeable advantage in capital.
A second common observation is that in such situations otherwise effective expenditures of effort and money are made ineffective by the matching expenditures of competitors. Comment.
Once again rising expenditures are the price of learning and the price of staying "with the pack" in a competitive situation. One can debate whether they were effective or ineffective, but all will agree they are necessary. The US/USSR arms race is a good example -- the USSR bankrupted themselves trying to match US defense spending over the decades.
If competitors can sometimes pool their expenditures toward common evaluations of nascent technologies (like they do in Japan), all competitors can thus increase the effectiveness of their expenditures. Sometimes fostering competition can be a beneficial strategy to an organization.
Students of institutions are interested in the legitimacy of organizational practices and forms. One thing that affects legitimacy positively is the number of organizations using the practice or form. What are the implications of such a coevolutionary perspective with respect to legitimacy?
Coevolution theory suggests that while having more organizations practice the form will increase legitimacy and thus increase the rate of form adoption, there may be feedback effects on itself (self-effect) or on other organizations. For example, the increasing rate of adoption of JIT by auto manufacturers had a dramatic effect on the organizational forms of auto suppliers they had to adopt a JIT system to remain legitimate as well. Both systems thus co-evolved.
Coevolutionary theory would also suggest that the increased adoption of one form will impact the adoption of other forms as well. The Olinkowski paper showed that adoption of the Lotus Notes system caused positive feedback toward adoption of a new work system as well. Technology and organizational structure both co-evolved.
What would happen if you combined such an idea with the idea that there is competition among practices and forms, so the more organizations using a particular practice or form, the less the effectiveness of it?
This notion means that as more firms adopt the new practice or form, the competitive advantage to adopt it becomes less and less. Eventually there may be no competitive advantage to adopt the form, so that the "self-effect" feedback causes the system of organizations to reach a stable adoption % less than 100%.
Firms may then adopt another new practice or form, which may prove to be more effective and thus begin the cycle again.
Topic 4: Situated Change
What is the essential idea of situated change?
Change is not always planned but is the result of evolving situations and on-going practices of individuals. Organizational transformation is an on-going improvisation enacted by organizational actors trying to make sense of and act coherantly in the world. Organizations are not stable but are constantly changing and "improvising".
How does that conception differ from other ideas about organizational change?
Planned change models assume managers deliberately initiate and implement changes (force field analysis, contingency theori, innovation theories,). It is criticized as a discrete event separate from ongoing processes or organizing, and too much emphasis on manager's rationality.
Technological imperitive models assume that technology is the main driver, and managers have little imput. Adoption of technologies creates predicatable changes in organizations. This model undermines proactive change ideas. The model is also incompatible with the customization available by end-users.
Punctuated equilibrium models assume change is rapid, episodic, and radical. Situated change sees incremental changes that occur gradually.
How is it related to what has been called here an ecological conception of change?
Situated change is an evolutionary change. An environmental disturbance (variant) causes slow, emergent adaptation. New practices are tried (variants) and those practices best suited to the new environment are adopted and become dominant. New practices are constantly replacing less efficient practices as the entire organization incrementally changes.
How is the idea exemplified by Orlikowski's Zeta Corporation?
At Zeta the introduction of Lotus Notes technology in the Customer Support Department set off a series of emergent changes in support practices that eventually changed the entire organization and reward structure. Knowledge began to be captured in the database and evolved into a vital source of training and information to specialists. It fostered collaborative behavior among specialists and improved customer service immensely.
Five main changes were:
* shift to electronic capture of support knowledge
* redistribution of work from individual to shared responsibility
* emergence of proactive collaboration among specialists
* expansion to global support
* controlling access of information to outsiders
The Notes technology didn't cause the change but enabled it. It would probably evolve differently in a different organization or situation.
To what extent and in what ways might a situated changed perspective provide the micro level foundations for a more macro theory or organizational change?
I believe it shows that all revolutions are made up of small evolutions in time. It shows what happens at the micro level can have significant effects at the macro level. This perspective acknowledges pro-activeness of actors in a macro world.
Explain Figure 6 in the paper on exploration and exploitation. What is the model?
In this model, the competitive consequences of learning depend on learning by others. It considers ways in which competitive advantage is affected by the accumulation of knowledge.
Assume an organizations realized performance is a draw from a distribution of mean X and variability V. But the relative change in performance vs past is often less important than the resulting relative position vs other learning organizations.
In this case, assume that an organization competes against N other organizations, each with performance average X and variance V. The chance an organization will become the best is 1/N+1. P* is the probability that the (x',v') organization will outperform all the rest.
What are x' and v''?
X' is the mean and V' the variance of a reference organization. There is a competitive advantage if P(x',v') is greater than 1/N+1. (chance of primacy is greater than mere chance).
What are the results?
For N=1, any increase in x is a competitive advantage. Learning always pays off regardless of variance.
Figure 6 is a plot of x' and v' assuming all other competitors have x=0 and v=1. The lines (for N-100, 10, 1) are the points where P(x'v') = 1/(N+1)^2 = probability of primacy by mere chance.
For N>1 increases in either the mean or the variance improve competitive advantage, and sufficiently large increases in either can offset decreases in the other. The contribution of the variance increases as N becomes larger.
What is the intuitive idea that they illustrate?
It shows that if learning reduces variability any gains in knowledge may actually reduce your chance for primacy, especially in situations of many competitors.
What is the effect of learning on performance distributions?
If learning improves both mean and variance, it improves competitive advantage, especially in environments with many competitors. However if learning increases performance but reduces variability, competitive advantage may actually decrease.
What the consequences in competition for primacy? What about competition for avoiding finishing last?
Increasing variability improves the chances of finishing first. But it also increases the chances of finishing last as well.
What are the strategic issues?
In technology, an alternative strong enough to overcome the effects of unfamiliarity will have competitive advantages. New technologies may have higher performance but increase variability in the initial adoption period . This can be good or bad.
Thus learning that decreases variability may help prevent finishing last but may also prevent you from finishing first as well. It depends on how learning affects the two tails of the performance distribution. The strategic question becomes whether you can do exceptionally well without leaving the confines of conventional action.
Organizations with lower performance but who can change their variance can get advantage by increasing variance, provoking a "right-hand tail race".
Social welfare issues?
I guess the models suggest that sometimes too much knowledge of a situation can reduce the chance of making a big improvement. Sometimes ignorance and the resulting increase in variance can be a good thing. In social welfare issues, knowledge of existing technologies may reduce the chance you will try something radical (and potentially useful).