Facilitating Local Ownership Through Paradoxical Interventions

Tough Issues

In this article, tough issues are defined as problems or challenges perceived as essential by those involved that, despite repeated efforts to solve them, often persist in organizations because of their complexity. This complexity is multifaceted. Mitroff and Sagasti’s (1971) “ill-structured problems” refer to content complexity characterized by ambiguity, multidimensionality, and dilemmas. Horn (2001) labels them as “social messes,” emphasizing the social complexity of multi-actor involvement, contrasting interests, and no central coordinating body. Duke and Geurts (2004) use the term macroproblems in recognition of contextual complexity because of institutional pressures, interrelated networks, and shifting belief systems. Churchman’s (1979) “wicked problems” point to the psychological aspect that can depress people, especially when they mistakenly strive to control such issues. In my experience, these different aspects often reinforce each other. Despite these difficulties, tough issues are potentially also the most rewarding to address, which is why Harpaz (2005) labels them “fertile questions.” They are often related to an organization’s primary process, where complexity of issues is greatest, and value is added to the outside world. They invoke professionals’ pride and passion. This upside is important, as dealing with such complexity requires active local participation.

Local Ownership

Tough issues often require a collaborative effort, working across tiers and organizational barriers in temporary activity systems (Engeström, 2004). Contributions from different actors are needed to understand and address the many aspects of the issues. This implies dealing with contrasting viewpoints, actions, and intentions—that is, the ability to create optimal conflict levels related to the task at hand (De Dreu, 2006). One learns about tough issues by addressing them, not by thinking about them beforehand. The latter leads to a “paralysis by analysis,” where people cannot take action until they have more information, but they cannot get good information until someone takes action. Thus, a process of “small wins” makes more sense: micro-level changes that actors enact as they make sense of and act in the world (Orlikowski, 1996). Design and execution are iterative processes, and change is shaped incrementally, incorporating emerging insights and contextual shifts. Each small win is a stable building block, often pointing to new ones, to be aggregated opportunistically into a robust change. When issues persist, the existing repertoire needs to be questioned by reflective practitioning: to unravel “espoused theories” and to uncover and revise “theories in use” (Schön, 1983). All such (un)learning takes place on the job. It underscores that the primary change agents addressing complex issues are those who deal directly and daily with them. The ideas and practices that need to be challenged often persist more widely within the organization, actively conserving the status quo and maintaining existing power balances.

The paradox of embedded agency is that doing what is deemed appropriate within such a context may be ineffective, but doing what works is not legitimate and will invite opposition (Seo & Creed, 2002). “Positive deviants” run the risk that their lack of conformity is not regarded as constructive by their peers. For this reason, De Geus (1997) emphasizes that innovations best take place at the margins of an organization away from prying eyes. Once success is proven in the margins, innovations can spread by organic diffusion rather than by institutionalization. This allows it to happen at a speed where the backlash of nonadopters is manageable, capacity can be built, and quality does not erode (Krackhardt, 2001).

Helping Relationship

Dealing with tough issues might be rewarding, but it also involves a lot of uncertainty. It requires unfamiliar repertoires, unpredictable processes, and arouses opposition (Brown & Starkey, 2000). Consequently, those involved often look for help from third parties. There is always apprehension about whether such help will be effective, especially as participants cannot picture the new approaches or assess the consultant’s ability well in advance. This creates a mixed message (Hendriks, 1987): Participants ask for help while being hesitant to accept it. What complicates matters is that both messages are rarely expressed simultaneously. A request for help might be on the table but the doubts and fears less overtly so. Discussion often restricts itself to content: the issues at hand, the suggested change approach, the consultant’s track record, and so on. Process aspects, such as the level of trust in each other, unease with power differences, and the awkwardness of unpredictable change, may create cold feet and remain largely implicit. Kegan and Lahey (2001) refer to these as “hidden competing commitments.” Handling mixed messages is part and parcel of helping relationships. It requires that both be addressed: the desire to innovate and the fear to do so and the explicit demand for help and the implicit doubt to accept it. Only in combination do they represent a sense of reality needed to deal with tough issues.

A common confusion is to regard the uncertainties as personal limitations to be dealt with privately rather than as intrinsic traits of complex problems to be dealt with at work. This is well recognized in the psychodynamic literature where French (2001) suggests that successful change depends on a combination of “positive capabilities” needed to deal constructively with the problem’s content and “negative capabilities” needed to deal with the associated emotions. In Gestalt, this is referred to as the paradoxical nature of change: Change occurs when the people involved abandon, at least for a moment, what they would like the situation to become and attempt to accept how things are (Beisser, 1970). It implies the need for paradoxical interventions that enable facilitators and participants to have two contrasting types of conversations simultaneously (Watzlawick, Baevin Bavelas, & Jackson, 1967).

This intervention paradox is recognized most strongly in consultancy approaches that regard the consultant’s role to be to empower clients to diagnose and address their own issues. This is at the heart of organization development practice (Bushe & Marshak, 2007) and process consultancy (Schein, 1999, 2010). Schein stresses that, to build a helping relationship, you have to be a process expert first and foremost to handle the unbalanced nature of that relationship. He mentions inevitable clients’ traps such as initial mistrust, attempts to look for reassurance, and hanging on to relief or resenting its absence. These can be reinforced by facilitators’ traps such as dispensing wisdom prematurely, reassuring clients, pressuring them into action, or keeping them at arm’s length. Processes of transference and countertransference can further frustrate the helping relationship by unconsciously making others responsible for one’s feelings. Process expertise allows for sufficient exploration of such inequalities, traps, and transference within the helping relationship. To this end, a “holding space” is created where feelings can be understood, filtered, and handled (Hirschhorn, 1988). This space is “contained” to prevent emotions from eclipsing the rest of the work. The helping relationship itself can be such a space, or—in mature teams—the group context itself. Within such relationships feelings can be examined effectively as soon as one of the parties does not act out his or her emotions and slows down his or her (re)actions. Typically, this is seen as a key skill for facilitators.

These theoretical notions, when taken together, suggest that a decentered agency is needed when dealing with tough issues, but it also puts higher demands on both the positive and negative capabilities of participants. This makes facilitating ownership more challenging. A cherished organization development notion such as clients’ free and informed choice to be helped becomes problematic, as clients cannot fully assess what is beyond their field of expertise (e.g., Bushe & Marshak, 2007). The same holds true for the ideal that clients take full responsibility for their actions, which requires repertoires of which they are not aware. Heated discussions over what this might mean for effective and ethical facilitation (e.g., Worley, 2009) inspired the following research question: How does the helping relationship enable or frustrate participants’ ownership throughout the process of dealing with tough issues?

Research Setting

The Dutch Ministry of Foreign Affairs has been the channel through which the Dutch Government communicates with other governments and international organizations. It carries out Dutch foreign policy ranging from promoting international stability and furthering European integration to assisting poverty reduction. The Ministry employs 3,000 people—many for their entire working life. The headquarters are located in The Hague where most of the staff (about 60%) work. The rest work abroad in one of the 155 embassies, consulates, and permanent representations. The organization has a well-developed “esprit de corps”: a sense of prestige associated with diplomatic service.

At the turn of the millennium, the Dutch government shifted its policies with regard to foreign aid to developing countries in line with discussions in multilateral platforms. It wanted to move away from supporting a multitude of small projects in a multitude of countries with Dutch funds and Dutch agencies. The government concluded that such projects were more viable when executed by local agencies, with active involvement of the civil society and embedded in regional/national policies and programs of the recipient country. Recipient countries were not happy with the bureaucratic burden of dealing separately with dozens of donor countries, each with their own agendas and requirements. The new policy proposed to pool donor countries’ funds to support countries to manage their own policies, programs, and projects. This would cut down transaction costs and make efforts more sustainable, hopefully bringing closer the fulfillment of the Millennium Development Goals.

In 2000, the heads of development cooperation (HDCs) of three embassies, together with the Deputy General of Development Cooperation and his “think tank” in The Hague, held informal discussions on the implementation of this policy. I was invited to join as an external consultant. The policy implied a shift in how embassy staff needed to think and act. They would need to move away from being experts in terms of project management or content (as in AIDS) to being process experts, getting multi-actor arenas to collaborate and local agencies to take the lead in addressing tough issues. The dominant approaches toward change at the Ministry could be characterized as “policies and procedures” directed from the top, respectively changed by political and rational–empirical strategies. The experiences so far led us to conclude that these two change strategies would not be effective. A normative reeducative strategy looked like a more congruent choice, and the three HDCs volunteered to experiment at their embassies. The initiative was labeled “learning & development” (L&D) and developed into a tapestry of interrelated activities over a period of 6 years that can be regarded as “action research.”

Overview of Activities

To help people experiment with tough issues implied an action orientation. It warranted a research orientation as both the tough issues and the change dynamics puzzled people. Over time, an internal network of “organizational change agents” within the Ministry wanted to learn to facilitate their own changes, which also implied a developmental orientation. In short, change efforts, academic research, and professionalizing were all needed (as well as managing the organizational context). All three were highly participative processes that were coupled throughout the 6-year period. This was action research being done within, about, and for social systems.

Some have pointed out that these three types of activities cannot be integrated well as they represent very different orientations, creating contrasting loyalties to, for example, academic rigor versus practical relevance (Schein, 2010). Others plead that action research can and should combine contrasting orientations but in line with dialectic theory, not by balancing them but by using the tensions between them (Vermeulen, 2005). Instead of integrating the three types of activities into one monolithic endeavor, this can be achieved by loosely coupling divergent types of action research of which there are many available (e.g., Reason & Bradbury, 2001). Some people (like myself) would participate intensely across these several arenas, whereas others more selectively. Experiences and expertise from one arena would feed activities in another. I regard such an interrelated tapestry of activities as a purposeful way of doing action research, cocreating with participants how to best serve contrasting aims. Figure 1 shows the types of activities undertaken in the 6-year period:

OPEN IN VIEWER

Figure 1.

Overview of arenas and activities

Research Steps

During the 6-year research project on change dynamics, we collected extensive data, ranging from transcriptions, participative observations, and reflective notes to a paper trail of evaluations, mission reports, emails, surveys, and so on. The data from the first four embassy projects in Pretoria, Sana’a, Lusaka, and Addis Ababa were collected and interpreted in great detail. The data collected elsewhere were more selective, focusing on “rich descriptions” of “meaningful events”: events pinpointing repetitive dynamics enabling or frustrating addressing tough issues. A few hundred of such descriptions formed the core of the data. Each change project and each meaningful event was coded, and contrasting interpretations were sought, which were written down in reflection documents. Out of these documents emerged elaborated sensitizing concepts.

Once the first four embassy projects were analyzed in detail, the list of concepts was refined and used to examine the wider data set to identify which of them appeared to explain the data most convincingly. This lead to about 13,000 coded passages and about 250 reflection documents. What emerged were key concepts grounded in data (Strauss & Corbin, 1998). We looked for the interconnections between these concepts and went back to the data to test them. A literature review allowed us to connect the concepts to theory and sharpen our reasoning.

Throughout this process, triangulation was sought (Eden & Huxham, 1996). Different people collected data, selected meaningful events, made rich descriptions, interpreted and coded the data, and conceptualized findings. We contrasted and discussed their input at each stage. Preliminary findings would find their way back to change settings where their resonance and actionability could be tested. We moved the research forward when we felt saturation was achieved and change dynamics could be explained sufficiently.

Each step of the research was discussed, shaped, and documented with those involved but not with everybody to the same degree. Researchers and facilitators were involved in all the research steps throughout the whole period, from framing the research question and collecting and interpreting data to gaining insight and ability to deal with complex change. Academic peers were only consulted to challenge our interpretations. The research question resonated with participants and champions, who helped collect and interpret data to further their own insight and ability, but their contribution limited itself to the 1- to 2-year duration of their own change process. They engaged the research question primarily for pragmatic reasons: to figure out which specific actions would help their situation. The facilitators and especially the researchers engaged the research questions more to figure out generalized change dynamics.

Initiation Paradox: Demand-Driven Versus Supply-Driven

At first glance, a demand-driven facilitation appeared desirable when we wanted people to take responsibility for their own development. They could state what issues mattered to them, what they wanted to learn, and the type of support they would like and commit themselves to all three. However, tough issues cannot be defined clearly beforehand, let alone how to tackle them. Furthermore, any innovative approach would break with the organization’s history, implying that they might overlap with blind spots or allergies of those involved. This would make participants hesitate, and they would shift their attention to simple problems instead. This often concerned dissatisfiers. People would agree that “someone” should do something about such problems, but nobody would be eager to put much work in and, by definition, motivation would decrease as soon as the symptoms lessened. They sometimes tried to resolve this by asking us for quick fixes, such as give us “a best practice to copy” or “a course in better communication.” Predictably, such fixes have limited innovative power. They might do some good for simple problems but do little to tackle tough issues, and, in either case, participants’ motivation remains low. This means that any facilitator who is hung up on a “demand-driven approach,” leaves participants to their own devices, refusing to help them address their tough issues and blind spots. Let me illustrate this dynamic with the L&D start-up in Yemen:

The Dutch HDC in Sana’a telephones with a request to start an L&D process. He says there are lots of real issues, such as a shift away from small Dutch development projects to larger sector programs managed by Yemen, which is difficult given its weak institutional context. It also poses challenges for the embassy: Policy departments and control departments would have to work more closely and strategic thinking would need to take precedence over everyday operational pressures. When we arrive in Yemen, enthusiasm seems to have lessened. The evening before the kick off the HDC asks us if we can do the start-up in half the time: everybody is busy, other change efforts are demanding time, and he feels unsure about the program. We react strongly and, consequently, he takes his request off the table. At the opening the next day, the HDC says he is unsure if it will be worthwhile. Some around the table agree. A few mention they were not informed properly. Others bring their issues and hopes to the table: more efficiency, team building, better translation of policy into implementation, and coherence of activities. Lively discussions in small groups do ensue: people start to feel heard, some real concerns are shared, first insights emerge, and the mood improves. When a plenary inventory is made the next morning, however, each group presents a cleaned up version. The sharp edges are gone, and safe container terminology is used such as “better communication.” Now, dissatisfiers such as “clear priorities” and “orderly archives” top the list. More complex issues such as “how do we interact with our local counterparts” have sunk to the bottom. Those at the bottom are at the heart of the embassy’s work, and innovation there would be worth the effort. When we point this out, some people shift their priorities, but most don’t, unsure such efforts will pay off. This is not remedied when we explain more about L&D processes in a PowerPoint presentation. We notice people’s eyes glazing over, and a few mutter it sounds like jargon. They are right: The principles behind L&D correspond with strategies least used at the Ministry. When we say they can choose issues they want to work on, they tell us that is not how things are done at the Ministry and ask what the hidden agenda is from headquarters. When we explain that they will manage this change themselves with our support, they ask what it will entail in terms of activities, time, and outputs and if we can guarantee that. We cannot supply them with all the answers, as the approach needs to be shaped incrementally with the embassy based on the issues they select. Benchmarks do not exist as the L&D projects differ greatly between embassies. Clearly, this does not put the participants’ minds at ease. (To be continued . . .)

To hang on to the demand-driven approach, some facilitators tried to dig a bit deeper to find out “what the real issues were.” But this was often seen as stirring things up. Even in this case, we see participants wondering if they will end up with more problems than they bargained for. This leads some to deny they have problems in the first place. After first requesting support, they back away saying, “It’s not that bad or urgent” or suggest others should take on the issue. “Help is needed but elsewhere.” The case also illustrates how this dynamic may evoke the other end of the initiation paradox: a supply-driven approach. At each embassy, there were notable critics who would grill us about our approach. This could partly be explained by projection: The critics seemed to recognize their own doubts in us and our approaches. The problem was that only tangible, predictable, and familiar approaches are persuasive, but those qualifications make them unsuitable for exploring tough issues. As a result, no reply could take their doubts away. When we made our designs more cutting edge, participants became more skeptical. Yet, when we came up with an approach more in line with participants’ expectations, they were disappointed we had so little to offer. Asking for more clues to fine-tune the approach fared only marginally better and was often interpreted as limitations at our end: “We are paying you to be experts, so surprise us.”

This can unnerve facilitators, and they may introject the insecurities and regard themselves as having not enough to offer. A last resort to reduce uncertainties is to collectively buy into a “magical solution”: to idealize an approach or person that discharges others of responsibilities (Hirschhorn, 1988). This can be in the “fatherly” guise of expert gusto or in the “motherly” guise of reassuring care. Expertise can be demonstrated by criticizing the participants’ organization, claiming more would be possible “if only . . .” One participant illustrated familiarity with this tactic: “The blood is still dripping off the Ministry’s walls from diagnoses done by previous consultants.” Expertise can also be implied by technocratic posturing in the form of multimedia presentations, power suits, client references, benchmarks, and the like. Such shared illusions, however, obscure reality. The expectation that magic will do the work for you cannot last; active involvement is needed when issues are interwoven with daily work. Facilitators who hype their work and camouflage what they have to offer soon lose participants’ trust. Learning dips, inevitable and even desirable during innovation, as people discover what they do not know, breaking the magic spell (Geschka, 1978). When the spell is broken, participants may pull out, or worse lose faith in dealing with tough issues and new approaches. Once the magic is gone, and their credibility with it, facilitators are rarely able to restore trust.

Handling the Initiation Paradox

Instead of basing contracts on reassurance, facilitators can go in the opposite direction. They can discuss with participants what issues matter enough to make the discomfort of unclear help and an unfamiliar approach worth enduring. The help requested/offered here is not the leading factor here but rather which issues are most suitable for innovation. Three criteria helped determine this at the start of each L&D project. (a) Are the issues “hot” enough for the participants? People have grown used to the persistence of tough issues and see them as both everybody and nobody’s problem. Working on them thus becomes a voluntary matter. This implies that lasting involvement depends on participants caring enough about these issues and getting energy out of addressing them. (b) Are the issues relevant and complex enough? Relevance increases when issues are part of the primary process and outside stakeholders are affected. Complexity implies a need to invest in learning. Together they legitimize the extra effort. In contrast, innovating because “it’s nice” never lasts. (c) Do the participants declare the issues within their domain? Since tough issues do not fit into delineated tasks, system limits are open to debate. There is always a good argument to be made that others should act first. The flipside is that participants are as free to define issues within their domain. As long as they take responsibility to do what is necessary despite a hostile environment, this is perfectly fine:

Back to the start-up visit to Sana’a where people had cold feet, even though energy increased when they discussed concrete challenges they encountered in their work. At the end of the second (and last formal) day of our visit a go or no-go decision was to be taken. As facilitators, we put discussion items on the agenda, reflecting the three criteria outlined above. The participants’ answers did not differ much from the facilitators. (a) Enthusiasm has increased during the 2 days but doubts are abundant as well. Half the group (mostly policy staff and a few control staff) seem eager enough. (b) There are issues that matter, but the most popular ones are internally focused and not very complex. (c) Conditions are not great. The embassy just found out that a large competing change would take up much time in the coming half year. Also, half of the embassy staff (including the ambassador) is to be replaced in a year’s time. Based on this, both the facilitators conclude that we should not move forward “unless . . .” The “unless” refers to possible shifts over time with regard to the three criteria. The group seems taken aback. “Now that we have identified issues, we can hardly leave things hanging,” says the HDC. A controller responds, “Outsiders cannot teach us how to put our house in order.” A colleague wonders if a failed start-up will not reflect badly on them in The Hague. Fifteen minutes later, the HDC closes the session. We stick around the day after to informally assist those who request it. Surprisingly, there is high demand to be coached on sticky external and internal issues (such as the restructuring of the Yemeni government together with the World Bank and troubled communication within the management team). Energy is high throughout that day, and some show disappointment about the “no-go” decision. A few weeks later, the HDC calls: The embassy staff no longer shares “our” conclusion. “Both discontinuity and competing changes are normal in a foreign affairs arena, not reasons to pull the plug. And if everybody does not want to put in as much effort, we will work parallel in smaller groups at different speeds.” He adds that the third informal day lasted a little more. He promises that the embassy intends to do serious prepwork and expects a second chance. Within 2 months, they analyzed their main issues and made an overview of possible remedies. This time, relevant issues figure more prominently. Together, we revisit the decision and start the L&D project. The HDC asks us afterward whether saying no was a trick to make them eager. He never quite loses that suspicion, even when he finds out that no-go decisions elsewhere often remained unchanged. In any case, the L&D in Yemen keeps running for 3 years.

The case illustrates how enthusiasm does not stem from talking about change but from taking something on such as figuring out causes behind and strategies for concrete issues. This corresponds with the notion of “swift trust.” Meyerson, Weick, and Kramer (1996) conclude that people choose this tactic when there is passion for the issue (it is worth the risk), when the road ahead is unpredictable (better to test than to talk), when contributions from others are indispensable (you need a “deal”), when group composition is temporary (you are done with them soon enough), and when there is time pressure (it is now or not at all). Tough issues have all or most of these traits. Surprisingly, swift trust happens when people put more at stake, not less.

In contrast, when issues are manageable, responsibilities clear, and relationships stable (such as in bureaucratic contexts), trust builds slowly as a precondition for collaboration (Worline & Quinn, 2003). “Swift trust” is thin, more a working hypothesis than a solid commitment. People contribute without needing prior reassurance and deal with obstacles along the way until they find it no longer delivers results. Commitments are provisional and doubts about the collaboration and the change are acknowledged and reflected on. It is this openness that allows mixed messages to be dealt with. This testing behavior also applies to the proposals of facilitators. Guarantees beforehand, though requested, are rarely trusted because facilitators are contested sources of information. A more suitable sales pitch focuses not on “what to do”; it problematizes first “what not to do any more.” A convincing argument can be made that “more of the same” of prevailing practices will not tackle tough issues. So there is not much to lose in trying something new. I label this “antiselling,” whereby you deconstruct fake certainties instead of offering new assurances.

The best way for participants to test facilitators is to work together “here and now.” As illustrated in the case, we worked as much as possible on concrete issues in the first contacts; we diagnosed together, debated contrasting visions, and introduced contacts and know how. There is no alternative for direct experience. In time, we did away with the idea of making big “L&D plans.” This did not mean that we would depart leaving the group mystified about what was to follow. Instead, we made provisional deals about the next steps. This fits with Wierdsma’s (2004) notion of “temporary workable agreements.” He underlines that a group cannot do what everybody wants, but neither is consensus needed. Consent will suffice: a deal that is just good enough to enable actions and ideas to be tried out until better ones come along. This way those involved incrementally shape the change one agreement at the time.

Execution Paradox: Letting Go Versus Taking Over

Once there is enough agreement, people start working on tough issues. At first glance, it seems desirable that participants are in the lead in designing, experimenting, and reflecting: It allows them to learn the most. The problem is that the more innovative an approach is, the less favorable the conditions will be and the more skillful participants need to be to pull it off on their own. A given change strategy only comes truly to life by executing it in depth: knowledge and experience count but also congruent ideals and values, diagnostic models and interventions, and roles and identities matter (Dunphy, 1996). Letting people take on a change strategy without such a background can easily lead to mediocrity. This allows them to prove to themselves that such alien approaches disappoint, underscoring the paradox of embedded agency.

Lacking ability creates problems, but also a lack of effort by participants may slow down the innovation. In politically charged environments such as the Ministry, it is easy to be pulled into the “issue of the day.” Voluntary learning and exploration may have come to be regarded a luxury item you have time for when all your routine affairs are in order (which is never). At the embassies, it led people to show up late or not at all, to push meetings back, and put little effort in. They often did not consider this problematic, as skipping internal meetings was neither unusual nor interfered strongly with the Ministry’s routines such as policy formation. With regard to such routines, they knew they could always pick things up where they had left them. They often had not discerned that it is a different matter when you collaboratively experiment with new repertoire around tough issues. It made a lack of effort less a sign of ill will but rather more of an inability to self-manage their innovative project. It would nevertheless slow things down:

At the start-up visits for any L&D project, we struck an agreement that the facilitators would periodically visit the embassies to catalyze the process: to take stock of what had and hadn’t worked, to work closely on chosen tough issues, to introduce new knowledge, to replan the change process for the next half year, to coach key players, and to resolve conflicts. Timely visits were key so that learning dips could be dealt with in time. Still, dates for visits would be hard to agree on and trips were regularly moved or cancelled. In the mean time, local L&D efforts would slow down because of participants’ blind spots or newly emerged “issues of the day.” The blind spots made it difficult for local champions to clarify what kind of help they needed. The long distance communication between facilitators and embassy would become more haphazard. Both participants and facilitators felt like they had lost their grip on the L&D process. In some cases, facilitators tried to pressure the embassy staff to be more disciplined, but this was rarely effective. Their suggestions were not always comprehended, and there were always competing demands on the staff’s time. Facilitators were faced with the choice to pull the plug or to make the best of an unprepared (late) visit, neither of which they liked. In most cases, facilitators would show up to discover that they had misinterpreted participants’ tameness. Counter to expectations, participants had ample issues to work on and made use of all the improvised help they could get. Many put in long hours, and sessions ran into the night. However, much of the learning focused on immediate concerns rather than on managing their own change. This made the “three step forward, two step back” experience a hard one to overcome.

There were moments when facilitators felt they could not go along with the situation any further. In the Pretorian Embassy, a small team had enjoyed a yearlong L&D project and wanted to triple its size to include all the other embassy staff. However, they seemed not to consider the necessary “homework,” such as dealing with the skepticism of new participants to bring them on board. Neither was the time ideal; the embassy was being restructured, causing considerable infighting. As facilitators, we felt they were setting themselves up to fail, and we did not go along: “Saying the program needs to be bigger does not equal enlisting colleagues. If you can’t see yourself doing that work, we can’t see ourselves facilitating it.” It made one participant accuse us of being “sudden sticklers for invisible rules.” Another remarked that we had finally showed our “true colors”: We were not nice and supportive after all.

The case illustrates different facilitators’ responses to an L&D project slowing down. The first was to demand more of participants in terms of time, effort, and discipline. Such demands could cause irritation “It’s not like it is clear what you expect(ed) us to do.” This would be a valid point because in an explorative process it is both difficult to explain and to forecast what will be required. A second response was to step in and take on many roles needed to save the change program. Not only is this a tiring prospect but also a dysfunctional one. Expert solutions can be outsourced but not innovation around tough issues. Many roles cannot be fulfilled other than locally. Also, the more facilitators took over, the less participants learned to run it themselves and the more innovation slowed down. This is a vicious cycle leading to overworked facilitators and consumerist participants. When it doesn’t work, they only look to the facilitators to fix it. Even when it works, they wouldn’t know why (it’s magic). In both cases, the required activities, ingredients, roles, and ideas are obscured. The more these are obscured, the harder it is for participants to follow their facilitators, especially when they switch a lot between many different roles. They become uncertain whether to expect support or direction, inquiry or advice, detachment or collaborative thuggery (Huxham & Vangen, 2005), or even to distinguish them from one another. The facilitator’s third response included one of those less expected roles of being very directive, and that took participants by surprise. Some participants have used ownership ideology as a defense mechanism against such pushy behavior of facilitators: “If we are the owners, than we make the decisions.” Lacking understanding, it leads some to blindly trust their facilitators, and others to blindly distrust them. Neither helps coproduce the change, and both can erode facilitators’ credibility.

Handling the Execution Paradox

The challenge during execution was to have participants gradually take on more leadership roles. Tough issues need many contributions to tackle them: know-how from different disciplines, political entrepreneurship, an ability to unravel issues and plan changes, and a capacity to reflect. Instead of monopolizing leadership in one manager or facilitator, we tried to distribute and share leadership. Complementary contributions tend to enhance each other. However, similar contributions tend to compete. This requires stretching participants’ contributions to be less one-sided in three ways: more communication in terms of process (rather than content), more facilitative in behavior (rather than directive behavior), and more self-organizing of changes (rather than reacting to them). We found that these three less habitual contributions make coproduction robust, but awareness of them and experience with them has to build over time. To this end, facilitators would prioritize support to participants who took on such facilitative roles. This can be regarded as creating temporary support systems to allow for deeper learning in selected subgroups. It is in line with the idea that multiplicity of learning processes is more effective, especially around tough issues, as it allows learning in different compositions, speeds, and depth (Dodgson, 1993).

The delineation between what external facilitators would do and what participants would do shifted over time. An effective rule of thumb for facilitators appeared to be as follows: (a) whatever participants can take on themselves, let go of as facilitators and (b) what participants forget or botch up, facilitators should take on selectively and partially to the extent that the change does not stall. This implies that facilitators take more responsibility for those contributions that are least fulfilled. The qualifications “selectively and partially” are meant to counter the pitfall of leaving little room for participants to step into the facilitators’ shoes:

In each L&D project, we distributed facilitation roles. Different participants would try their hand at facilitating separate learning activities. Each embassy also had a core team co-organizing their L&D project as a whole. Last, in each project, someone from the Ministry headquarters was coupled with an experienced consultant to serve as lead facilitators. This way, facilitation roles could be learned at three different levels. We prioritized our support to these cofacilitators: We coprepared, co-evaluated, and sometimes cofacilitated activities. Delineating what to take on and what to let go was relatively easy “on site”: We could adjust it flexibly. It was harder when we were designing together for the following months when the embassy would run their L&D project in our absence. Especially in the early stages of each L&D project, facilitators felt they could not leave the change design over to a fresh core team. In order for the analysis, design, and interventions to be robust enough for the next half year, they often set up a few structured meetings and led the core team through a “pressure cooker” design process. It needed to be fast as the embassy’s time was always limited for such an unfamiliar activity. The upside of this “pressure cooker” approach was that it would shock each core team. The experience reduced their blind spots and made them aware of what such a change process takes. One core team commented, “We only worked on it during three short meetings, but we are completely beat. This is hard work, and it is all new.” The downside was that the faster facilitators moved, the more apprehensive and passive the team could become. More than once the lead facilitators would haggle over this balance, with one feeling the learning design needed to be stronger and the other feeling that if we pushed any more “it would turn into our party instead of theirs.”

Pushy facilitators’ behavior could lead to heated discussions with participants, the earlier example of tripling the L&D project in Pretoria being a case in point. Half the participants were adamantly in favor and half were undecided, and the facilitators were against it. For a while, it looked like it might turn into a fight about what people thought should happen instead of what could work and about who is in charge instead of what contributions each would vouch for. We forced the discussion to shift from the first (ideals and mandates) to the latter (possibilities and abilities). Each person had a turn sharing their reasoning, and collectively, we weighed our options and decided on a gradual enlargement of the program.

Whenever possible, facilitators deliberately fell back and dared participants to pick up the slack. Having participants take on new roles in an innovative change process creates frictions, but as long as these are manageable or reversible, such frictions accelerate learning. Vermunt and Verloop (1999) label them as constructive frictions in contrast to negative frictions that arise either when mishaps put the change process in jeopardy or when third parties interfere needlessly where participants were capable enough. This implies quite a bit of balancing between steering and supporting, optimizing the change and constructive friction, and local implementation and contextual politics.

How facilitators managed this balance is a matter of judgment on which they could disagree as shown in the case. They did generally agree that to enhance the abilities of participants to self-manage their change in the future, it is worth having a less perfect change program in the present. What an expert might regard as a perfectly designed change program allows for little empowerment for participants to learn to create it themselves. What facilitators also agreed on is to share their professional insights whenever possible. Chameleon-type behavior of facilitators who shift quickly between roles could frustrate participants’ expectations. Though facilitators generally felt they needed to be true to their best professional judgments regardless of such expectations, they could share the reasoning behind their choices. This enabled professional insights to be shared in connection to the participants’ experiences. The more such insights were shared, the less intense intervention paradoxes became.

Trust in a change or a facilitator is then based on understanding rather than on predictability. Such insights could however only be shared piecemeal as participants gained experience. When too much is “taught” at one time, it loses relevance, and participants lose interest quickly. By sharing their reasoning, facilitators also modeled “reflective practitioning” (Schön, 1983) as a flexible way of steering change in contrast to the more familiar monitoring of plans and policies. This also implied that any participant who came up with another, more convincing line of reasoning, could shift the course taken, which would then be tested in practice as illustrated in the case of Pretoria. The more coproduction is advanced this way, the deeper the change can become.