On predictive entrepreneurial action in uncertain, ill-structured conditions

Causation, effectuation, and enactment and its derivatives (see, for instance, Welter et al. 2016 on the interrelations between effectuation and creation and bricolage theories) have been among the processes discussed most often in the context of entrepreneurial decision-making. Therefore, we will concentrate on these three processes and investigate them in light of our framework of decision problem structures. To do so, it is necessary to first outline their respective features and to delineate them from one another, at least analytically, even though they might co-occur, intertwine, and overlap in actual human reasoning (Bhowmick 2015; Sarasvathy 2001).

In her seminal paper, Sarasvathy (2001) introduces a theory of effectuation and contrasts this decision-making process with its counterpart causation. Fundamentally, she distinguishes between the two based on the underlying logic of each of the processes. While she associates causation with prediction, she emphasizes the role of control in effectuation. Causation is linked to well-structured decision problems (Sarasvathy 2001) and, in short, rests on the idea that entrepreneurs choose “optimal” paths of action in the present which are believed to maximize possible future returns as predicted by the entrepreneur. The entrepreneur is hence aware of the desired end and selects the alternative from the option set that he or she believes will deliver that end best possible.

Effectuation has been argued to take the opposite stance (Bhowmick 2011; Dew et al. 2008; Dew and Sarasvathy 2002; Grégoire and Cherchem 2020; Read and Sarasvathy 2005; Schneider 2018; Tumidei et al. 2020) and has been juxtaposed with causation in different respects (Dew et al. 2009; Fisher 2012; Read and Sarasvathy 2005; Shirokova et al. 2020). Effectuation describes a process apt for facing an environment which is uncertain and ill-structured to a degree that does not make it possible to predict the future; alternatively, effectuation leans on the idea that to “the extent that we can control the future, we do not need to predict it” (Sarasvathy 2001: 252). Therefore, effectuation processes take a limited set of means as given (rather than the aspired effect as in causation processes) and—predetermining a particular level of affordable loss—focus on the various effects which can be created through control of the available means while making use of strategic alliances and exploiting contingencies over time. Therefore, the “distinguishing characteristic between causation and effectuation is in the set of choices: choosing between means to create a particular effect, versus choosing between many possible effects using a particular set of means. Whereas causation models consist of many-to-one mappings, effectuation models involve one-to-many mappings” (Sarasvathy 2001: 245). The concept of bricolage, first elucidated by Lévi-Strauss (1967)—and first applied to the field of entrepreneurship by Baker and Nelson (2005)—while differing in detail, largely resembles the notion of effectuation. It does so in particular because of its central tenet of “making do with what is at hand” (Baker and Nelson 2005: 329). Accordingly, it too starts from consideration of the constrained resources available to the entrepreneur. Given the strong ties between both bricolage and effectuation (Fisher 2012), we focus the following discussion expressly on effectuation; however, the findings derived are not limited to effectuation per se but apply similarly to related approaches too.

Enactment, which borrows from Weick’s sensemaking conception (Weick 1979, 1995; Weick et al. 2005), essentially describes the idea of non-directed action that is made sense of only after it took place (Bhowmick 2015). Therefore, enactment in some sense resembles effectuation and contrasts with causation (Sarasvathy 2001). Both effectuation and enactment rest on the premise that the dynamic environment is too uncertain to be predicted and, hence, does not fall into the scope of causal reasoning with its optimization analyses. Both effectuation and enactment emphasize the constructive feature of eventual outcomes rather than the predictive feature. Consequently, Bhowmick (2015: 519) states: “In a fundamental way, both concepts give primacy to action over pre-planning in explicating organizing and exploring.”

While enactment and effectuation are similar in some respects, they are also distinct from each other, in particular with regard to the directedness of action (Bhowmick 2015). While the enactment concept puts non-directed action before sensemaking, the effectuation logic implies that entrepreneurial action targets control. Moreover, the given set of means is central to the effectuation logic, while enactment gets along without explicit reference to particular means, thus leading Bhowmick (2015) to refer to effectuation as resource-dependent action and to enactment as resource-independent action.

In sum, both effectuation and enactment are designed to match uncertain conditions and preclude the making of meaningful predictions, whereas causation considers well-structured environments under risk and largely builds on predictions of future states of affairs. Figure 1 illustrates the inchoate range of processes that leaves room for a process of entrepreneurial decision-making capable of both incorporating uncertainty and building on prognoses.

Bhowmick (2015: 528) has argued that “under uncertainty, entrepreneurial actions have a non-causation basis” and thus classes causation as a tool for entrepreneurial action under risk rather than uncertainty. Similarly, causation has been directly linked to well-structured, risky decision problems (Sarasvathy 2001). We follow both Sarasvathy (2001) and Bhowmick (2015) in their classification of causation, and, thus, deviate from Packard et al. (2017) in this respect who characterize causal approaches as a means of transforming absolute uncertainty into creative uncertainty aiming to arrive at a manageable decision problem.

If causation falls into the scope of well-structured decision problems, then structural defects are meaningless and can neither be recognized nor addressed. Unlike causation, neither effectuation nor enactment have been subject to an investigation of whether the concepts recognize structural defects with which the decision problem they are designed to solve may be afflicted and, if so, how they try to address them. This is surprising, given that both effectuation and enactment are two ways of dealing with uncertain, and hence ill-structured, decision problems.

Effectuation rests on the idea that entrepreneurs start with a given set of means, particularly with who the entrepreneur is, what he or she knows, and whom he or she knows (Sarasvathy 2001, 2008). The entrepreneur then imagines potential effects that can be created with the available means and eventually selects one of those effects in light of his or her (contingent) aspirations based on what he or she considers an affordable amount of loss.

Effectuation implicitly recognizes at least three structural defects and incorporates particular ways of addressing them: First, the effectuator lacks a consistent goal to be targeted and hence suffers from a goal defect. Effectuation proposes the entrepreneur refers to general, contingent, and, thus, vague human aspirations as its reply to this dilemma. Second, the effectuator is quite unaware of how to appraise different imagined effects and consequently suffers from an appraisement defect. Effectuation, again, provisionally suggests a quite general and vague criterion for the appraisement and selection of one particular effect, namely affordable loss as its attempt to resolve the ill-structuredness. Third, while the entrepreneur’s set of means is taken for granted, the particular outcomes of available options are uncertain. Therefore, the effectuator is confronted with a causality defect. The “solution” to this issue as implied in effectuation is to imagine possible effects that can be created with the set of means and, therefore, to artificially close the initially open set of outcomes. Figure 2 illustrates effectuation’s relation to structural defects and its attempts to resolve them.

In sum, effectuation indeed implicitly recognizes various structural defects and, hence, the ill-structuredness of the decision problem it is designed to deal with and includes different ways to resolve the various defects. Therefore, effectuation can be understood as a means of partly reducing the perceived degree of ill-structuredness and, eventually, of arriving at a manageable decision problem through a particular degree of pre-structuring.

Like effectuation, enactment too is a process of entrepreneurial decision-making in the face of uncertainty and, hence, ill-structuredness. Unlike effectuation, however, enactment does not include any pre-structuring of the decision problem and instead emphasizes the initial non-directed action and the sensemaking of what was done only ex post. Therefore, even though decision problems suitable for enactment are afflicted with structural defects, enactment itself remains silent on them and, thus, does not address them. Rather than addressing the ill-structuredness, enactment simply bypasses its implications through action preceding thought (Weick 1979, 1995).

As outlined above, decision problems may either be well-structured or ill-structured, the latter to different degrees, depending on both the amount and magnitude of structural defects associated with the ill-structuredness. To this point we have discussed three existing decision-making processes and related them to either well-structuredness or ill-structuredness, respectively. We distinguished each of these processes based on whether or not it recognizes structural defects.

Moreover, they can be differentiated according to the role of prognoses in the processes. While effectuation at least implicitly recognizes various structural defects, it denies prognoses a key role. Enactment, in contrast, skips recognizing and addressing structural defects while similar to effectuation, prognoses are not central to the process either. Even though it is indeed true that both effectuation and enactment reject the logic of prediction on a fundamental level, the degree to which they do so varies. While it can be argued that every action is aimed at improving the state of affairs (von Mises 1949) and, hence, implies an expectation and, thus, a prediction of how the action can improve this state, enactment in some sense largely gets along without any reference to explicit prediction putting action before sensemaking.

Effectuation, however, while emphasizing the logic of control and differentiating itself from causation through rejecting the logic of prediction, still requires a certain degree of prediction and planning. It is, for instance, necessary when trying to imagine effects possibly creatable with the given means or when deciding on the level of affordable loss. That is because a particular loss can only be affordable in light of other obligations the effectuator faces in the future. Sarasvathy (2001: 260, emphasis added) seems to acknowledge the fact that the process she suggests requires some planning when she states, “effectuation does not involve elaborate planning.”

Causation bypasses structural defects simply because it is designed to deal with decision problems in well-structured situations where there are none, while relying heavily on prognoses which are key to causation.

Analyzing the range of existing decision-making processes in light of both the potential recognition of structural defects and the role they give to prognoses reveals that the range is incomplete. It lacks a process which recognizes the ill-structuredness of the decision problems entrepreneurs usually face yet puts prognoses center stage. Therefore, this paper outlines a process incorporating these attributes that we term execution. Table 1 summarizes the positioning of effectuation, enactment, causation, and execution with regard to the recognition of structural defects and the role of prognoses.

Objectivity is not a property of the structure of decision problems. While there are objective facts associated with each particular problem, there is no objective measure of whether or not it is ill-structured and, if so, to what degree. Characterizing a problem as ill-structured must necessarily result from an individual entrepreneur’s personal perception. Because entrepreneurs differ in their subjective perceptions of the environment and the general circumstances relevant to their actions (Karhu and Ritala 2020; Shepherd et al. 2015), perceptions about the structure of a particular decision problem will necessarily differ from one entrepreneur to another. Hence, the selection of a decision-making process can only be suitable in light of a particular perception of the underlying problem (Beckert and Bronk 2018). Neither process is generally superior to its alternatives, but its usefulness depends on the circumstances the entrepreneur faces and interprets subjectively (Gigerenzer 2008; Gigerenzer and Gaissmaier 2011; Sarasvathy 2001).

Nevertheless, decision-making processes designed to deal with uncertainty and ill-structuredness respectively, namely effectuation, enactment, and execution, differ fundamentally from each other with regard to how they suit different perceived decision contexts. They can particularly be distinguished by the degree of specificity of certain decision parameters, namely end(s), means, and the selection criterion.

Execution is the most specified process among the three mentioned. It specifically recognizes the role of ends, means, and the selection criterion and incorporates them into the process. Effectuation does so to a lesser degree: While it relies heavily on the means available to the effectuator, both ends (in the form of general human aspirations) and the selection criterion (affordable loss however determined) are quite vague. Enactment requires decision parameters to be even less specified. There is no clear place for either ends or a selection criterion and even if means must play some role (as one has to employ some means to do something) there is no direct reference to the role of means, which is expressed quite well in Bhowmick’s (2015) depiction of enactment as resource-independent action.

Table 2 provides an overview of the degree of specificity of the decision-making processes designed to address uncertain, ill-structured decision problems with regard to ends, means, and the selection criterion.

Execution, effectuation, and enactment are three different tools applicable to deal with ill-structured decision problems. Even though the answer to the question of which of these processes fits which particular decision best, only depends on the personal perception of the acting entrepreneur about the decision circumstances, in general, we can ascertain that the three processes constitute a range reflecting options suitable to deal with different degrees of ill-structuredness. We can tell that the more a problem is perceived as ill-structured and, as such, the more a problem is perceived as vague and unspecified, the more superior effectuation and enactment might be to execution (albeit slightly ill-structured problems that are perceived as risky rather than uncertain might be a good fit for effectuation, too, as argued by Welter and Kim (2017)).

The less the decision circumstances allow for a sufficient amount of planning through pre-structuring the ill-structured problem, the better effectuation and particularly enactment seem to fit. Accordingly, the range of possible processes generally goes from execution via effectuation to enactment with regard to the degree of ill-structuredness, vagueness, and lack of specificity of the underlying decision problem. However, focusing on the degree of detailed planning associated with each of the processes, the range goes from enactment via effectuation to execution and, hence, vice versa. It follows that less ill-structured, less vague, more specified problems allow for a more sophisticated planning and, hence, for the application of execution.

Figure 3 illustrates the range of possible decision-making processes in light of the degree of ill-structuredness, vagueness, and lack of specificity at one extreme and their inherent degree of detailed planning at the other.

Entrepreneurial decisions like that to enter or literally create a specific market with a particular (revolutionary) product, can barely be described in terms of isolated one-time decision problems that can be solved just once and permanently. Instead, judgmental entrepreneurial decision-making must be viewed as a continuous and recursive process (McMullen 2015; McMullen and Dimov 2013; Packard et al. 2017) and, as such, in terms of a period of time rather than a point in time.

The temporality of the entrepreneurial journey implies that entrepreneurs’ perceptions of their circumstances alter as time passes. Corbett (2007) emphasizes the importance not only of what one knows, but also of how one processes new information. Such new information, and particularly that acquired through learning and experience (Mitchell and Shepherd 2012), will naturally make entrepreneurs change how they regard their decision environment (Knight 1921). While it has been argued that Bayesian learning is at odds with Knightian uncertainty and, thus, unsuited for situations regularly faced by entrepreneurs (Packard et al. 2017), its fundamental idea of differing decision bases a priori and a posteriori respectively, can be transferred into the sphere of the entrepreneurial journey in a less formalized fashion.

We agree with Beckert and Bronk (2018) and Knight (1921) who argue that most decisions with regard to their degree of unknowingness are neither black nor white but various shades of gray. Hence, while the entrepreneur utilizes his or her specific knowledge (Wood and Williams 2014) about some things, other aspects are unknown and possibly unknowable. Therefore, entrepreneurs imagine potential future scenarios (Klein 2008) by drawing on their specific individual state of a priori knowledge. Once the initial decision has been taken (i.e., a posteriori), the foundation of the decision will be changed owing to the improved information base. Therefore, entrepreneurs might also subsequently reevaluate how they perceive the structure of a particular decision problem in their ongoing judgment and, consequently, the process they apply to that problem (similarly Jiang and Tornikoski 2019 who analyze the relation between perceived uncertainty and process selection).

Different directions can be imagined: It seems quite obvious that particularly when a product never seen before shall be launched, processes emphasizing the role of acting itself without much reference to structured pre-planning and prognoses can serve as a reasonable starting point, namely enactment and effectuation, which might even be applied complementarily (Bhowmick 2015). Effectuation has been directly linked to new ventures, new products, new markets (Sarasvathy 2001) and it therefore has a narrower field of application than execution. As time goes by, entrepreneurs learn and acquire experience that might impel them to switch from enactment or effectuation toward execution if they perceive the altered structure of the problem suitable.

However, even ventures that will create a new market are not limited to effectuation or enactment; but might instead be a good fit for execution. For instance, although Jeff Bezos was about to pioneer e-commerce with Amazon, he did so with structured financial planning and an emphasis on maximizing future return rather than with recourse to a certain amount of affordable loss (Stone 2013). Moreover, he was equipped with a range of quite specific goals rather than merely vague general human aspirations (Stone 2013).

While learning and experience might cause entrepreneurs to switch from effectuation or enactment to execution in their subsequent judgment, the direction might run vice versa too. Research finds that entrepreneurs can be prone to overconfidence (Busenitz and Barney 1997; McKelvie et al. 2011; Shepherd et al. 2015; Townsend et al. 2010; Wu and Knott 2006; Zacharakis and Shepherd 2001) and one result of such overconfidence can be the application of execution to a problem unsuitable for the process. An inexperienced entrepreneur entering a market for the first time with a novel product, for example, might falsely—which she can learn only ex post—perceive the decision problem as ill-structured to a degree that allows pre-planning and predicting. However, if it turns out that the execution process was inappropriate, the entrepreneur might alter her ideas about applying effectuation or enactment, if the initial decision did not prove fatal to the business. Regardless, the selection of a process to be applied to entrepreneurial decision-making is related to the life cycle of a product and/or an entire firm.

If an entrepreneur is not eager to enter a market with an innovative product but seeks to imitate other competitive products, execution seems to be a good starting point given the history of other products and firms manufacturing them. In that case, the entrepreneur essentially does not have to learn and experience everything alone and can benefit from the observable actions of others (Rae and Carswell 2000).

Different decision-making processes can also certainly be used within the same firm simultaneously. For instance, if a particular business has offered a specific product in one market for decades and can therefore address pending decisions with execution, it might at the same time plan on penetrating a foreign market that has never seen a similar product. Given that the potential consumers in the foreign market might be quite different from the consumers in the domestic market, perhaps owing to fundamentally different cultural values and norms (Hofstede 1980; Steenkamp et al. 1999), it can be wise to start off with an enactment- or effectuation-based process rather than confidently applying experience acquired in the domestic market in a new foreign one.

The literature offers different concepts of who is an entrepreneur or who acts entrepreneurially and the range spans from very broad understandings of entrepreneurship to quite narrow ones (Klein 2008; Salerno 2008). It has been argued (justifiably in the authors’ opinion) that most human action is in some sense entrepreneurial as it is undertaken in the face of uncertainty (Herbener 1992; von Mises 1949). At the other end of the scale, some literature prefers to very tightly define entrepreneurship as relating only to actions with reference to start-up firms (e.g., Kim 2015). Neither extreme is particularly useful for our discussion, and we therefore follow Klein (2008), Salerno (2008), and others in focusing on the middle ground between the two, incorporating far more than only start-up ventures while not considering any human action specifically entrepreneurial.

Moreover, entrepreneurship has inter alia been discussed as an (at most) analytically distinct function (“pure entrepreneur”) (Kirzner 1973) and based on a more practical approach characterizing entrepreneurs as having ownership of the means of production and, as such, being the ultimate decision-makers (Foss and Klein 2012; Klein and Klein 2001; Salerno 2008). The latter concept reveals that “the entrepreneur is nearly always also a capitalist, and the capitalist is also an entrepreneur” (Klein 1999: 21).

While execution can be the ideal decision-making process for any entrepreneur independent of the market she acts in while contingent on the actual decision she faces and her perception of the problem structure, it seems particularly suited for the specific cases of the financial-market entrepreneur (Klein 1999) or “financiers as entrepreneurs” (Klein 2008). Investors must not be misunderstood as mere passive funders of entrepreneurial ventures; rather, they are entrepreneurs themselves (Landström 1998), exercising judgment on the allocation of their scarce means among their investment opportunities, doing so under conditions of uncertainty, aiming to make profit. Scholars in the Austrian tradition such as von Mises (1949), Rothbard (1962), and Klein (1999) have even argued that the “capitalist-entrepreneur” plays the most important role in the market economy and can be viewed as the driving force behind it.

While there is no natural necessity for this particular class of entrepreneurs to make use of execution, this process tends to be the primary vehicle suitable to address investors’ decision problems. In general, they would rather not build on processes that bypass sufficient pre-planning along with prognostic elements, which means that neither effectuation nor enactment naturally seem the best fit. That is also because investors usually act on their personal anticipation of future returns (Klein 1999), a feature that is missing in these processes.² Therefore, even when particular entrepreneurs appear as effectuators or “enacters” creating their venture, execution might still play an important role in those ventures if the effectuators or “enacters” seek external financing by a financial-market entrepreneur. As Hsu (2004: 1805) points out, “convincing external resource providers such as venture capitalists to provide financial capital may be challenging” and to do so might actually be a function of how plausibly effectuators or “enacters” can communicate how they imagine their business will develop in the future. A mere reference to vaguely imagined effects—or in the words of Sarasvathy (2001: 262), “when destinations are unclear and there are no preexistent goals”—in case of effectuation or to the general and loose idea of action before sensemaking in enactment might be insufficient to convince the financial-market entrepreneur to invest. This, in fact, might have been the reason why Jeff Bezos deliberately worked out a business plan, including “typing revenue projections into an Excel spreadsheet” (Stone 2013: 29), before pitching his venture to potential investors (Stone 2013).