Constructing Vision with Scenario Planning
Terry R. Schumacher
Rose-Hulman Institute of Technology, Engineering Management Dept., Terre Haute, IN 47803 USA Abstract Strategic vision is often included as an important component of leadership. Yet there is relatively little guidance offered in the management literature on how to acquire vision. This paper describes practices that facilitate scenario planning so that it becomes a process for creating shared vision. Most of the work on scenarios addresses the mechanics of scenario construction. Those authors adopt a planning perspective and suggest scenario planning can benefit organizations by stimulating creative thinking about the future or improving forecasts. The Scenarios-to-Strategies (S2S) approach is presented, and scenario planning is considered from a communication perspective. Facilitation practices that enhance traditional scenario-building processes are presented which support the social processes of constructing shared vision. These operate on the layer of participants' cognitive processes, to integrate the different participants' views into a unified, shared framework that heightens understanding and commitment. Example scenario planning projects from two industries, electric utilities and software research, are summarized to demonstrate lessons learned that enhance the facilitation of scenarios as a group process. Introduction Scenarios are a commonly used management technique. There is wide variation in the details of their application. This paper briefly summarizes the growing attention to vision as a necessary component of leadership in organizations, and then describes the Scenarios-toStrategies method (S2S) and its benefits. Features of the S2S are presented in contrast with another approach that is here labeled 'Level One Scenarios'. An example of a Level One Scenario approach is available as a published case [11]. The S2S is considered to be a Level Two Scenario method in the comparison table below. Describing the contrasts between the approach used in the UPS case, and in the S2S, helps distinguish the attributes of the later. Recently the “Strategy-as-Practice” perspective has emerged and is receiving growing attention. It conceptualizes strategy as a social activity, seeing strategy not as something an organization has, but rather something that people do. The S2S method is consistent with this view bringing the Constructivist communication theory into scenario facilitation practices.

Leaders as Vision Providers “When there is no vision, the people perish.” There has been growing attention to the concept of vision in management literature for more than 20 years and there is growing awareness that shared vision contributes to enhanced performance. Zaleznik [45] distinguished mangers and leaders, and attributed creating new
Schumacher Constructing Vision with Scenarios ! PICMET 2012

approaches and imagining new areas to leaders. He observed that, in contrast to managers, leaders' key concerns were the meaning of events to participants, and that they were "Shaping ideas instead of responding to them". This emphasis on creating a new vision is also the core of the S2S approach. Peters & Waterman suggested the importance of vision 20 years ago [26]. In part, the success of high-performing organizations is due to the high level of commitment among those who undertake necessary tasks. Peters questions the effectiveness of a vision that is portrayed in numeric form as he believes this cannot capture the emotional involvement needed for high levels of motivation [28]. Plans and purely analytical perspectives appeal to the intellect, and this logic does help people understand WHY a goal has been selected. He suggests that vision, through its appeal to emotion can reach the will, and can create a much greater commit to goals, something greater than a mere understanding of the goals. Kouzes and Posner [17] include "inspiring a shared vision" as one of "five fundamental practices (that allow) leaders to get extraordinary things done." They observe "You cannot command commitment, you can only inspire it" and believe a shared vision is at the root of creating commitment. The leader’s own belief in, and enthusiasm for the vision are "the spark that ignites the flame of inspiration." Senge [32] comments at length on the important role of vision, and its place in leaders' activities, "Building shared vision must be seen as the central element of the daily work of leaders. It is on-going and never-ending." According to Senge, shared vision creates a common identity that supports trust and communication. It connects people and provides an exhilaration for undertaking important work. He uses a metaphor of a rubber band stretched between one's hands, the existing organizational reality the lower hand, and the vision characterizing an idealized future as the upper hand. When people accept a vision, the 'creative tension' between existing and desired states acts like cognitive dissonance and pulls people into resolving that tension by achieving the vision. For Senge, a shared vision is not just an idea, but "a force in peoples hearts". That force motivates them to accomplish the vision. Conger [9,10] puts articulation of vision at the center of his 3-stage Charismatic Leadership model. The first stage is to assess critically the existing situation through observation, reflection, need and ability assessment, and satisfaction with the status quo. His second stage is to formulate and articulate an idealized future vision. This vision needs to solve key organization problems and he offers suggestions including, using images and stories, that the vision should contrast strongly with the status quo, and that the vision should be presented as positive (an attractor) and not as negative. Conger also notes leaders need to demonstrate their commitment to the vision, because "trust begins when followers perceive, beyond a shadow of a doubt, their leader is unflinchingly dedicated to the vision and is willing to work even at considerable personal risk." This author's own experience in an extremely successful high-tech start-up witnessed the central role of shared vision in creating a world-class competitor. Leaders there spent considerable time repeating the vision and evolving it as needed to accommodate changing circumstances. They frequently explained their decisions by describing how these were linked to the vision. Although not academics, and largely unaware of the above mentioned authors, these leaders acted consistently with the advice in the literature. The level of commitment and accomplishment among employees was impressive.
Schumacher Constructing Vision with Scenarios ! PICMET 2012

The above authors are in general agreement that a vision is an ideal, an imagined future state that highlights differences between existing conditions and goals [19]. It offers solutions to problems. It provides the seven properties of sense-making [40] by offering a language that members use both to understand their situation and to communicate about it. A vision provides guidance in making decisions, yet is open ended enough to allow individuals freedom to choose how they will contribute and how they might elaborate the vision in light of changing circumstances. These authors also agree vision must address and endure for the long term to be effective. Like an organizational culture, and strategy, the large effort required to inspire many people to commit to a vision implies that the core of the vision must remain relatively stable if it is to be absorbed and implemented in executing significant tasks. Despite the importance attributed to vision in the management literature, there is relatively little advice on how to acquire vision. Conger's model - assessing the existing situation, then formulating a vision - offers an obviously correct framework, but falls short of offering a detailed process to guide someone in formulating a vision for an organization. The S2S scenario process described here is such a process. It offers the further benefit of testing a vision's viability. Communication Models The most widely known model of communication is Information Theory, sometimes called the 'S-M-C-R' model because it describes how a Sender transmits a Message through a Channel to a Receiver. It was developed by engineers at Bell Labs [33] as they worked on problems with telephone signals. The models assumes the sender and receiver are similar thus allowing any meaning the Sender wishes to communicate to be embedded in the message and ensuring the Receiver will be able to extract that meaning from the message. When the focus is on telephones or computers, the assumption of similarity is very accurate. Protocols are well defined and information is reliably sent and intensions interpreted. Anthropologists studying diverse cultures found the assumption of similarity to be inaccurate. They gradually came to adopt an "assumption of difference" [3] as they encountered peoples who understood differently such basic dimensions of experience as space and time [13, 14]. Spradely [33] comments, "All knowledge depends on categorization… In contrast to popular opinion, categorization is not a discovery of the natural groupings of objects in the environment. It is, rather, an invention of ways to classify and organize experience." Constructivism is a communication theory which sees interacting people as creating shared realities as they construct and exchange their language symbols [2, 21, 35, 38, 40]. Anthropologists developed ethnographic methods with emphasis on the nuanced differences language, and the corresponding meanings these communicate, within a group. The Constructivist model offers tremendous leverage for understanding how leaders impact organizations through vision. Leaders are not simply sending bits of information to receivers. Through communicated vision they postulate an alternative reality for an organization. Because of its incompleteness (in detail) and its distance in the future, people can accept the possibility of its potential existence. If they find the envisioned situation preferable to the status quo, they may undertake substantial efforts to achieve it. A leader seeking extraordinary performance can describe a vision that is worth the climb. There is broad agreement that it is the presentation of contrasts - the status quo versus a vision of an ideal - that is key to the motivational power of
Schumacher Constructing Vision with Scenarios ! PICMET 2012

vision. Indeed Schein [29] frames the task for transformational leaders as changing the balance between followers' anxiety about existing problems ("survival" pressure to undertake change) and their "learning anxiety" about the uncertainty they face in undertaking change (willingness to abandon the known situation and reach for the unknown). Building Shared Vision The importance of organizational vision is widely accepted, and we can conceptualize how vision acts to improve organizational effectiveness, yet the development of a shared vision within an organization remains problematic. Where does one find a viable vision? How does one convince stakeholders of its necessity? How might a particular vision be tested to build assurance that it can deliver its promise? Westley and Mintzberg [39] describe five examples of leaders who successfully used their vision to change organizations. They describe the variety of circumstances in which those leaders acquired their organizational vision and conclude there is not a single pattern. Their 'salient capacities' that surrounded vision creation include inspiration, imagination, foresight, and sagacity, and different leaders had different combinations of these. They question the effectiveness of what they label the "Linear Needle" model in which the leader with a vision injects it into an organization. They indicate a number of weaknesses in this common approach, not the least of which is an absence of any mechanism to develop a vision. They suggest an alternative "Drama" model with the leader deeply engaged, responding in real time, producing a heightened level of commitment. Their drama model has much to offer for getting a vision adopted, it offers less insight into its creation. Mitzberg's lengthy critique of planning [24] discusses several fallacies which have limited the contribution of strategy to organizational effectiveness. Formal planning methods have produced fragmented analysis, only incremental change, risk-adverse investment, and detachment of workers from organizational goals. Mintzberg labels this pattern "management by remote control". He believes it has driven out integration, consistency and, especially, commitment, attributes he believes are critical to successful planning. He suggests that "crafting a vision" could overcome the fallacies of formal planning methods. A vision - as suggested above - reaches beyond incremental change, offers an integrated perspective, and invites emotional commitment to the goals it outlines. Beech [1] offers a number of insights in his discussion of the relationship of vision and planning. His definition of vision is consistent with the above authors. Vision is the sense of direction, the agenda that the organization perceives itself to be pursuing, "vision defines the ideal future". Vision is a dream, the steps to the dream are vague in vision. This incompleteness allows individual creativity and contribution and Beech sees this incompleteness as a key attribute, "it is this quality of being within the organization's grasp, but requiring serious effort, that allows the vision to inspire, motivate and unite the organization’s members." Plans are also incomplete and achievable, yet are usually not described as inspiring. Beech observes an important process attribute is balancing the amount of constraint. "The process must be structured or it wanders, and then meaningful results are difficult to get. But, at the same time, it has to be freewheeling, or it fails to generate anything of much value." Mintzberg's critique [24]

Schumacher

Constructing Vision with Scenarios

!

PICMET 2012

argues that much corporate planning is too structured and that this suppresses strategic thinking and innovation. "Planning may impede itself from functioning as its proponents claim it should." Scenario Planning Scenario Planning has been used in businesses at least since the 1950s and the technique is widely used in military and disaster planning. At the core, the various scenario approaches support defining possible futures. There are many variations in applying the technique with the common benefits suggested by scenario proponents including, stretching participants' mental models, supporting creativity, and surfacing assumptions [22, 23, 25]. Building a shared language to discuss strategic issues is also frequently mentioned [27], an advantage also attributed to shared vision. Wack [36, 37] described his work at Shell Oil and explained how its use caused managers to consider a broader set of alternate futures and think through possible responses. He claims this allowed the company to adjust more quickly to the OPEC oil shocks of the 1970s. Jefferson [16] has recently revisited Shell's experience. Wilson [43] defines scenarios as "alternative plausible futures", clearly indicating that they are not predictions. Scenarios are "significantly, often structurally, different views of the future", a point that is stressed in the S2S approach. For Wilson, scenarios are "not variations around a midpoint base case". Finally, he sees scenarios to be "specific decision-focused views of the future". This suggestion to narrow the arena for scenario construction is a key insight which much scenario planning fails to appreciate. Discussion of this point is expanded below. More broadly, Wilson believes scenarios are a tool to develop a "strategic culture", which he sees as as more advanced and more effective than the culture of "traditional, long-range planning". This distinction is a step toward alleviating some of the weaknesses Mintzberg finds in most planning processes. Facilitation techniques are offered below to ensure a process that achieves the results Wilson advocates. Coates [8] is a long time contributor to the strategic planning and futures field. He notes that one of the biggest divides among scenario practitioners is whether or not scenarios should be seen as forecasts. In approaches where scenarios are taken as forecasts, the immediate problem is that countless possible scenarios can be created. How to select from this multitude is a challenge, and these approaches often include attempts to assign probabilities to the constructed scenarios, thereby reducing the focus to a manageable number [12]. Under conditions of Risk (all possible states and their probabilities can be known) efforts to define scenario probabilities and outcomes may allow calculation of likely results. This could be a valuable input to decision making. This author's industrial scenario experience has always encountered conditions of Uncertainty (unable to define all possible states), and therefore attempts to assign probabilities was always a distraction from capturing what benefits the scenario process could offer. Further, human judgment is the standard input to probability assignment, and here our abilities are wanting. Consider a brief example: About 5 years ago two students graduate students in a Strategy course selected GM as their topic and developed scenarios with a 10 year time horizon. (These students have been working full time for a decade the automobile industry.) They were made aware that The Economist had mentioned the possibility of a GM bankruptcy, however they did not include this in any of their scenarios. They did consider oil prices a major
Schumacher Constructing Vision with Scenarios ! PICMET 2012

scenario factor and concluded that $100 per barrel was the upper bound that they needed to consider. Only half of their forecast period has passed, and both critical events have occurred. These were very good students with years of work experience in the industry. Their performance is typical, we are not good at estimating the likelihood of many events. Assigning probabilities for future events appears easy, but achieving even tenable estimates is uncommon. Consistent with this observation, Brooks [5] comments on the incredibly poor estimates made in planning software project schedules, these typically have an operations timeframe much shorter that strategic issues. He notes that our estimates are chronically weakened by "gutless estimating", overly-optimistic thinking, and by the pressure to reduce the uncertainty we face in strategic planning. Reducing the uncertainty we acknowledge in our plans does not reduce the uncertainty we face in reality. The S2S Approach The Scenarios-to-Strategies approach (S2S) does not consider scenario probabilities, instead it moves away from the central-tendency thinking they imply and addresses possibilities the boundary conditions which can best test strategy viability. As a simple analogy, consider testing the operating range of an electronic device which is designed to function between zero and 100 degrees. Central tendency thinking might lead to testing at 'the most likely conditions', perhaps room temperature. Approaching the task from a perspective of exploring boundary conditions leads to testing at 0 degrees and at 100 degrees. The exploration of boundary conditions is central to the S2S and is termed a 'possibility' approach, in contrast to the common 'probability' approach. It emerges from two observations, that we are poor at assigning probabilities to critical future events, and the pragmatic need to limit the number of scenarios considered. In testing the electronic device, we do not need to test at 1 degree and at 2 degrees and at 3 degrees... if we test at the boundaries (0 and 100 degrees) we are relatively confident of performance between these endpoints. Defining these boundary conditions greatly narrows the number of scenarios that must be considered while avoiding the weakness of a probability approach. Probabilities and related scenario construction issues are summarized in Table 1. Scenario Attributes and Levels - Factors and Choices "Strategic Planing at United Parcel Service" [11] describes their scenario planning process. In the 1990s they realized their environment was changing. The CEO initiated a series of steps over a few years, including consultant-facilitated scenario workshops. Management developed a 2 x 2 matrix that captured their assessment of the future in 4 scenarios (Fig 1.) This is an example of what will be called here "level 1" scenarios. The 2 x 2 matrix is a common format for displaying scenarios. It is organized using two dimensions, what UPS termed "axes of uncertainty", here "Market Environment" and "Demand Characteristics". The 2 x 2 format limits the process to 4 scenarios, though the details of each scenario have some flexibility to accommodate variation. The larger process constraint is that only 2 dimensions provide the over-arching definition of possible futures.

Schumacher

Constructing Vision with Scenarios

!

PICMET 2012

Table 1. Comparison of scenario approaches.
Level 1 Scenario approach Level 2 Scenario approach (S2S)

Envisioned “Scenarios” include both external trends & Scenarios are constructed of ‘Factors’, forces organization’s decisions. Distinguishing the beyond the organization’s control. Strategies futures relative impact of decisions is problematic. are sets of organization Choices. Combining a Strategy with a Scenario yields one outcome (possible future). Probability Participants often asked to estimate scenario likelihood. Participants seek to find, then focus on a most probable future. This often results in 'surprise' when another future emerges. Discussion of probability suppressed; the goal is to define boundary conditions, ‘extreme possibilities’ which are explored in ScenarioStrategy confluence.

Process goal

Central Tendency - the process goal is Boundary Exploration - Factors & Choices are often determination of most likely scenario. defined as dimensions; process focus is the end points of each dimension, the inputs to Scenarios & Strategies. Sharing data & opinions is the primary level of learning for participants. Conflicts emerge when different views seek to agree on what is ‘most likely’. Conflicting views are encouraged in exploring boundaries. Potential participant conflicts are minimized as facilitators push boundaries to extremes needed in creating Scenarios & Strategies. Each strategy is tested against each scenario and the result is rated using the criteria. This offers a second level of team learning (beyond sharing data) when insights emerge. Evaluating each Scenario-Strategy combination yields one possible futures (typical S2S has 10 to 20). Criteria are used to rate these. Participants then select strategy.

Conflicting views

Strategy test

None (no separate strategy defined)

Criteria

Not needed; production of ‘most likely’ scenario defines planning future.

A key characteristic of level 1 scenarios is that there is no distinction between external Factors over which the participants have no control (e.g., energy prices, competitor actions, terrorism), and the specific Choices the planning entity makes that comprise potential strategies (e.g., investing in technology, concentration in specific segments, or merging with another entity). The four quadrants in Fig. 1 represent what could be termed "Outcomes", the result of the interaction of organizational choices with that subset of the external factors included in each 'scenario'. The Choices made in each quadrant are assumed to be 'reasonable' sets of decisions that best fit the unfolding future. Because we observe a different set of choices interacting with a different environmental context in each outcome, it is difficult to assess the relative influence of the Choices.

Schumacher

Constructing Vision with Scenarios

!

PICMET 2012

The blurring of Factors and Choices is reflected in the participants' comments about their 'scenarios'. "We all liked... the Brave New World scenario, and didn't like... Tangled Paths". The reality of organizational life is that factors beyond an organization's control substantially influence outcomes. (e.g., Airlines would like fuel prices to be low, but liking this does not make it so.) It is through strategic choice that organizations attempt to optimize their outcomes - their future performance - in forecasting and adjusting to the constraints imposed by the significant external factors that define their environment. When Factors and Choices are not separated in scenario planning, the superiority of some Choices over others is not obvious, although this is the core of all strategic planning effort. The assessment of the relative strengths of strategic choice can be a key benefit of Level 2 scenario planning processes.

Figure 1. UPS 1997 scenarios. [11] The Factors and Choices distinction is at the core issue defining scenarios approaches, What are scenarios? The literature often frames the question: Are scenarios to be considered forecasts? Or only different possible futures? The S2S offers a third option: Scenarios are to be used to test Strategies, after which the Scenarios can be discarded. The ultimate goal is defining and selecting Strategy. A second weakness of the Level 1 approach is the critical issue of the timing of strategic actions, what Hambrick calls Staging [15]. Sequence of action matters. If execution of some choices requires substantial lead time, a Level 1 approach assumes sufficient decisions were made early enough to allow for the scenario outcomes. This type of process tends not to show a situation in which preparatory work was not foreseen, and at a later date, 'best' choices were not possible. (For example, a disruptive technology emerges that redefines markets. By not investing early, the firm loses its position and risks failure; think RIM and the touch screen phones.)
Schumacher Constructing Vision with Scenarios ! PICMET 2012

The UPS case indicates their benefit from the Level 1 scenario development process. Benefits included development of a common vocabulary, a guide for investing in growth opportunities. Of course there may have been specific strategies that emerged that are not revealed in the case. A "mind-set shift" was a subtle impact that some participants experienced, acknowledging "we got much richer outcomes than if we had only done competitive analysis...without scenario planning, you don't see the big changes". Scenario processes offer many benefits and the claim here is that certain scenario processes offer greater benefits. Capturing those benefits requires elaboration of the process. The distinction between Factors - beyond the control of the organization (typically external), and Choices - options which the organization does control, is a core attribute of what is here labeled "Level 2 scenario approach" in table 1. This distinction is not new [20] but apparently has not been widely adopted as Level 1 scenario processes are commonly reported. In this author's electric utility planning work, Factors included primary energy prices, interest rates, regulations. Choices included selection among available technology (coal, gas, wind), timing and scale of projects to add capacity, and lobbying efforts. (These examples only a subset of Factors and Choices addressed.) The set of Factors considered is brought together to formulate Scenarios, environments which the organization may face, but which it cannot choose or even influence. Numerous Scenarios can be created from a list of Factors. As discussed above, to reduce the large number of possible Scenarios to those of greater usefulness, the S2S method seeks to define those Scenarios that capture boundary conditions, and avoids attempts to estimate the most probable scenarios. In similar process, sets of Choices are the inputs for creating Strategies. Again expediency directs that the total number of Strategies considered should be reasonably small. Exploration of boundary conditions again helps pare the number of possibilities. In addition, one hopes for some level of internal consistency within strategic choices in a Strategy and this can eliminate some possibilities from consideration. In Scenario development internal consistency offers a weaker screen. We might eliminate a Scenario with a booming global economy and oil prices over $200 a barrel as the high energy price would act to suppress economic growth. However there are many instances when seemingly 'inconsistent' scenarios occurred, that is, we were surprised (e.g., Stagflation in the 1970s, the continued robust growth of China after the 2008 crisis, the failure of the Russian economy to growth with the arrival of democracy when the USSR disbanded.) Boundary conditions remain the stronger screen to reduce the number of Scenario possibilities to consider. Participants often have specific Strategies they wish to test and this interest is another filter to reduce the number of Strategies to consider. (This consistent with Wilson's [42, 43] suggestion that scenarios should address "specific decision-focused" contexts.) Placing one Strategy into one Scenario yields an Outcome which may look similar to what are called scenarios in Level 1 approaches. Importantly, the Factor/Choice separation yields critical benefits. First, participants will fit alternate Strategies against a Scenario int the S2S to see if a preferred Outcome can be accomplished. Alternatively, individual Strategies will be fit against a number of Scenarios to assess the resiliency of those Strategies. The S2S approach explores both of these, evaluating each Strategy against each Scenario, developing an Evaluation Matrix which allows participants to develop a sense of the strengths and weaknesses of the
Schumacher Constructing Vision with Scenarios ! PICMET 2012

Strategies. This benefit is discussed by Sharpe [34] in using scenarios in NPD planning. To support this evaluation, Criteria are developed by the participants and used in the assessment of Strategies. The process flow and elements of the S2S approach are shown in Figure 2.

Choices alternate actions available to the organization

Factors forces outside the organization's control

Criteria for judging strategy success

Strategy
(set of choices)

Scenario
(set of factors)

Set

of

Scenarios

A
Set of Strategies

B
Evaluation Matrix

C

1 2 3

Test each strategy, in each scenario, using all criteria.

Figure 2. Elements of the Scenario-to-Strategy-(S2S) approach. Typical S2S planning efforts have created 3 to 5 Scenarios and 2 to 5 Strategies. Participants have developed 3 to 5 Criteria (e.g., marketshare, earnings, customer satisfaction). Completion of the Evaluation Matrix would require the participants to make perhaps 4 x 4 x 4 = 64 judgments. A full day is usually set aside for this process and the accompanying discussions generate insights and continue the process of building a shared language among participants. The Social Process "Plans are nothing. Planning is everything" - Napoleon A number of authors have described various scenario methods at various levels of detail. Wilson [44] breaks the process into 18 steps, Linneman [20] has 10, and Nowack [23] has 3 steps in "Scenario Development". (Nowack also offers 3 steps in "Scenario Transfer" i.e.,
Schumacher Constructing Vision with Scenarios ! PICMET 2012

implementation, issues often seen as beyond the scenario process and within the realm of leadership and organizational change). The S2S is broadly consistent with each of these authors and the intent here is to offer additional insight into facilitating process success, along with highlighting method contrasts that differentiate the S2S. There are several goals when facilitating the development of Scenarios and Strategies. The development of a viable strategy, along with high levels of understanding of that strategy and commitment to it, are the primary goals. To support these, the facilitator wants to create a group process in which participants develop a common language for exploring the issues and causal forces influencing the organization. The process should capture the available data, filter and structure it to reveal any insights that come from 'connecting the dots', and result in agreement on strategy selection. Face to face interaction is central to development of a common language and is the core of the S2S approach. S2S development acknowledges that communication operates at 2 levels, relationships among participants, and the content of issues, data and decisions. Watzlawick [39] observes that the relationship level always dominates the content level and this suggests the need for attention to facilitation of the scenario process. Selection of participants, setting expectations, and support for the evolution of the content are all important aspects of facilitation. Participants & scheduling Participant selection considerations include familiarity with relevant content and organizational position especially regarding implementation. Diversity of function and viewpoint is regarded as a goal in most scenario processes because search for creative organizational strategies is a common goal. Group size influences the process, larger groups take more time to discuss issues and even in situations with minimal time pressure, participants become anxious when they perceive slow or no progress toward completion. I have conducted the S2S process with groups ranging from 7 to 25 participants. Groups larger than 15 challenge facilitation because progress is slower and has greater potential for conflict or loss of focus. The increase in the number of possible participant interactions is combinatorial so careful consideration is warranted in contemplating larger groups. When the fraction of participants who are familiar with a scenario process is large, facilitation is easier. The common situation in which no participants have much scenario experience is more difficult and larger group size compounds this. One process variation with large groups is to begin with a single group to discuss and define issues, Factors, Choices and Criteria, then divide the group in two, one subgroup developing Scenarios from the Factors, the other developing Strategies from the Choices. The entire group is brought together for creating the Evaluation Matrix. Using two half-size groups to separately develop Scenarios and Strategies is a much easier facilitation task. This can promote something of a competitive gaming mindset for participants which was very positive in the few trials observed. The subgroups can be run in parallel if sufficient facilitation is available. A final consideration relating to group size is scheduling and attendance. Larger groups make scheduling more difficult. Further, attendance is critical to the development of shared understanding. If a participant misses a single session in which a major issue is resolved, then questions the group's resolution when rejoining, the loss of momentum can be catastrophic. The
Schumacher Constructing Vision with Scenarios ! PICMET 2012

working out of shared definitions and understanding can seem slow and painful to participants. When hours of work is undone by the questioning of an issue believed settled, participants begin to question the process and enthusiasm is depleted. The first line of defense is for the facilitator to set expectations for attendance when determining membership. I explain the "2 strikes" rule in the beginning. If a participant misses 2 sessions, they should withdraw from the process. A metaphor compares the development of a group's shared language and understanding with building Sand Castles on a beach. The wind and waves gradually erode anything built, there must be continual effort to build up a structure against those forces. Likewise, any shared understanding is subject to erosion until the process achieves some stability in its final structure. The facilitator can view their function as supporting the creation of components and knitting those together where they are mutually supportive, similar to how a house gains much more stability once the foundation, walls and roof are tied together. There are two opposing forces to balance in scheduling sessions. The Sand Castle phenomena argues that the entire process should be completed as quickly as possible. Ideally this would be a single long session in which there is no group turnover, no forgetting, no 'erosion'. The S2S projects I conducted ran from 36 to 90 hours of group time. Obviously too long for a single session. Sessions of 2 to 4 hours a day avoid fatigue, allow participants to gather information outside the group session that they realize is important to the discussion, and permit them to continue to respond to normal aspects of organizational life, email, phone calls, etc. so the process is not too disruptive to other work activities. Creating the Evaluation Matrix is best conducted in a single long session, typically a full day. This minimizes participants' perception that their judgments 'drift' during the evaluation, which can lead to some questioning of the results. (A facilitator may observe that changing perceptions among the group reflects learning, which is a central process goal.) The recommendation is for a single, long Matrix-development session. Conducting the entire process within a short calendar period remains a goal with 2, or at most 3 weeks being an achievable goal. Facilitation support techniques Facilitation activities that promote building shared understanding include timely and accurate record keeping, attention to nuances in language use, promoting group interaction, and dedicated space. Projects sometimes have a dedicated 'war room' which only the team uses. This allows collection of artifacts, such as completed flip charts, and a sense of comfort and ownership of the space. Conversation among some members may continue for some time after a session formally ends. Those conversations may be cut short if participants are forced to the leave the room to make way for another group. In general the facilitator wants to maximize participant interactions, both in-session dialogue on issues and out-of-session social interaction. Serving meals before, during or after sessions increases this interaction, as does asking participants to review and comment on session notes prior to subsequent sessions. A 'tag-team' style of facilitation is effective in most intensive group processes and recommended for the S2S, especially for larger groups. One facilitator is 'on stage', asking participants questions to stimulate input and clarify contributions, answering questions about the process, and recording participant comments on white boards or flip charts. The other facilitator monitors the process at a 'meta-level', looking for misunderstandings, inconsistent language use,
Schumacher Constructing Vision with Scenarios ! PICMET 2012

ascertaining group perceptions and making few, but key suggestions such as calling for a break or recommending small process changes. This person can play the lieutenant role effectively, offering process observations from within the group. The on-stage facilitator is often fully engaged in their tasks and may not notice deeper patterns. The facilitators occasionally exchange roles as the process progresses, if the on-stage facilitator tires, or to bring a different personality to the present task. The monitor role may be able to accomplish real-time record keeping, capturing and structuring the group discussion on a lap top, taking photographs of diagrams constructed and other participant contributions. It is desirable to have a separate, dedicated recorder in order that session notes can be distributed to the participants within a hour or so after each session. Content Emerges in Process There is a general pattern to the development of the content. The discussion begins with listing and defining issues facing the organization. The facilitator may encourage discussion at this level for an hour or more, or until participants are satisfied they have mentioned the key issues. Special attention is paid to trends, is there patterned change is issues over time? (e.g., Are regulations becoming more stringent? Is R&D productivity falling? Are customer expectations changing?) With a large number of disparate issues the facilitator may ask if some subset may be put aside to be addressed is a separate project (e.g., in utility planning, the group mentioned, then suspended any discussion of deregulation, of diversification into non-utility businesses, and concerns about human resource capabilities that might limit strategic choices. The focus was on adding capacity in a period of rising cost and rising environmental concerns.) Another focus for the facilitator is to uncover decisions that concern the organization. A repeated group pattern is open 'brain-storming' of content until that process slows, then examining the list in an attempt to meld similar or redundant items, or place some items within broader issues. If the list remains large, more than a 12 or 15 items, then participants are each given several votes (e.g., 4 votes with a list of 12 items) and they assign their votes to the issues they consider most important. This typically allows the list to be reduced in size. Facilitators ask if it is acceptable to the group to remove items that received few or no votes, careful to keep any items that someone is unwilling to drop. This pattern of Generation, then separate Evaluation, (G/E) has a base in our understanding of human psychology. Once a list of issues is defined, the facilitator asks these to be classified into Factors issues or forces the organization cannot control, or Choices - items where the organization makes decisions or allocates resources. (For the utility, oil prices and interest rates were Factors. Generation technology, timing of plant additions, and location were key Choices. For a software research team, the type of operating systems in use and hardware capability in the future were Factors. Choices included functions to include in their next release, decisions on whether to pursue collaboration, and which industry standards to support.) This distinction in not always obvious and the discussion should refine the group's understanding. For example, a group may decide not to invest in a particular new technology, but another organization may do so and the technology development could accelerate. The Choice is not to develop the technology, but only to invest in it. The Factor is that the technology may, or may not, be developed by others regardless of our investment choice.
Schumacher Constructing Vision with Scenarios ! PICMET 2012

It is necessary that the Factors and Choices must 'fit' for scenario planning to produce meaningful results. If the software teams includes oil prices in their set of Factors, and their Choices address adding pixel manipulation functions to their program, there is insufficient fit. Participants may be concerned about rising oil prices, but it has no meaningful impact on their program features. To achieve fit, the facilitator begins with a discussion of Choices using the generation/evaluation process. Then moves to a discussion of Factors. A third discussion surrounds what Criteria the organization will use to evaluate its success in the future. Indicators such as sales, profit, market share, product innovation and customer satisfaction are commonly used Criteria. Similar to the Balanced ScoreCard [18] approach, 3 to 5 Criteria are common in the S2S. To achieve good fit, the process iterates, discussing Choices, then Factors, then Criteria, then back through Choices, Factors and Criteria now with more shared understanding of the other components. Each of these topics may take an hour or two in the first iteration, maybe less in the second. Language and boundaries The Constructivist communication perspective sees the essence of language, the naming of experience, as creation of and negotiated agreement on boundaries. These boundaries are made very explicit in defining ranges for Factors, Choices and Criteria. During all S2S discussions the facilitator must be sensitive to nuances in participant language. Individuals will use terms ("categories for experience" for Constructivists) whose definitions may not be shared by all of the group. The facilitator should focus attention to work toward a shared understanding of key terms. This can be asking for definitions, asking if there is agreement, and if necessary suggesting special definitions that the group will adopt for the duration of this project. (Example: the utility group discussed "a good economy" implying this had a positive impact on the firm's performance. The economist had a very specific definition, but others were vague. Did "a good economy" mean low unemployment? Or rapid regional GDP growth? Or record new home construction?) The common tendency is for participants to not notice subtle differences in terms or to let any differences pass to avoid appearing disagreeable. The development of a deeply shared vision is inhibited if such differences are not resolved. Language is the medium for constructing the mental models that participants will use to extrapolate how changes in a Factor or Choice will influence an outcome. If there is insufficient language coherence among participants, there will be impassable disagreements when they begin the Matrix evaluation. Once the group is comfortable with their list of Factors, Choices and Criteria, the next step is to define likely ranges - high and low values - for each. The easiest to understand are those that have simple numerical indicators. Examples include oil prices or interest rates. The group must agree on what time frame their scenario effort will address. (Assume 10 years for discussion purposes - What will be the pattern of oil prices over the next 10 years? Will it gradually rise to a maximum of $130? $170? Will it decline for several years (how far?) and then rise? Will it oscillate between $90 and $100?) There will often be substantial differences of opinion among participants on such questions. These could become contentious, but the S2S process requires that participants define the high and low boundary conditions for each Factor, not that they agree on a 'most likely' outcome. The process can thus diffuse conflict over 'what will happen'. Instead they seek 'reasonable' boundary
Schumacher Constructing Vision with Scenarios ! PICMET 2012

conditions, 'what could happen'. The general goal of the facilitator is to push the group to select more extreme values for range endpoints, especially further into the future. The student example above typifies our psychological limits. They could not accept (in 2002) that oil prices would surpass $100 per barrel in a decade, nor did they think a GM bankruptcy should be addressed in their scenarios. Humans are not particularly good at addressing the uncertainties they face looking into the future. The S2S compensates for this shortcoming by driving participants into 'thinking outside the box' about the future. This is a point that usually requires facilitator forcing. Some Factors and Choices are not easily defined on a numeric continuum. Some are best modeled as discrete events - a Republican president after the election or a Democratic president can be a two-state range for a Factor "Presidential outcome". Multi-state discrete ranges are often used in the S2S. (Factor states for the Software team: 1. Operating Systems remain like today, 2. A few distributed languages emerge but only in specialized niches, 3. An explosion of distributed languages cause a major shift in application architectures.) In the Level 1 scenarios there are two primary dimensions the define the future (usually a blend of Factors and Choices). In the S2S the number of dimensions for Factors and Choices is limited only by the willingness of the participants to commit the time and effort to address them. Practical considerations usually produces 4 to 10 Factor and slightly fewer Choice dimensions. Three to Seven Criteria have been observed though three is common and usually sufficient. These numbers still represent tremendous variety and the formulation of Scenarios and Strategies allows this complexity to be addressed efficiently. The reduction process is the same for Factors into Scenarios and Choices into Strategies, only the content is different. Two rules guide formulation of the Scenarios and Strategies. 1. In defining each Scenario (or Strategy) participants must include a point on each Factor (or Choice) range. If the Factors are oil prices and interest rates, then each Scenario must have an oil price value somewhere along the defined range of possible oil prices, and an interest rate value somewhere along the defined range of possible interest rates. This would still allow an infinite number of possible Scenarios. Rule 2, the set of Scenarios must include all the endpoints of all Factor ranges at least once. Continuing the example, one Scenario could have the highest possible range of oil price and highest range endpoint for interest rates, and a second Scenario could have the lowest possible range of oil price and lowest range endpoint for interest rates. The two rules are met. Participants can, and usually do create more Scenarios to include other combinations, these usually include non-endpoint values on the ranges. Creating more Scenarios (or Strategies) makes completion of the Evaluation Matrix more time consuming as participants will address the product (evaluate each Strategy in each Scenario). Similar rules guide development of Strategies from the Choice dimensions and their ranges. A 'Business as Usual' strategy, built of the continuation of the current set of Choices is a good starting point for developing scenarios. I have yet to facilitate a group that concluded this is their optimal Strategy. Typically participants develop 3 to 7 Scenarios and 2 to 5 Strategies (so Evaluation Matrices are typically 5x3 or 4x4 or 6x2, etc.) The variety of the numerous Factor and Choice dimensions are maintained but the effort required to assess all of these is diminished by aggregating Factors into Scenarios and Choices into Strategies. In part the justification for this reduction is that ultimately we are concerned with the boundary Scenarios (We don't test at zero degrees and at 1 degree and at 2 degrees .. to 100 degrees, we test only at zero and at 100, the
Schumacher Constructing Vision with Scenarios ! PICMET 2012

boundary conditions.) Creation of the Scenarios can be guided by judgments of internal consistency - though we sometimes observe conditions such as Stagflation in the 1970s which previously were considered inconsistent. In the end Scenarios are scaffolding and after they have been used to evaluation the Strategies, their importance is greatly diminished. Strategies are different than Scenarios on this point. Internal consistency is a primary concern. A starting point can be to define the existing 'Business as usual' strategy the organization now pursues. Then substantially different Strategies can be around themes (e.g., decentralize operations, grow international sales) or in incorporating all Choice endpoints, new combinations can be described. More than one iteration at defining Strategies and Scenarios from the Choices and Factor ranges can be made, and this helps tighten their fit. After full discussion and acceptance of these, participants are asked to divide the Strategies and Scenarios among them and write 1 to 2 page text descriptions of each. These are shared and feedback requested. This supports greater agreement on their definitions. The final step in the S2S is to complete the Evaluation Matrix. A strategy is selected, and it is evaluated using the Criteria with the first Scenario. Participants discuss and vote on the results for this Strategy in that Scenario for each Criteria. This is repeated until each Strategy and Scenario combination had been evaluated. Usually this is done for two time periods (for the software team, it was 12 months and 24 months in the future. For the Utility, it was 3 years and 10 years.) This can be an exhausting exercise and disagreements often emerge. A well facilitated process will have prepared participants so that language differences, or intolerance of contrasting opinions, are not barriers to learning from the evaluation process. Patterns of Strategy performance in the completed Matrix are usually very obvious. Some Strategies are dominated by others. Some Strategies perform well further in the future. Some Strategies may fail in some Scenarios but excel in others. Usually the group concludes that one Strategy is preferred over others. Occasionally there are two subgroups who favor different Strategies. Groups often make minor modifications to one of the Strategies after observing how it performed across the Scenarios and then accept this as their preferred Strategy. Uniformly they agree that they understand with far greater richness the Choices they face, and the linkages between those Choices and the Factors that they perceive comprise their environment. Conclusions The S2S is one of many variants in the family of scenario planning approaches. It is primarily recommended for small groups and facilitation guidelines are offered here to enhance its effectiveness. The S2S distinguishes Factors and Choices as fundamentally different inputs to the future. It is unique in emphasizing Strategy creation and assessment and discarding the Scenarios after they have been used to evaluate the Strategies. It also avoids the practice of assigning probabilities to Scenarios, moving away from a central tendency approach, instead adopting a framework of possibilities in Strategy evaluation. The S2S accommodates a larger number of dimensions, and greater complexity, than is commonly represented in Level 1 scenario methods. It does so efficiently by addressing primarily the endpoints of dimension ranges, the most contrasting conditions which offer the starkest learning environment. The S2S
Schumacher Constructing Vision with Scenarios ! PICMET 2012

has been used in a number of organizational settings and participants report satisfaction with the process and results. The S2S process can produce a shared vision among participants, fulfilling a need commonly expressed in the leadership and strategy literatures. Adopting a Constructivist communication model offers guidance for the facilitation techniques that have been applied in its practice.

References
[1] Beech, Lee Roy, Making the Right Decision, Prentice Hall, 1993. [2] Berger, Peter and T. Luckman, The Social Construction of Reality, Doubleday & Company Inc. 1966. [3] Bennett, Milton, “Towards Ethnorelativism: A Developmental Model of Intercultural Sensitivity”, in Cross Cultural Orientation: New Conceptualizations and Applications, Paige, Michael (ed.), University Press of America, 27-69, 1986. [4] Bodwell, Wendy and Chermack, Thomas, (2010) "Organizational ambidexterity: Integrating deliberate and emergent strategy with scenario planning", Technological Forecasting and Social Change, 77, 193-202 February 2010. [5] Brooks, Frederich, The Mythical Man Month, 20th anniversary edition, Addison-Westley, 1995. [6] Burnside, Robert M., "Visioning, building pictures of the future", in Managing Innovation, Henry, Jane and D., Walker, Eds. Sage Publications, 1991. [7] Chermack, Thomas, "Studying scenario planning: Theory, research suggestions, and hypotheses", Technological Forecasting and Social Change, Vol. 72, No. 1, 59-73, January 2005. [8] Coates, Joseph F., "Scenario Planning (From My Perspective)", Technological Forecasting and Social Change, 65, 115-123, 2000. [9] Conger, Jay A. and R. Kanungo, Charismatic Leadership in Organizations, Sage Publications 1998. [10] Conger, Jay A., "Inspiring others: the language of leadership", Academy of Management Executive, V:1 31-45, 1991. [11] Garvin, David & L., Levesque,, "Strategic Planing at United Parcel Service", Harvard Business School case 9-306-002, June 2006. [12] Godet, Michel, "The Art of Scenario and Strategic Planning: Tools and Pitfalls”, Technological Forecasting and Social Change, 65, No. 1, September, 3-22, 2000. [13] Hall, Edward T., The Hidden Dimension, Anchor Books, 1966. [14] Hall, Edward T., The Silent Language, Anchor Books, 1959. [15] Hambrick, Donald and J., Fredrickson, "Are You Sure You Have a Strategy?" Academy of Management Executive, Vol. 15, No. 4, 2001. [16] Jefferson, Michael, "Shell scenarios: What really happened in the 1970s and what may be learned for current world prospects", Technological Forecasting and Social Change, 79, No. 1, September, 186-197, 2012. [17] Kouzes, James and B. Posner, The Leadership Challenge, Jossey-Bass Inc. 1987.
Schumacher Constructing Vision with Scenarios ! PICMET 2012

[18] Kaplan, Robert & Norton, David, "Using the Balanced Scorecard as a Strategic Management System", Harvard Business Review, January-February, 75-85, 1996. [19] Larwood, Laurie, F., Cecilia, M., Kriger, and P., Miesing, “Structure and Meaning of Organizational Vision”, Academy of Management Journal, 38, 3, 740-769, June 1995. [20] Linneman, Robert E., and J., D., Kennell, "Short-sleeve approach to long-range planning", Harvard Business Review, March-April, 141-150, 1977. [21] Maturana, Humberto, "Reality: The Search for Objectivity or the Quest for a Compelling Argument", in the Irish Journal of Psychology, special issue, 1988. [22] Masini, Eleonora, and Vasquez. “Scenarios as seen form a human and social perspective”, Technological Forecasting and Social Change, Vol. 65, No. 1, September 2000. [23] Nowack, Martin, Endrikat, Jan and Guenther, Edeltraud, "Review of Delphi-based Scenario Studies: Quality and design considerations", , Technological Forecasting and Social Change, Vol. 78, No. 9, p 1603-1615, November 2011. [24] Mintzberg, Henry, The Rise & Fall of Strategic Planning, Prentice-Hall, 1994. [25] van Notten, Ph. W. F., A. M., Sleegers, and M. B. A., van Asselt, “The Future Shocks: On Discontinuity, and Scenario Development”, Technological Forecasting and Social Change, Vol. 72, No. 2, February, 175-194, 2005. [26] Peters, Thomas and R., Waterman, In Search of Excellence, Harper & Row, 1982. [27] Peters, Thomas, Thriving on Chaos, Alfred Knopf, 1987. [28] Postma, Theo and F., Liebl, “How to Improve Scenario Analysis as a Strategic Management Tool?”, Technological Forecasting and Social Change, Vol. 72, No. 2, February 2005, 161-174. [29] Schein, Edgar, "Transformative Change: Unlearning & Relearning Culture", Chapter 6, The Corporate Culture Survival Guide, 1999. [30] Schoemaker, Paul, "Multiple Scenario Development: Its conceptual and Behavioral Foundation", Strategic Management Journal, Vol 14. No 3, March 1993, 193-213. [31] Schwartz, Peter, The Art of the Long View, Doubleday 1991. [32] Senge, Peter M., The Fifth Discipline, Doubleday 1990. [33] Shannon, C.E., & W., Weaver, The mathematical theory of communication. Urbana: University of Illinois Press, 1949. [34] Sharpe, Paul, and Keelin, Tom, "How Smithkline Beecham makes better resource-allocation decisions", Harvard Business Review, March-April 1998. [35] Spradley, James, The Ethnographic Interview, Wadsworth Group, 1979. [36] Wack, Pierre, "Scenarios: uncharted waters ahead", Harvard Business Review, 73-89 September-October 1985. [37] Wack, Pierre, "Scenarios: shooting the rapids", Harvard Business Review, 139-150 November-December 1985. [38] Watzlawick, Paul, The Invented Reality (Contributions to Constructivism), Norton, 1984. [39] Watzlawick, Paul, The Pragmatics of Human Communication, Norton,1967.

Schumacher

Constructing Vision with Scenarios

!

PICMET 2012

[40] Weick, Karl, Sensemaking in Organizations, Sage Publications, 1995. [41] Westley, Frances, and H., Mintzberg, "Visionary leadership and strategic management", Strategic Management Journal, 10, 17-32, 1990. [42] Wilson, Ian, “From Scenario Thinking to Strategic Action”, Technological Forecasting and Social Change, 65, No. 1, September, 23-29, 2000. [43] Wilson, Ian, The Subtle Art of Strategy: Organizational Planning in Uncertain Times, Praeger Publishers, Westport CT, 2003. [44] Wilson, Ian, and Ralston, W, (2006) Scenario Planning Handbook: Developing Strategies in Uncertain Times, Southwestern Educational Publications. [45] Zaleznik, Abraham, "Managers and Leaders: Are They Different?", Harvard Business Review, May-June, 1977.

Schumacher

Constructing Vision with Scenarios

!

PICMET 2012