Causation and Effectuation PDF

Causation and Effectuation: Toward a Theoretical Shift from Economic Inevitability to Entrepreneurial Contingency Author(s): Saras D. Sarasvathy Source: The Academy of Management Review , Apr., 2001, Vol. 26, No. 2 (Apr., 2001), pp. 243-263 Published by: Academy of Management Stable URL: https://www...

Causation and Effectuation: Toward a Theoretical Shift from Economic Inevitability to Entrepreneurial Contingency Author(s): Saras D. Sarasvathy Source: The Academy of Management Review , Apr., 2001, Vol. 26, No. 2 (Apr., 2001), pp. 243-263 Published by: Academy of Management Stable URL: https://www.jstor.org/stable/259121 JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms Academy of Management is collaborating with JSTOR to digitize, preserve and extend access to The Academy of Management Review This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms t Academy of Management Review 2001, Vol. 26, No. 2, 243-263. CAUSATION AND EFFECTUATION: TOWARD A THEORETICAL SHIFT FROM ECONOMIC INEVITABILITY TO ENTREPRENEURIAL CONTINGENCY SARAS D. SARASVATHY University of Washington In economics and management theories, scholars have traditionally assumed the existence of artifacts such as firms/organizations and markets. I argue that an explanation for the creation of such artifacts requires the notion of effectuation. Causation rests on a logic of prediction, effectuation on the logic of control. I illustrate effectu- ation through business examples and realistic thought experiments, examine its connections with existing theories and empirical evidence, and offer a list of testable propositions for future empirical work. I now am eagerly striving, for example, to get this truth which I seem half to perceive, into words which shall make it show more clearly. If the words come, it will seem as if the striving itself had drawn or pulled them into actuality out from the state of merely possible being in which they were. How is this feat performed? How does the pulling pull? How do I get my hold on words not yet existent and when they come by what means have I made them come? Really it is the problem of creation; for in the end the question is: How do I make them be?... values is the set of values in terms of which he wants to act? (March, 1982: 74). Walk into an MBA classroom anywhere in the world. Chances are the discussion revolves around a decision or a set of decisions to be made. For example, classes with a more economic bent (e.g., managerial economics, market- ing, strategy) might be discussing the pricing decision. The standard formal approach to this decision involves setting the marginal revenue ... Sustaining, persevering, striving, paying with effort as we go, hanging on, and finally achieving our intention-this is action, this is effectuation in the only shape in which, by a pure experience-philosophy, the whereabouts of it anywhere can be discussed. Here is creation in its first intention, here is causality at work (James, 1912: 181, 183). equal to the marginal cost; a more adaptive approach might involve doing market research to discover the shape of the demand function and to arrive at a price that the market will bear. In another example, classes with a more psychological bent (e.g., human resources management, organization behavior, leadership) might be discussing personnel decisions, such as hiring the best person for the job or managing and/or leading a team. Approaches might range from psychometric measurements to avoiding well-understood biases, such as anchoring, escalation, groupthink, and so on. These decisions in economics and management may be discussed at several levels: individual, firm, industry/market, and economy. But underlying almost every one of these decisions is the assumed existence of the central artifacts and contexts of business within which the decisions take place. In other words, none of these decisions involves the creation of artifacts such as firms, markets, and economies. For example, the following are rarely, if at all, addressed in We know how to advise a society, an organization, or an individual if we are first given a consistent set of preferences. Under some conditions, we can suggest how to make decisions if the preferences are only consistent up to the point of specifying a series of independent constraints on the choice. But what about a normative theory of goal-finding behavior? What do we say when our client tells us that he is not sure his present set of I thank the Ewing Marion Kauffman Foundation for funding the empirical work that led to the development of the ideas in this article. I also thank Herb Simon, Anil Menon, and Lester Lave for their invaluable conversation, and the anonymous AMR reviewers, Edward Conlon, Marilyn Gist, Tom Jones, Tom Lee, Benyamin Lichtenstein, Scott Shane, S. Venkataraman, and Andy Wicks for commenting on earlier versions of this paper and for helping me improve it. our curricula: 243 This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms 244 Academy of Management * How do we make the pricing decision when the firm does not yet exist (i.e., no revenue functions or cost functions are given) or, even more interesting, when the market for the product/service does not yet exist (i.e., there is no demand function)? * How do we hire someone for an organization that does not yet exist? How do we even get able people to apply to a contingent organization-an organization whose existence is contingent upon acquiring employees (e.g., a knowledge-intensive firm, such as a software company)?! * How do we value firms in an industry that did not exist five years ago and is barely forming in the present (e.g., internet companies)? More interesting, how would we have valued them five years ago, when internet companies were barely emerging? * At the macro level, how do we create a capitalist economy from a formerly communist one? Or, more interesting, what should a postcapitalist economy look like?' A very large and growing fraction of people in business struggle with such decisions every day. Business all over the world is becoming more free-market oriented and more entrepre- neurial. Almost half the companies on the Fortune 500 list did not exist fifteen years ago. Emerging technologies, such as those on the internet, are not only creating rapid change but also fundamentally redefining how we truck and trade and how we interact with one another in every sphere of human action. Therefore, as March points out in the quote at the beginning of this article, questions such as the ones listed above ought to be an important part of our re- Review April to simply pursue an interesting idea that seems worth pursuing. Similarly, if we clearly knew which particular market to capture, we could presumably use techniques of market research and formulate strategies to penetrate it. In areas such as e-commerce, however, most markets are nascent or simply nonexistent. Marketing to markets that do not yet exist involves understanding how markets come to be. Similarly, valuing and financing a firm that does not yet exist involve understanding how firms come to be. And creating a firm in a market that does not yet exist involves understanding how to make decisions in the absence of preexistent goals. March sets out three justifications that re- searchers have used to ignore phenomena involving ambiguous, changing, and constructed goals and values: The first is that goal development and choice are independent processes, conceptually and behaviorally. The second is that the model of choice is never satisfied in fact and that deviations from the model accommodate the problems of introducing change. The third is that the idea of changing goals is so intractable in a normative theory of choice that nothing can be said about it. Since I am unpersuaded on the first and second justifications, my optimism with respect to the third is somewhat greater than most of my fellows (March, 1982: 72). In the past couple of decades, researchers have been struggling in March's spirit of optimism to take on these seemingly intractable questions. I hope to make a contribution here toward that effort by identifying and developing search endeavors. a decision model that involves processes of efEach of these questions involves the problem fectuation, rather than causation, and showing of choosing particular effects that may or may its use in the creation of new firms. Although a not implement intentional goals. For example, if general theory of effectuation could be develwe knew precisely what type of firm we wished oped to address all four types of questions listed to create, we could use existing theories and above, in this article I develop only a special principles to create the firm. But usually all the theory to explain the creation of new firms. entrepreneur knows when he or she starts out is After a brief definition of effectuation as consomething very general, such as the desire to trasted with causation, I explicate the processes make lots of money or to create a valuable leginvolved through two thought experimentsacy like a lasting institution, or, more common, one hypothetical and the other historical-and then I succinctly review several relevant streams of research in order to delineate the 'All four questions listed here can be addressed through space for effectuation processes in the literature a general theory of effectual reasoning, the main elements of which are explicated in this article. However, given the and to develop a rudimentary theory of effectucognitive and spatial limits of a single journal article, I focusation. Thereafter, I suggest connections to the on the first question alone. I address this question in conseminal works of three eminent researchers in siderable detail to illustrate (what is for now) a special management who have taken the lead toward theory of effectuation in the creation of firms in nonexistent new horizons in our discipline (March, 1982; or not-yet-existent markets. This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms 2001 Sarasvathy 245 Mintzberg, 1994; Weick, 1979). Following that, I so on. These are obviously oversimplified exam- examine some recent empirical evidence that ples. To bring the definitions closer to reality does not fit with the traditionally accepted par- through, say, the dinner example, we have to adigm of causation models and, finally, develop add elements of dynamism and contingencies of propositions based on effectuation at all four various kinds, including multiple interacting levels of phenomena: macro, industry/market, chefs and hosts and dinner guests. But the point here is that in each example the generalized firm, and individual. end goal or aspiration remains the same both in PROCESSES OF CAUSATION AND EFFECTUATION Definition: Causation2 processes take causation and effectuation-that is, to cook a meal, to build some wooden artifact, or to create a painting. In fact, an effect is the operationalization of an abstract human aspiration. The a particular effect as given and focus distinguishing characteristic between causa- on selecting between means to create tion and effectuation is in the set of choices: that effect. Effectuation processes take choosing between means to create a particular a set of means as given and focus on effect, versus choosing between many possible selecting between possible effects effects using a particular set of means. Whereas that can be created with that set of causation models consist of many-to-one map- means. pings, effectuation models involve one-to-many A simple example should help clarify and distinguish between the two types of processes. Imagine a chef assigned the task of cooking dinner. There are two ways the task can be or- ganized. In the first, the host or client picks out a menu in advance. All the chef needs to do is list the ingredients needed, shop for them, and then actually cook the meal. This is a process of causation. It begins with a given menu and focuses on selecting between effective ways to prepare the meal. In the second case, the host asks the chef to look through the cupboards in the kitchen for possible ingredients and utensils and then cook a meal. Here, the chef has to imagine possible mappings. Both causation and effectuation are integral parts of human reasoning that can occur simultaneously, overlapping and intertwining over different contexts of decisions and actions. Yet in this article I deliberately juxtapose them as a dichotomy to enable clearer theoretical exposition. Before embarking on a literature review to delineate the space for effectuation models, I present two realistic examples from business to illustrate and compare the two types of decision processes (i.e., causation and effectuation). The first thought experiment is a hypothetical onethat of creating an imaginary restaurant-and the second is historical-the story of U-Haul. menus based on the given ingredients and uten- sils, select the menu, and then prepare the meal. This is a process of effectuation. It begins with given ingredients and utensils and focuses on preparing one of many possible desirable meals with them. A variety of such simple examples can be imagined: a carpenter who is asked to build a desk, versus one who is given a toolbox and some wood and asked to build whatever he or she chooses; an artist who is asked to paint a portrait of a particular person, versus one who is given a blank canvas and some paints and re- quired to paint anything he or she chooses; and Thought Experiment #1: Curry in a Hurry In this example I trace the process for building an imaginary Indian restaurant, "Curry in a Hurry." Two cases, one using causation and the other effectuation, are examined. For the purposes of this illustration, the example chosen is a typical causation process that underlies many economic theories today-theories in which it is argued that artifacts such as firms are inevita- ble outcomes, given the preference orderings of economic actors and certain simple assump- tions of rationality (implying causal reasoning) in their choice behavior. The causation process used in the example here is typified by and embodied in the procedures stated by Philip 2A brief outline of the philosophical underpinnings of Kotler in his Marketing Management (1991: 63, causation is provided in a later section, titled "Future Theoretical Work on Effectuation." 263), a book that in its many editions is consid- This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms 246 Academy of Management Review April ered a classic and is widely used as a textbook eat out at least twice a week. That would help in MBA programs around the world. her determine her menu choices, d6cor, hours, Kotler defines a market as follows: "A market and other operational details. She could then consists of all the potential customers sharing a design marketing and sales campaigns to in- particular need or want who might be willing duce her target segment to try her restaurant. and able to engage in exchange to satisfy that She could also visit other Indian and fast food need or want" (1991: 63). Given a product or a restaurants and find some method of surveying service, Kotler suggests the following procedure them and then develop plausible demand fore- for bringing the product/service to market (note casts for her planned restaurant. that Kotler assumes the market exists): In any case, the process would involve considerable amounts of time and analytical effort. It 1. Analyze long-run opportunities in the market. 2. Research and select target markets. * Identify segmentation variables and segment the market. * Develop profiles of resulting segments. * Evaluate the attractiveness of each segment. * Select the target segment(s). * Identify possible positioning concepts for each target segment. * Select, develop, and communicate the chosen positioning concept. 3. Design marketing strategies. 4. Plan marketing programs. 5. Organize, implement, and control marketing effort. would also require resources both for research and, thereafter, for implementing the marketing strategies. In summary, the current paradigm suggests that we proceed inward to specifics from a larger, general universe-that is, to an optimal target segment from a predetermined market. In terms of Curry in a Hurry, this could mean something like a progression from the en- tire city of Pittsburgh to Fox Chapel (an affluent residential neighborhood) to the Joneses (specific customer profile of a wealthy family), as it were. Instead, if our imaginary entrepreneur were to use processes of effectuation to build her restaurant, she would have to proceed in the opposite This process is commonly known in marketing direction (note that effectuation is suggested as the STP-segmentation, targeting, and posi- here as a viable and descriptively valid alternative to the STP process-not as a normatively tioning-process. Curry in a Hurry is a restaurant with a new superior one). For example, instead of starting twist-say, an Indian restaurant with a fast food with the assumption of an existing market and section. The current paradigm using causation investing money and other resources to design processes indicates that, to implement this idea, the entrepreneur should start with a universe of all potential customers. Let us imagine that she wants to build her restaurant in Pittsburgh, Pennsylvania, USA, which will then become the initial universe or market for Curry in a Hurry. Assuming that the percentage of the population of Pittsburgh that totally abhors Indian food is negligible, the entrepreneur can start the STP the best possible restaurant for the given mar- process. Several relevant segmentation variables, such as demographics, residential neighborhoods, ethnic origin, marital status, income level, and patterns of eating out, could be used. Based on these, the entrepreneur could send out questionnaires to selected neighborhoods and organize focus groups at, say, the two major universities in Pittsburgh. Analyzing responses to the questionnaires and focus groups, she could arrive at a target segment-for example, wealthy families, both Indian and others, who ket, she would begin by examining the particular set of means or causes available to her. As- suming she has extremely limited monetary resources-say $20,000-she should think creatively to bring the idea to market with as close to zero resources as possible. She could do this by convincing an established restaurateur to become a strategic partner or by doing just enough market research to convince a financier to invest the money needed to start the restaurant. Another method of effectuation would be to convince a local Indian restaurant or a local fast food restaurant to allow her to put up a counter where she would actually sell a selection of Indian fast food. Selecting a menu and honing other such details would be seat-of-the-pants and tentative, perhaps a process of satisficing (Simon, 1959). Several other courses of effectuation can be imagined. Perhaps the course the entrepreneur This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms 2001 Sarasvathy actually pursues is to contact one or two of her friends or relatives who work downtown and bring them and their office colleagues some of her food to taste. If the people in the office like her food, she might get a lunch delivery service going. Over time, she might develop enough of a customer base to start a restaurant or else, after a few weeks of trying to build the lunch business, she might discover that the people who said they enjoyed her food did not really enjoy it so much as they did her quirky personality and conversation, particularly her rather unusual life perceptions. Our imaginary entrepreneur might now decide to give up the lunch business and start writing a book, going on the lecture circuit and eventually building a business in the motivational consulting industry! Given the exact same starting point- but with a different set of contingencies-the entrepreneur might end up building one of a variety of businesses. To take a quick tour of some possibilities, consider the following: Whoever first buys the food from our imaginary Curry in a Hurry entrepreneur becomes, by definition, the first target customer. By continually listening to the customer and building an ever-increasing network of customers and strategic partners, the entrepreneur can then identify a workable segment profile. For example, if the first customers who actually buy the food and come back for more are working women of varied ethnic origin, this becomes her target segment. Depending on what the first customer really wants, she can start defining her market. If the customer is really interested in the food, the entrepreneur can start targeting all working women in the geographic location, or she can think in terms of locating more outlets in areas with working women of similar profiles-a "Women in a Hurry" franchise? Or, if the customer is interested primarily in the idea of ethnic or exotic entertainment, rather than merely in food, the entrepreneur might develop other products, such as catering services, party planning, and so on-"Curry Favors"?3 Perhaps, if the customers buy food from her because they actually enjoy learning about new cultures, she might offer lectures and classes, maybe beginning with Indian cooking and mov- 247 ing on to cultural aspects, including concerts and ancient history and philosophy, and the profound idea that food is a vehicle of cultural exploration-"School of Curry"? Or maybe what really interests them is theme tours and other travel options to India and the Far East"Curryland Travels"? In a nutshell, in using effectuation processes to build her firm, the entrepreneur can build several different types of firms in completely disparate industries. This means that the original idea (or set of causes) does not imply any one single strategic universe for the firm (or effect). Instead, the process of effectuation allows the entrepreneur to create one or more several possible effects irrespective of the generalized end goal with which she started. The process not only enables the realization of several possible effects (although generally one or only a few are actually realized in the implementation) but it also allows a decision maker to change his or her goals and even to shape and construct them over time, making use of contingencies as they arise. Furthermore, even the generalized aspiration of starting a business is not a necessary starting point for effectuation processes. Several successful businesses and even great companies have begun without any conscious initial intention on the part of the founders. To cite but one example, the waste management giant Browning Ferris Industries (BFI) began as the result of contingent problem solving. In 1967, while presiding over a community association meeting, Tom Fatjo, a respected professional in Houston, Texas, listened to members whine about the garbage problem in their subdivision. Exasperated, he suggested that maybe the community should haul its own garbage. The community turned to him and dared him to do it himself. After physically hauling garbage while continuing his professional career for over a year, he realized the potential in garbage and went on to build BFI. In a similar vein, the Curry in a Hurry entrepreneur's journey of effectuation might also be the result of any one of a wide variety of serendipitous events. For example, a chance suggestion made by a friend after tasting her food on a social occasion might have started the process or, as happens in the case of many entrepreneurs today, an unexpected misfortune might 3 I apologize for the cheesy names, but, hopefully, they get the message across. have forced her to earn a living on her own. This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms 248 Academy of Management Thought Experiment #2: U-Haul The following example also consists of a thought experiment to trace the processes used in the creation of U-Haul. Although there is no detailed history of U-Haul in which the actual processes used by its founder are traced, the thought experiment uses extracts from the company's history posted on its website, combined with Silver's study: Review April If we examine each of these entrepreneurial decisions using only causation processes, the best current theory and practice within each functional domain will fail to lead us to a good decision. For instance, if we examine the marketing decisions using current theories and practice in marketing, we have to figure out the potential universe of customers for U-Haul and develop a marketing plan that targets the segment with the highest potential for return on Like many other successful ventures, the con- investment. Even if Shoen could somehow find a cept for U-Haul? was provoked by need. After way to figure all this out in 1945 without ex- World War II the population of the United States became more mobile and migratory. There existed an obvious widespread need for do-ityourself moving equipment on a one-way, nationwide basis. It was the visionary approach of U-Haul that recognized this need, acted upon it and literally created an industry. With $5,000, L. S. Shoen, his wife Anna Mary Carty Shoen and their young child moved to the Carty ranch in Ridgefield, Washington. There, with the help of the Carty family, the Shoens built the first U-Haul trailers in the fall of 1945, using the ranch's automobile garage (and milk house) as the first manufacturing plant for the budding U-Haul Co. By the end of 1949, it was possible to rent a trailer one-way from city-to-city throughout most of the United States (Silver, 1985: 387-390). The historical facts are that in four years Shoen transformed his perception of an obvious widespread need, $5,000, and access to an automobile garage into a nationwide firm with a complicated production function, thousands of stakeholders, and what was essentially 100 per- hausting his initial capital of $5,000, there was presumably no way he would be able to realistically convince any potential investor to put up the enormous outlay called for in such a marketing plan. This has been tested through class discus- sions using the creation of U-Haul as a case study. Students typically come to one of two conclusions: 1. This project is not financially viable-the resource requirements are very large (esti- mates range between $20 million and $50 million in current dollars) and overwhelm any attempt to price the service viably; OR 2. This project is not viable psychologicallyeven if it were financially viable and potentially profitable, the initial resources required would be so large as to raise the question of why anybody with control over $20 to $50 million would want to invest it in this relatively mundane but risky project consisting of buying trucks and renting locations across the country. cent market share in the newly created do-it- Yet U-Haul was created with an almost instan- yourself moving industry. In our thought exper- taneous national presence for a very small fi- iment we can now examine the minimal set of decisions that he had to make in effecting this transformation: nancial outlay! Shoen used processes of effectuation that involved his seizing and exploiting contingencies 1. How many moving vans/trucks should he buy or make? 2. How many locations would he need to open? 3. How many employees should he hire? (One per location or more?) 4. From whom should he raise the capital? 5. Should he open a few locations regionally or go national at once? 6. How should he establish his market presence-advertise? If so, how? 7. Putting it all together, how should he price the product? 8. Given the fact that all he had was $5,000 to begin with, should he move to Ridgefield and begin building the trucks? through an expanding network of human alli- ances. Instead of trying to raise the money to buy a large number of trucks or trying to start the company with very few locations, he did the following:4 . He began by establishing an identity. The trailers were painted bright orange. The name "U-Haul Co." was established. Trail- ers were imaged on the sides and back with a sales message: "U-Haul Co., Rental Trail- ' The sources used were the company's own historical records and Silver (1985). This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms 2001 Sarasvathy ers, $2.00 Per Day"-always advertising them, whether on the road or on display. * He convinced friends, family members, and customers (who then convinced others close to them, and so on) to individually make down payments on trucks and then lend him the use of the trucks. * He contracted with service station outlets (including national chains) to merchandise trailer rentals, eliminating the need for buying space in cities across the country and for recruiting employees to man the spaces. * He offered early customers a discount on their trailer rental for establishing a U-Haul rental agent at their destination and established a commission structure for dealers. Thus, with hardly any employees and a ridicu- lously small outlay of funds, U-Haul came into being. Furthermore, in the case of U-Haul, in the initial stages of implementing processes of effectuation, the firm appears almost to have been in the business of selling livelihoods to potential U-Haul franchisees (before the idea of franchising was developed), rather than in the oneway rental business. This case study particularly highlights the unique role of the decision maker in solving the existence problem through effectuation. Characteristics of decision makers, such as who they are, what they know, and whom they know, form the primary set of means that combine with contingencies to create an effect that is not preselected but that gets constructed as an integral pc.rt of the effectuation process. The effectuator merely pursues an aspiration and visualizes a set of actions for transforming the original idea into a firm-not into the particular predetermined or optimal firm, but a very generalized aspiration of a firm. The commitment to such a tentative set of actions includes proceeding with no a priori guarantees or even strong potentialities for success. The effectuator more often than not proceeds without any certainties about the existence of a market or a demand curve, let alone a market for his or her product, or a potential revenue curve. In cases involving spectacular successes (Silver, 1985), the effectuating entrepreneurs' vision appears to involve more than the identification and pursuit of an opportunity; it seems to include the very creation of the opportunity as part of the implementation of the entrepreneurial process. The latent market for U-Haul, consisting of the obvious widespread need for one-way rentals, was only a necessary condition for its 249 actualization. Sufficiency is provided by active implementations of imagined solutions that seize and build on several types of contingencies that ultimately carve out the structure and shape of the market. Currently, markets on the internet are being created in this manner, through contingent interactions between the imaginations of effectuators and the aspirations of their partners in the process, whether the partners consist of customers, investors, and/or various types of alliances. A RUDIMENTARY THEORY OF EFFECTUATION PROCESSES IN BUSINESS Before developing a theory for decisions in- volving effectuation and delineating its space within the literature, it is necessary to emphasize that effectuation processes are not posited here as "better" or "more efficient" than causation processes in creating artifacts such as firms, markets, and economies. Under what cir- cumstances which types of processes provide particular advantages and disadvantages is an issue to be resolved through future empirical studies. For example, in the thought experiment of Curry in a Hurry, presented above, if the entrepreneur clearly wants to build an up-scale Indian restaurant, she presumably will be better off using causation processes than effectuation. But if she has only the generalized aspiration of building a successful business of her own with relatively limited access to resources, she should consider effectuation processes. Summarizing from the literature on decision making, the anatomy of a decision involves . a given goal to be achieved or a decision to be made (usually well structured and specific), . a set of alternative means or causes (that can be generated through the decision process), . constraints on possible means (usually imposed by the environment), and . criteria for selecting between the means (usually maximization of expected return in terms of the predetermined goal). Clearly, this structure assumes a decision process involving causation. A decision involving effectuation, however, consists of . a given set of means (that usually consists of relatively unalterable characteristics/ circumstances of the decision maker), This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms 250 Academy of Management . a set of effects or possible operationalizations of generalized aspirations (most- ly generated through the decision process), . constraints on (and opportunities for) possi- ble effects (usually imposed by the limited Review April dynamic process involving other decision makers interacting with one another. Table 1 presents a list of the distinguishing characteristics of the two types of processes. means as well as by the environment and its Causation processes are effect dependent. Ef- contingencies), and fectuation processes are actor dependent. Cau- . criteria for selecting between the effects (usually a predetermined level of affordable loss or acceptable risk related to the given means). sation processes are excellent at exploiting knowledge. Effectuation processes are excellent at exploiting contingencies. Nature abounds in particular events with regular causes that can Entrepreneurs begin with three categories of "means": they know who they are, what they know, and whom they know-their own traits, tastes, and abilities; the knowledge corridors they are in; and the social networks they are a part of. At the level of the firm, the corresponding means are its physical resources, human resources, and organizational resources, ax la the resource-based theory of the firm (Barney, 1991). At the level of the economy, these means become demographics, current technology re- gimes, and sociopolitical institutions (such as property rights). One could speculate that effectuation processes are more general and more ubiquitous than causation processes in human decisions. For example, on most nights most people cook dinner using an effectuation process-that is, they look around in their kitchen cupboards for what's available and fix themselves something. Only rarely do they decide to throw a dinner party and carefully develop a causation process for accomplishing it (i.e., choose a menu, shop for specific ingredients, and follow precise rec- ipes). It stands to reason, then, that effectuation processes may not be very helpful for throwing a great dinner party. But human life usually comes stocked with cupboards that open and close at unexpected moments, often containing unspecified ingredients that the decision maker has little choice over; grocery shops are typically too far away or closed; and cookware often has to be borrowed from neighbors. To put it more mundanely, in cases in which a particular effect has been preselected by the decision maker, causation processes can be applied to choose the best, the fastest, the most efficient, or the most economical method to achieve the cho- sen effect; imagining possible effects and choosing among them, however, involve characteristics of the decision maker(s) and his or her (their) ability to identify and use contingencies over a be analyzed and understood, and, therefore, causation processes are excellent when dealing with natural phenomena. Human life abounds in contingencies that cannot easily be analyzed and predicted but can only be seized and exploited, and, therefore, effectuation processes are far more frequent and very much more useful in understanding and dealing with spheres of human action. This is especially true when dealing with the uncertainties of future phenom- ena and problems of existence. An examination of existing research on decisions dealing with uncertainties pertaining to the future (even if the research predominantly involves causation processes) should be useful in delineating the space for processes of effec- tuation. Researc'hers in areas ranging from mathematics, statistics, and economics to psy- chology, sociology, and business have grappled with decisions involving future phenomena. His- torically, the research on decision making under uncertainty can be divided into (1) the development of normative, rational decision models (e.g., Focardi & Jonas, 1998; MacCrimmon, Wehrung, & Stanbury, 1986; Shapira, 1997) and (2) empirical investigations into bounds on that rationality in actual decision makers (e.g., Einhorn & Hogarth, 1981; Kahneman & Tversky, 1990; Taylor 1984; Zey, 1998). The normative development is rooted in the conceptual distinction between "risk" and "uncertainty" (Knight, 1921). The commonly used statistical metaphor of the urn containing different colored balls serves to illustrate the difference between the two (Kamien, 1994). Problems involving risk are akin to a speculative game involving an urn containing five green balls and five red balls. Whoever draws a red ball is awarded a prize of $50. For any given draw, we can precisely calculate the probability of getting a red ball, because we know the underlying distribution of balls inside the urn from which we are making the draw. Problems involving This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms 2001 Sarasvathy 251 TABLE 1 Contrasting Causation and Effectuation Categories of Differentiation Causation Processes Effectuation Processes Givens Effect is given Only some means or tools are given Decision-making Help choose between means to achieve Help choose between possible effects that selection criteria the given effect can be created with given means Selection criteria based on expected Selection criteria based on affordable loss return or acceptable risk Effect dependent: Choice of means is Actor dependent: Given specific means, driven by characteristics of the effect choice of effect is driven by the decision maker wants to create and characteristics of the actor and his or her his or her knowledge of possible ability to discover and use contingencies means Competencies Excellent at exploiting knowledge Excellent at exploiting contingencies employed Context of More ubiquitous in nature More ubiquitous in human action relevance More useful in static, linear, and Explicit assumption of dynamic, nonlinear, independent environments and ecological environments Nature of Focus on the predictable aspects of an Focus on the controllable aspects of an unknowns uncertain future unpredictable future Underlying logic To the extent we can predict future, we To the extent we can control future, we do can control it not need to predict it Outcomes Market share in existent markets through New markets created through alliances and competitive strategies other cooperative strategies uncertainty involve the same award of $50 for the draw of a red ball, except we do not know how many balls are in the urn, what colors they are, or even if there are any red balls at all in the distribution. In statistical terminology, decisions involving the first type of urn, with the known distribution, call for classical analytical techniques, and decisions involving the second type of urn, with the unknown distribution, call for estimation techniques. Once the underlying distribution is discovered through estimation procedures, the urn with the unknown distribution is transformed, as it were, into the urn with the known distribution, and it becomes susceptible to analytical techniques. Real-life examples of risk include all types of insurance, some areas of the stock markets, and gaming of various types. Forecasting demand for very well-established products, such as Coca-Cola or personal computers nowadays, also falls within this category. Some real-life examples of uncertainty include dealing with environmental pollution, global warming, ge- netic cloning, and commercialization of innovations-particularly radical innovations. Experiments by researchers developing normative models have demonstrated that human beings in general prefer the "risky or known distribution" urn over the "uncertain or unknown distribution" urn (Ellsberg, 1961). But some researchers, such as those studying creative problem solving (Getzels & Csikszentmi- halyi, 1976), scientific discovery (Kulkarni & Simon, 1986), and entrepreneurship (Dickson & Giglierano, 1986; Kamien, 1994), have speculated that since creative problem solvers like entrepreneurs have been shown to have a high tolerance for ambiguity, they will have a preference for the urn with the unknown distribution. Both normative approaches have been qualified by other researchers, who have shown that human beings in general are not strictly rational (Simon, 1959). Instead, their rationality is bounded by cognitive limitations, such as physiological constraints on computational capacity (Payne, Bettman, & Johnson, 1993), and psycho- This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms 252 Academy of Management logical limitations, such as biases and fallacies (Bar-Hillel, 1980; Tversky & Kahneman, 1982). Yet this does not imply that decision makers are irrational. Rather, the evidence suggests that within certain bounds, decision makers use heuristics and inductive logics that often lead to very effective decisions (Gigerenzer, Hell, & Blank, 1988). The arguments from both perspectives- unbounded rationality and bounded rationality-can be summarized as follows. If the decision makers believe they are dealing with a measurable or relatively predictable future, they will tend to do some systematic information gathering and invest some effort on a reasonable analysis of that information, within certain bounds. Similarly, if they believe they are dealing with relatively unpredictable phenomena, they will try to gather information through ex- perimental and iterative learning techniques aimed at first discovering the underlying distri- bution of the future. This logically implies that the decision makers' underlying beliefs about the future phenomena that impact a particular decision can be deduced by examining the types of heuristics and logical approaches they use in making the decision. In terms of the urn metaphor used to describe causation processes of risk and uncertainty, the process of effectuation seems to suggest the following conjecture about a decision maker's logic: "I do not care what color the balls are in the urn or what their underlying distribution is. If I am playing a game where drawing a red ball wins $50, I will go acquire red balls and put them in the urn. I will also look for other people who have red balls and induce them to put them in the urn and play the game as my partners. As time goes by, there will be so many red balls in the distribution as to make almost every draw a red ball. Furthermore, if neither I nor my acquaintances have red balls, but only green ones, we will put enough of them in the urn so as to make the original game obsolete and create a new game where green balls win." In sum, this conjecture can be embodied in the following four principles that form the core of a rudimentary theory of effectuation, graphically depicted in Figure 1: 1. Affordable loss rather than expected returns: Causation models focus on maximizing the potential returns for a decision by selecting optimal strategies. Effectuation predetermines Review April how much loss is affordable and focuses on experimenting with as many strategies as possible with the given limited means. The effectuator prefers options that create more options in the future over those that maximize returns in the present. 2. Strategic alliances rather than competitive analyses: Causation models, such as the Porter model in strategy, emphasize detailed competitive analyses (Porter, 1980). Effectuation emphasizes strategic alliances and precommitments from stakeholders as a way to reduce and/or eliminate uncertainty and to erect entry barriers. 3. Exploitation of contingencies rather than exploitation of preexisting knowledge: When preexisting knowledge, such as expertise in a particular new technology, forms the source of competitive advantage, causation models might be preferable. Effectuation, however, would be better for exploiting contingencies that arose unexpectedly over time. 4. Controlling an unpredictable future rather than predicting an uncertain one: Causation processes focus on the predictable aspects of an uncertain future. The logic for using causation processes is: To the extent that we can predict the future, we can control it. Effectuation, however, focuses on the controllable aspects of an unpredictable future. The logic for using effectuation processes is: To the extent that we can control the future, we do not need to predict it. This logic is particularly useful in areas where human action (locally or in the aggregate) is the predominant factor shaping the future. For example, instead of defining a market as the universe of all possible customers as Kotler defines it, an effectuator would define his or her market as a community of people willing and able to commit enough resources and talents to sustain the particular enterprise. In the former case, the market is assumed to exist in- dependent of the firm or entrepreneur, and the task of the entrepreneur becomes to grab as much of that market as possible. In the latter case, however, the founder, along with others, creates the market by bringing together enough stakeholders who "buy into" the idea to sustain the enterprise. Since the structure of what exactly the enterprise is is left open and is dependent upon the particular commitments made by the stakeholders, the need for prediction is greatly reduced, if not completely obliterated. In This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms 2001 Sarasvathy 0 U) 0~ 0 0 0 0 .0 0 253 , h~~~~~ 0 U) U) U) U) U) ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~41 0 Contingencies U2 o U~~~~~~~~cici 0U)~~~~~~~~~~~~~~~~~ -0~~~~Cninece o~~~~~~~~~~~~~~~~~ 0 .J 0 0 )0c 0 -U0 0~~~~~~~~~~~0~ 0 0 0 0 0 ci~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~o~ U). . -' -- I Ics I I 1"~~~-00 1- 1- 1 1' . ,0#' U) 0UolU S0I Ol 0 0 u 0.2.0 3 o Ev u z0)< 0) Q U~~~~~~~~~~~~~~~~~~~~~~~~~~~~() 3 X 3 i z < Ev 3 Q O i O < <)A U ? ? L)~~~~~e ConinenIe ;. 0 0UU U) W 00 CIQ CY) ~ ~ ~U .) - (~~~~~~~~~~~~~~~~~~~~~jO~~~ ~~~~~~~~~~~~..... - S 0~~~~~~~~~~~~~~~~~~~~~~~ ci)~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~a 8 E a t g 2 X s j:; X E 8 F X E Y g i O; ! n~~~~~~~~~~~0 .q a 0 4S, j /,.0 2 0 U-. U0 . . .... .......... UOII1DUIB3:II .0 -~~~~~~ U) .~~~~~~~~~U 0 O 0 ~ ::S (1) ~ . ~0 On 0a.ci .n 0W i) 0~ J This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms c)0 0 Q) 0~ 0 00 O 254 Academy of Management other words, the particular firm created becomes the residual of a process of constructing a network of partnerships and precommitments (Burt, 1992), and the market itself is an aggregated taxonomy of such sustainable sets of partner- Review April ... We should indeed be able to develop better techniques. Whatever those techniques may be, however, they will almost certainly undermine the superstructure of biases erected on purpose, consistency, and rationality. They will involve some way of thinking about action now as occurring in terms of a set of unknown future values ships and commitments. (March, 1982: 75). CONNECTIONS WITH EXISTING THEORIES AND EMPIRICAL EVIDENCE A theory, however rudimentary, should not only identify gaps in our existing understanding of phenomena but should also be able to integrate existing theories and evidence that do not quite fit the current paradigm and, ultimately, should provide new hypotheses and predictions to be tested through future work. In this section I examine an important subset of existing theories and evidence that does not quite fit into the current paradigm of decision making using causation processes. In the next section I set out a list of propositions for decision making at all four levels of business phenomena. The intellectual lineage of the ideas influenc- ing the theory of effectuation presented in this article includes a very large and impressive list of thinkers, ranging from the pragmatic philos- ophers at the turn of the century to current leaders of thought in economics and management: Peirce (1878), James (1912), Knight (1921), Lindblom (1959), Simon (1959), Vickers (1965), Allison (1969), Weick (1979), Nystrom & Starbuck (1981), Buchanan & Vanberg (1991), March (1982), Burt (1992), and Mintzberg (1994). I examine a limited subset of the theoretical work done by three of these researchers here and follow with some additional empirical evidence that does not fit into the current paradigm. I also present brief outlines of future theoretical development out- side the scope of the current paper. Along the way, I highlight connections to effectuation. From the pioneering work Organizations (March & Simon, 1958) to several recent articles in management journals, March has created a substantial body of theories and empirical evidence on how human beings behave, make decisions, and interact with one another and with the external environment in organizations. Of particular value to building a theory of effectuation are his ideas on the tradeoffs between exploration and exploi- tation in organizational learning (March, 1991) and his inspirational exhortation to researchers to challenge the assumption of preexistent goals in decision making (March, 1982). Organizational learning involves decisions in which scarce resources (including attention) are allocated between the exploration of new possi- bilities and the exploitation of old certainties. These decisions are complicated by the fact that their costs and benefits may be dispersed over time and space and that they are subject to the effects of ecological interaction. Yet, balancing the allocation between exploration and exploitation is crucial to the survival and sustenance of the organization. Understanding the relationship between these two horns of a continuing dilemma in organizational evolution leads us away from a linear approach to such concepts as "success" and "sustainable competitive advantage." For example, by introducing a new technology, such as a computerized decision support system, an organization may decrease its chances of being the worst competitor, but it may reduce its chances of being the overall winner in the game (March, 1991: 84). It would be rather obvious to speculate that decision units of exploration would contain pro- James G. March To say that we make decisions now in terms of goals that will only be knowable later is nonsensical-as long as we accept the basic framework of the theory of choice and its presumptions of pre-existent goals. I do not know in detail what is required, but I think it will be substantial. As we challenge the dogma of pre-existent goals, we will be forced to reexamine some of our most precious prejudices .... cesses of effectuation, whereas causation models would dominate exploitation. But, more interesting, one could speculate that the problem of allocation of resources between exploration and exploitation might itself be modeled more effectively using an effectuation rather than a causation model. March's exposition on exploration and exploitation also brings out that causal reasoning and effectual reasoning need This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms 2001 Sarasvathy 255 not always pull in opposite directions. Instead, the considered opinions of mutual fund experts they can work in a complementary fashion, just (Borges, Goldstein, Ortmann, & Gigerenzer, as exploration and exploitation can both be 1999). Another example of the benefits of not used by a firm to sustain its market share over having expertise comes from the area of re- different spatial and temporal contexts. search methods, embodied in Gersick's discov- ery of a new model of group development. Eschewing the normative linear movement of Henry Mintzberg research-from careful literature review to de- I would like to introduce just one fact5 here. In one sense, it is the only real fact I know in all of the literature of strategic management... ... Honda's success, (in capturing two thirds of the American motorcycle market) if we are to believe those who did it and not those who figured it, was built precisely on what they initially believed to be a probable non-starter-namely the small motorcycle. Their own priors were that a market without small motorcycles would not buy small motorcycles. Had they had a proper planning process in place .. . this non-starter would have been eliminated at the outset-plan rationally and be done with it. But Honda was badly managed in this regard, and so a few Japanese managers, riding around on those little things in Los Angeles, were pleasantly surprised. They learned. (General Motors was apparently well managed in this regard, because a product development manager there once told me that they had a mini-van on the drawing boards long before Chrysler ever did but that this probable non-starter was scuttled in the planning process) (Mintzberg, 1991: 92). Success stories of probable nonstarters abound in the history of economics, management, and human affairs in general. For example, Polanyi comments on one of his contribu- tions to physics: adopted a mode of unconstrained curiosity and immersion in the phenomena, which led her to the element of surprise that was crucial to her discovery (Gersick, 1992). The prolificacy of successful nonstarters in hu- man affairs is matched only by sure things that fail disastrously. In a detailed review of the pre- dictive accuracy of forecasting by experts in various fields, including population, economics, technology, and so forth, Hogarth and Makridakis (1981) conclude that the evidence indicated fore- casting errors varied from a few to a few hundred percentage points and that forecasting was notoriously inaccurate. Also, using hundreds of studies in management and other areas of human behavior, Mintzberg makes a powerful argument that strategic planning is "not" strategy formation (Mintzberg, 1994). Once again, the evidence seems to suggest that a different model of decision making is required: one that does not focus on analysis and prediction but on synthesis and action. Effectuation provides one possible alternative, particularly in the problem of strategy formation, I would never have conceived my theory, let alone have made a great effort to verify it, if I had been more familiar with major developments in physics that were taking place. Moreover, my initial ignorance of the powerful, false objections that were raised against my ideas protected those ideas from being nipped in the bud (Polanyi, 1963: 1013). Researchers in cognitive science have explored the importance of so-called ignorance in the form of a recognition heuristic. This research into ignorance-based heuristics explains phe- nomena such as (1) the success of Benetton's ad campaign that conveyed nothing about the product but sought only to induce name recog- nition (Goldstein & Gigerenzer, 1999) and (2) the striking stock market returns generated by the recognition knowledge of pedestrians that beat 5The fact he is referring to is the title of the article, "Learning 1, Planning 0." duction of hypotheses, to careful operation- alization, to design, to inference making-she which belongs in the category of existence prob- lems such as those listed in the beginning of this article. Karl E. Weick I want to argue that one reason we theorize poorly about what matters most is because we use discourse that makes it hard to capture living forward. Living forward is a blend of thrownness, making do, journeys stitched together by faith, presumptions, expectations, alertness, and actions-all of which may amount to something, although we will know for sure what that something may be only when it is too late to do much about it. Unsettled, emergent, contingent living forward contrasts sharply with our backward- oriented theoretical propositions that depict that living as settled, causally connected, and coherent after the fact (Weick, 1999: 135). Weick's theory of enactment-retention-selection puts decision makers in organizations at This content downloaded from 131.155.94.236 on Sun, 21 Nov 2021 14:45:15 UTC All use subject to https://about.jstor.org/terms 256 Academy of Management Review April the center stage of the organization's evolution Effectuation in areas other than the creation of (Weick, 1979). Unlike in commonly accepted economic artifacts. I have primarily brought out models of evolution, where selection is exclu- the role of effectuation within the normative de- sively the prerogative of the environment, Weick cision theory literature, but there is a substan- argues that "decision makers in organizations tial body of theoretical and empirical work pro- intervene between the environment and its ef- viding alternative perspectives that need to be fects inside the organization, which means that examined in more detail. Connections of effec- selection criteria become lodged more in tuation with such theories as March's "garbage the decision-makers than in the environment" can" model, Weick's enactment processes, Lind- (Weick, 1979: 125). blom's successive comparisons framework, Gig- But this intervention is not coherently planned erenzer's ignorance-based heuristics, Simon's or causally prescribed as most mainstream re- bounded rationality, and several others (e.g., the search on organizations seems to emphasize. literature on improvisation and bricola