Levinthal escreve de uma forma muito atractiva.
"The question of “where do good strategies come from” is arguably a subset of the broader question of why things are the way they are. At its most basic level, as with respect to the question of life on planet earth, we have two general classes of answers: one invoking some form of intentional design and the other invoking a Darwinian process of descent with variation and a contested selection process. The argument developed here tries to set forth a middle-ground perspective of a Mendelian executive. This “Mendelian executive” operates with a degree of intentionality; but, unlike the conception of rationality in neoclassical economics, this intentionality is limited. Furthermore, the emphasis is more on the design of experimental processes than on the design of specific paths forward. [Moi ici: Desenhar caminhos específicos é demasiada arrogância. É menos inseguro fazer pequenas expriências e ir tacteando] While circumscribed, this intentionality and design sets this Mendelian executive apart from a pure Darwinian process. Both orientations highlight the role of path dependencies. However, the intentionality of the Mendelian executive allows for the conscious exploration of adjacent “spaces” rather than the happenstance of random variants. Furthermore, the argument developed here highlights the role of intentionality with respect to the selection and culling of strategic initiatives. The firm is viewed as operating an “artificial selection” environment in contrast to selection as the direct consequence of the outcome of competitive processes. [Moi ici: Tanto a paisagem competitiva evolui e afecta as organizações, como estas podem procurar espaços competitivos alternativos] While these outcomes may inform the artificial selection process, the two criteria need not be the same.Continua.
...
While not controlling these processes, perhaps to some degree we can engage in some intentional engineering of these evolutionary processes. Consider, for instance, the frequently discussed need for organizations to balance processes of exploration and exploitation. Such manipulation is unlikely to be a precise engineering of the evolutionary process, but rather a reflection of a broad awareness of the effect of alternative organizational policies on the dynamics of firm evolution, as these policies tilt the “needle” of the exploration/exploitation balance in one direction or another.
...
An important dichotomy in arguments regarding the specification of business strategies is the degree to which action is identified a priori as a result of “offline” reasoning and consideration versus an ex post assessment of “online” trials. In this regard, it is import to note that a Darwinian process is one extreme of online trials. The Mendelian executive straddles both online and offline forms of learning and strategy identification. The experimental variants are not random acts but rather conscious choices of potentially promising initiatives. Furthermore, the Mendelian view melds an offline sensibility with regard to the ex post evaluation of online trials. A purely evolutionary or Darwinian mechanism operates through a contestation of relative fitness and what is fundamentally a myopic selection process. By contrast, the Mendelian executive may make conscious choices of what might constitute the dimensions of merit by which initiatives are evaluated. Furthermore, that calculus of evaluation may constitute forward-looking indicators of merit. The ultimate value of the contemporaneous realization of any initiative is inevitably speculative, but speculation is conscious and explicit. From an evolutionary perspective, there are two basic mechanisms at work: processes of variation and selection. We can consider the work of the Mendelian executive with respect to both of these processes. We tend to associate variation as the consequence of some stochastic process, but variation may have considerable elements of planning and intentionality. [Moi ici: Variação ou variedade?] Variation may be considered from the perspective of individual ideation; the internal ecology of initiatives within the firm; and the role of path dependence, “pivots,” and adjacencies.
...
Yet, in the Mendelian perspective developed here—situated between classic images of design and intentionality and a Darwinian process of random variants and contested selection—the interpretation of experimental outcomes plays a prominent role. It is common to contrast learning processes with processes of selection. Our typical conceptions of learning processes are based on notions of reinforcement learning in which the proclivity to engage in a particular act is reinforced or diminished as a consequence of the association between that act and an observed outcome. However, in the strategy context, outcomes in terms of an ultimate consequence of a strategic action are typically rather distant in time from the “act.” Given this “distance,” the time between initiating a new product, mode of competing, entry into a new geography or market space, and the ultimate feedback regarding the wisdom of such an effort, interim judgments play a critical role in any adaptive efforts. [Moi ici: O papel dos indicadores do balanced scorecard como forma de avaliar se a execução da estratégia está a ocorrer como previsto e se os resultados estão a caminhar para o desejado]
...
when organizations define milestones and metrics around key success factors, they are constituting an artificial selection environment that guides the cultivation of initiatives within the firm. The virtue of such shaping rewards is that they may substantially speed up the feedback process relative to the feedback from the environment as to whether a given action or strategy is pushing forward along a promising track."
