By Authors:
Rodrigo Canales – Boston University
Mikaela Bradbury – Yale School of Management
Anthony Sheldon – Yale School of Management
Charlie Cannon – Rhode Island School of Design
Interviewers:
Linette Dawson – University of Tampa
Yanbo Song – INSEAD
Article link: https://doi.org/10.1177/00018392241241483
How did the research collaboration arise? What did the start of the project look like?
Rodrigo: The collaboration emerged quite organically through a series of workshops I was facilitating with the Hewlett Foundation. We were discussing the challenges around intersectoral collaboration and integrating evidence into practice. At some point, one of the participants from the foundation said something that stuck with me. They said, “We’ve been talking about intersectoral collaboration and learning across sectors for decades, and yet we always seem to fail.” That sentiment really resonated. Despite having capable actors, motivated teams, and good financing, these collaborations often fell apart. So, the foundation said, “We want to do a study to understand why we keep failing.” That became the starting point. We knew we wanted to focus on the paradox: why smart, motivated people working on urgent problems, with good resources, still often could not collaborate successfully across sectors. From there, we identified a set of projects — some that had failed and some that had succeeded — to try to understand what made the difference.
Charlie: My path to this collaboration came from a different direction. I am trained as an architect but have worked in industrial design and organizational consulting. My interests have always revolved around how design principles and design thinking can be applied to large, complex problems — what I often call the “small d, small t” approach to design thinking. When Rodrigo and I started talking, we quickly realized that his work on institutional change and my background in design were both addressing similar challenges from different perspectives. That made the partnership very exciting.
Rodrigo: And to Charlie’s point, this project really bridged both our domains. The institutional side needed the design sensibility to handle ambiguity and emergent solutions, while the design side benefited from the structural thinking we brought in through organizational theory. We were looking at the same problem but using different lenses. That synergy helped us formulate what eventually became the concept of scaffolding practices.
Charlie: And we kept coming back to the same question: What enables collaboration to work when so many forces seem to pull it apart? That is what kept the research going.
You highlight modular structure and programmatic scaffolds as key to making collaboration easier. Could you share real-world examples from your cases of how cross-functional and virtual teams tackled challenges across different cultures and geographies? What made these approaches effective in your understanding?
Rodrigo: What we found is that most collaboration breakdowns stem from a lack of basic organizing agreements. When people collaborate within a single organization, there are typically invisible structures already in place — norms, roles, accountability systems — that make collaboration possible. But when collaborating across organizations, sectors, or geographies, these assumptions fall apart. Simple but essential questions — Who is doing what? How do we stay committed? How do we hold each other accountable? How do we know what we’ve learned? — go unanswered. Without mechanisms to address those questions, teams falter. But in the successful cases, we saw that teams used scaffolding practices — deliberate frameworks to build shared agreements — to move forward.
One example is what we call the negotiation of currencies of exchange. Collaborations often break down when participants can no longer justify the investment of their time or resources. For instance, someone might say, “I love this project, but I have a quarterly performance target, and this work won’t help me meet it.” Or a politician might drop out, saying, “I’m up for re-election soon, and I have nothing tangible to show for my involvement here.” In successful cases, these conversations happened early. Stakeholders sat down and articulated what they needed to get out of the project and by when. An academic might say, “I need to publish within the next year to stay on the tenure track, so I’ll need access to data by then.” A teacher might say, “This is great, but I’m required to teach a set curriculum. I need permission to teach outside the plan if I’m going to participate.” And a policymaker might say, “I need a concrete story or pilot result I can present to voters in 12 months.” By aligning on what each person needed—and the timeline in which they needed it — the collaboration became more sustainable. Participants could look ahead and say, “If Linette doesn’t get data by month six, she’ll have to drop out.” That knowledge helped everyone stay committed and accountable. Managing these timelines was especially important, as each sector works on different rhythms: academic publishing cycles, electoral terms, school calendars, and so on. This scaffolding practice — negotiating currencies of exchange — helped stakeholders remain connected by making invisible constraints explicit and negotiable.
Charlie: If I may zoom out, what we have come to see is that process and organizational challenges are deeply interwoven. Many of the scaffolding practices we identified are ways to address both types of challenges simultaneously.
Take the example of a protective convening space. On one level, it is a process-oriented tool — creating a space for brainstorming and idea generation. But on another level, it plays a critical organizational role: shielding participants from the pressures of their day jobs. A teacher might say, “I’ve been invited by the National Department of Education to join this initiative,” and that invitation becomes their alibi or institutional coverage to step away from standard duties. The protective space gives participants the psychological and political cover to experiment with unfamiliar ideas, suspend hierarchies, and begin working in new ways — something especially crucial when navigating uncertainty or challenging the status quo.
From my experience working with design teams in innovation networks, this challenge is familiar. Designers tend to understand process ambiguity well but often lack the tools to navigate institutional complexity. Conversely, project managers may be skilled in organizational alignment but uncomfortable with the fluidity and uncertainty inherent in creative collaboration. Scaffolding practices bridge those gaps. They help create a structure that enables innovation while also providing the stability needed to sustain commitment. And yet, as Rodrigo noted, despite widespread failure being the norm, people are still surprised when these collaborations do not work. That is why articulating and sharing these scaffolding mechanisms is so important.
Your paper highlights both the stabilizing and extending functions of scaffolding. Prior research often focuses on extension — enabling experimentation and innovation — whereas you show scaffolding also plays a stabilizing role. Were there instances where one function was more critical than the other? How do organizations navigate the need for both flexibility and stability in uncertain environment?
Rodrigo: It is difficult to isolate a case where either the stabilizing or extending function was more important — what we observed is that both were needed simultaneously. One of our key insights is that the process and organizational challenges in cross-sector collaboration are deeply intertwined. These challenges arise together, and effective scaffolding practices help address both dimensions at once. We have noticed a tendency for people to lean toward one side or the other, depending on their background. For example, designers typically have a strong grasp of process challenges — they are comfortable with uncertainty, iteration, and ambiguity. But they often overlook the organizational side of things: how to align incentives, build structures, or navigate inter-organizational politics. On the other hand, project managers or bureaucrats might be adept at structuring relationships and managing stakeholder alignment, but they struggle to deal with process ambiguity or adapt when the path forward is unclear.
Let me give a concrete example. A common way to coordinate complex collaborations is through a hub-and-spoke model, where a central organization manages bilateral relationships with each partner via formal contracts. That model works when you can predict what is going to happen — when you know what each party will contribute and what outcomes are expected. But in many of the innovation-driven collaborations we studied, this level of certainty just did not exist. In one case — a youth employment initiative called PPE — we wrote contracts based on assumptions about what would happen. But once implementation began, we discovered that some of those assumptions were simply wrong. These were not things anyone could have predicted; they only became clear through doing the work. As a result, the contracts had to be renegotiated.
And here is the problem: once one contract changes, it triggers a domino effect. Every other partner who signed an agreement based on the original assumptions now has to revisit their terms. That kind of cascading renegotiation can derail a project completely. No collaboration can survive this kind of “whiplash effect” for long. A designer might anticipate this —“Of course, we didn’t know exactly where we were going!” — but the people organizing the collaboration are often working from a very different logic. They may feel blindsided by the uncertainty and unprepared to respond flexibly. Conversely, a process-oriented person might excel at embracing the unknown but fail to provide the accountability and structure others need to stay engaged.
That is where scaffolding practices come in. They provide a substitute for formal contracts — a stabilizing mechanism that gives participants enough structure to commit while still leaving room for adaptation. One such scaffold is the negotiation of currencies of exchange, which we discussed earlier. Another is the protective convening space, which creates an environment where hierarchical norms can be temporarily suspended, allowing stakeholders to brainstorm, challenge assumptions, and co-create freely. These stabilizing functions matter because they allow participants to explain to their home institutions why they are involved, why it matters, and why it is okay to take the time and resources to be part of something inherently uncertain. Without that, people revert to their default modes — project managers want tight contracts, and creatives want open-ended exploration — and the collaboration collapses under the tension.
Charlie: To build on Rodrigo’s point, I would say that the stabilizing and extending functions of scaffolding are often two sides of the same coin. For instance, a protective convening space is not just about having a room to brainstorm ideas — it also shields participants from the demands of their home organizations. It offers both psychological safety and institutional legitimacy.
In our own work, we have seen this play out repeatedly. Design teams, for example, tend to be fluent in navigating process uncertainty, but they are rarely trained to anticipate the organizational constraints that surround collaboration. Likewise, many development professionals are extremely skilled in navigating institutional politics but are not always comfortable working without clear plans or deliverables. So, these scaffolding mechanisms are critical for creating a shared structure that accommodates both ways of thinking. You need to de-risk the environment before you can ask people to take creative or strategic risks. That is why we emphasize that scaffolding is not just about enabling flexibility — it is about creating the stability that makes that flexibility possible.
In sum, we found that successful collaborations never relied solely on either the extending or stabilizing function. Instead, they employed practices that addressed both — often in ways that were deeply integrated. Prior work has rightly emphasized scaffolding’s role in enabling innovation, but we believe its stabilizing function has been underappreciated, and it is a critical part of making these collaborations work — especially when navigating uncertainty across sectors, disciplines, and geographies.
Your research offers valuable insights into enabling collaboration across diverse cultural and institutional contexts. Could you walk us through how the study was designed and how you selected and analyzed your cases to ensure the robustness and relevance of your findings?
Rodrigo: Our research began by interviewing seasoned international development experts — people with 30 years of experience in various stakeholder roles. These were open-ended interviews, and we simply asked, “Why do these cross-sector collaborations so often fail?” Their responses revealed common patterns in the breakdown of collaboration. From that, we identified key challenges that repeatedly emerged and then developed a theoretical sampling frame to explore these themes more deeply. We then sought out potential case studies that reflected these patterns — large, cross-sector collaborations that involved strong actors, sufficient funding, and ambitious goals. From a pool of 30 to 40 possible cases, we selected eight that were matched in terms of structure and objectives but varied in outcome: some succeeded, some failed. We deliberately ensured geographic diversity, spanning four different countries, to avoid having our findings reflect only one national or regional context.
For each case, we conducted extensive fieldwork — interviews, observations, and archival research. We wrote detailed narratives and constructed comprehensive timelines highlighting critical decision points, challenges, stakeholder interactions, and strategic pivots. Each of these eight timelines allowed us to identify what was similar across cases and, importantly, what distinguished the successful collaborations from the failed ones. What we discovered was that all collaborations encountered similar challenges. However, the successful ones handled these challenges differently — through what we came to call scaffolding practices. These included programmatic tools like protective convening spaces and organizing tools like negotiation of currencies of exchange. Importantly, all challenges had both process and organizational dimensions, and most failed cases missed addressing one or the other. Once we developed this preliminary framework, we hosted a three-day workshop with about 100 international development professionals not involved in the initial research. We shared our emerging insights, received feedback, and refined our understanding through collaborative discussion. We also conducted follow-up interviews to test the resonance of our findings with experienced practitioners. These steps helped us confirm that our findings were not only theoretically grounded but also robust across contexts.
How did you approach concerns around generalizability, and what made you confident that the scaffolding mechanisms identified could be applied across different geographies, sectors, and cultural settings?
Charlie: We deliberately avoided making our findings overly bound to any single cultural or geographic context. Instead, we focused on principles of process design and organizational theory that emerged consistently across cases. Each scaffolding practice we identified — such as negotiating shared goals, redistributing authority, or creating a learning orientation — emerged from real situations but reflects broader principles. These are not generic “best practices” but rather context-sensitive strategies grounded in how people actually work through complex challenges together.
Let us take the example of negotiating the problem frame. In every setting — whether among government officials in India, NGO leaders in Mexico, or donor coalitions in Ghana — people came into the room with different definitions of the problem and different experiences of what had or had not worked. These definitions were shaped by social norms, professional roles, cultural expectations, and institutional histories. The key was not to erase these differences but to create a process that enabled them to be surfaced and negotiated.
We do not claim that our model will play out identically in every setting. However, we do believe the scaffolding practices offer a useful structure for groups to navigate their own challenges without getting stuck in context-specific pitfalls. In essence, we are offering a framework that enables teams to work with, rather than against, contextual complexity.
Rodrigo: Yes, and one last point on that — these practices are rooted in foundational theories. Each one is derived from principles in organizational design or in managing under uncertainty. We were not trying to describe what happened in Mexico or in Ghana per se; we were asking, “What principle of process design or organizing theory does this illustrate?” That is why, even though the examples are specific, the practices themselves are applicable in other settings.
Of course, cultural variation matters. For instance, in a high-hierarchy culture like Japan, redistributing authority may look different than in more egalitarian systems. But we think the underlying need—to create space for distributed expertise — remains. In fact, in highly hierarchical systems like healthcare, where doctors often dominate, scaffolding mechanisms such as protective convening spaces are even more essential. They provide psychological and institutional cover to suspend hierarchy and allow shared problem-solving to occur.
Charlie: And that connects to another important point. One thing that enables people to suspend their usual roles and hierarchies is “failure.” Many of our cases involved efforts where previous attempts to address a problem had not worked. That shared experience of failure created the conditions for humility and openness. It pushed people to acknowledge that their standard tools were insufficient, and it made them more receptive to new ways of working together. But even then, this shift does not happen automatically. That is why one of our key scaffolding practices is establishing a learning orientation. We saw that in the successful cases, there was a moment when participants explicitly said, “We don’t know what we’re doing, and we need to figure it out together.” Without that moment, people defaulted to familiar hierarchies or disciplinary silos — what we call “muscle memory.”
So again, we are not offering a one-size-fits-all model. We are offering a set of principles and practices that can help diverse teams recognize and address the common pitfalls of cross-sector collaboration. How they apply those tools will depend on their context — but the need for those tools appears to be remarkably consistent.
With 226 in-depth interviews, what specific analytic strategies did you use to move from raw qualitative data to identifying scaffolding mechanisms? Were there patterns in the data that surprised you or challenged your initial assumptions?
Rodrigo: Initially, the article was written in a more traditional qualitative format. We drew on all eight cases and presented contrasting examples across them. However, the challenge was the sheer depth and complexity of each case. Our article became too lengthy to fit within the ASQ’s page limit. Mike Pratt, the editor, advised us to condense the narrative. His suggestion was to place the rich detail from all eight cases into comprehensive tables, which would demonstrate the rigor and breadth of the study while allowing the main text to focus in-depth on just two contrasting cases. While difficult, this decision allowed us to tell a coherent story within the constraints of the article format. As academic writers, it was painful to cut out so many vivid quotes and detailed insights — “our babies,” so to speak — but we believed readers could still trace the logic of our findings through the detailed methods and tables. The chosen two cases served to clearly illustrate the process model, while the broader data lent its robustness. Of course, this is why we are now writing a book — to do justice to the full richness of the cases and bring back the insights we had to leave out. We also have a full report available for those interested in more detail.
To manage and analyze 226 in-depth interviews across eight cases, we assembled a large team. Each country or case was assigned to two research analysts who conducted interviews, transcribed them, and conducted the initial coding. To ensure methodological rigor and consistency across the team, we held weekly meetings where we reviewed our labeling, coding strategy, and interpretations. From the transcripts, we engaged in traditional qualitative coding, identifying patterns like: “this is a challenge,” “this is a workaround,” “this is a breakdown of incentives,” or “this is a timeline misalignment.” The research analysts then constructed detailed timelines for each case, identifying key decision points, actions, turning points, and outcomes. These were based on stakeholder narratives, allowing us to reconstruct the trajectory of each initiative.
Each of these case timelines and narratives was then read and reviewed by all the researchers, using common language and labels. This collaborative cross-checking was crucial in ensuring that we applied a consistent lens across diverse geographies and contexts. Once we had the eight detailed case narratives and timelines laid out, we conducted a comparative analysis to identify what was common across all projects, and what differed between the successful and failed cases. The scaffolding mechanisms emerged from these comparisons. We then went back to the raw interview data to unpack the “how.” That is, once we identified a particular mechanism that enabled success, we returned to the transcripts to understand in detail what stakeholders said and did in those moments — how they navigated the challenge, what they negotiated, and how they reached agreement. This iterative process of theory-building and validation was key to distilling the scaffolding mechanisms.
Charlie: In the later stages of the analysis, we also developed consistent visual representations to aid comparison across cases. These included: 1) Stakeholder maps, which we designed to reflect the variety and roles of participants in each initiative. 2) Process diagrams showing how projects moved from problem framing to prototyping and scaling and where breakdowns occurred. 3) Comparative timelines to identify truncation, delay, or acceleration across cases. These visual tools helped us validate and refine our findings and have been included in the report. They also enabled others to see, at a glance, what patterns held across cases and where key differences emerged.
Were there any unexpected findings that warrant further investigation? What do you see as the next extension of this work?
Rodrigo: One recurring insight that continues to surprise us is the consistent pairing of process challenges and organizational challenges. They often appear together and are interdependent. Teams that only addressed one side — say, the procedural uncertainty of collaboration — often missed the organizational misalignments that ultimately caused breakdowns. This pairing of challenges led us to see scaffolding practices as dual-purpose: they extend what teams can do, but also stabilize collaboration under uncertainty. We are also continuing to refine this model in our book, including developing practical checklists for scaffolding practices like “negotiating currencies of exchange” or “creating protective convening spaces.”
While most scaffolding practices we identified are grounded in well-documented theory, two practices stood out as needing further exploration: “adaptive financing” and “mapping learning pitches.” Adaptive financing is critical, but still under-theorized. It requires funders and collaborators to build financial models that allow for course correction and adjustment — something that is currently rare. Groups like the Systemic Investing Consortium are starting to tackle this, and we believe research can contribute here. Mapping learning pitches is another area we are excited about. Borrowing from rock climbing, we conceptualize “learning pitches” as bounded zones of uncertainty. Collaborators need to anchor themselves, assess where they are, and determine the next step up toward their North Star or scale goal. This metaphor can help project teams navigate nonlinear progress. We believe this concept can be applied across sectors—from architecture and engineering to public health and climate resilience. But how people do this well, and how to teach it, remains an open question.
Charlie: Finally, a broader question we are exploring is: “Can you teach scaffolding?” In our study, we found people who had learned to scaffold through trial, error, and years of failure. But can this knowledge be taught systematically to more people? If yes, we can stop relying on exceptional individuals to make complex collaboration work, and begin to build the systems and tools that make it possible more broadly. That is a big question for both researchers and practitioners.
Do you have any reflection questions for readers — things you would want them to think about after reading your article or book?
Rodrigo: I think my answer would actually respond to one of the questions we mentioned earlier — which was about the practical applications or how this might be applied in different places, in different ways. And so, what I would ask readers is for help. Just as we want to make sure that these principles and scaffolding practices are applicable and useful in a variety of international development projects — because that is how we came to our understanding of this — I think what our colleagues and we have also felt is that the findings and insights we have developed from that field of practice, which is a broad, cross-sectoral field, seem to be incredibly valuable and applicable for folks who are working at that same scale. Like people who are working in climate action, or in greenhouse gas reduction, or thinking about environmental justice or environmental degradation, or policing and nonviolence, the insights seem to be very useful in those fields. I think part of what we are working through as we move into the bookmaking is to be hunting for and looking at examples, both successful and unsuccessful, of projects or initiatives in these other fields and in other places. That might help us reflect on our findings and test them more but also help us see how or where — and help other people see how and where—these findings might be useful in these other fields of practice.
In the preliminary work we have done so far — whether that is because of our own professional work outside of academic research — we are finding that these things are applicable. It seems like it happens at three different tiers of organization. The first tier is as broad as international development, where we are talking about really broad cross-sector engagement. Then, as we move down a level, we get into places where… maybe, like some of the examples you were asking about — could this work for entertainment or fashion or something like that? Or other kinds of politics where there might not be all of these deep cross-sectoral players. They may be working within a field of practice, but they themselves are coming from many different disciplines or backgrounds, right? So, there is kind of a value for that interdisciplinary — even if we are maybe not at the transdisciplinary level — there are valuable insights and applications in interdisciplinary practice.
And then, for sure, we are also seeing it in talking to not just our research colleagues but colleagues from different schools who are engaged in research projects looking at organizations—multinationals, large-scale enterprises—where the traditional siloing that we see so often could really benefit from many of these same kinds of scaffolding practices within an organization. It is not necessarily the area we are most interested in for applying these insights, but they do seem to be relevant.
Charlie: What I would love to hear from readers is: Can you share examples with us? Are there areas of practice where you see how relevant this is? Or do you have a counter — like, ‘Well, actually that seemed… we’ve got some good failures,’ or ‘What you’re saying doesn’t reflect our experiences in these places because it’s missing this or that,’ right?” That — ”it’s missing this or that” — would be especially interesting for us as we are moving forward with the book and trying to distill what is most useful about all this.
Interviewer bios:
Linette Dawson is a doctoral candidate in the EDBA program at the University of Tampa. Her interests lie in personality and cultural intelligence in organizational behavior and international business. She can be contacted at linette.dawson@gmail.com or https://www.linkedin.com/in/linettedawson/.
Yanbo Song is a doctoral candidate in Organizational Behavior at INSEAD. She is interested in the creative process and collaboration dynamics, especially in cultural industries. She can be reached at yanbo.song@insead.edu or https://www.insead.edu/phd-student/yanbo-song.