From Cross Purpose to Common Ground: Making Collaboration Work.

Have you ever been three sentences deep into a conversation before realizing you and the other person were talking about entirely different things? Not disagreeing, just orbiting different meanings of the same words. It happens easily when everyone in the room is invested, moving fast, and excited about what they’re doing or building. In literature this is called talking at cross purpose. The good version is when someone catches it and the group rewinds to find where the meanings split. The bad version is when nobody catches it, and everyone leaves believing they agreed on something they didn’t. 

I think this happens constantly in our circles, and I’ve watched it happen often around the word “collaboration.”

Everyone  agrees collaboration matters. Funders write it into calls for proposals. Researchers cite it in their theories of change. Innovators build it into pitch decks. Governments name it in policy frameworks. But agreeing that collaboration matters is not the same as agreeing on what it is, and this is where the stakeholders keep talking at cross purpose. My argument here is that: collaboration is not a shared format where everyone participates in the same way. It’s an exchange,  each stakeholder entering through a different door, bringing something the others need, and getting something in return that they need too. And because those stakeholders are rarely just holding different needs but also different futures, on different timelines, the exchange has to hold time, space as well as interest.

Why Collaboration Matters 

In a previous piece, I wrote about Why Climate Innovations Fail To Scale and how the first valley of death happens not from a lack of funding or technical talent, but at the seams. It breaks down at the handoffs: between the scientists who generate knowledge, the engineers who build prototypes, the financiers who manage risk, and the communities who live out the consequences.

Every framework innovations lean on, exists because single-actor innovation was never enough to meet a problem the size of climate change. Eric von Hippel’s lead-user theory established a foundational truth decades ago which states that critical innovation insights often originate outside the formal R&D lab, from people living closest to the problem who build their own workarounds long before a market recognizes the need. That intuition branched into user-driven innovation, user-centered design, and customer co-design.  Different names circling the same reality, that the end-user is a collaborator, not a passive recipient.

That outside-in view expanded further when Henry Chesbrough formalized open innovation in 2003, arguing that useful knowledge is widely distributed and organizations should manage knowledge in flow (moving inward and outward) rather than sealing it inside an institutional vault. That reframe laid the groundwork for the university-industry partnerships, corporate venturing, and platform ecosystems we rely on today.

 From there, the literature scaled into the innovation helixes i.e,  triple, quadruple, quintuple. These frameworks expanded the ecosystem to formally incorporate government, academia, and civil society, eventually recognizing the natural environment itself not as a backdrop, but as a primary stakeholder with hard ecological limits and legitimate claims on the process.

The theory has spent decades expanding who belongs in the room. What hasn’t kept pace is the practice of listening to why each of those actors is there, and what they actually need from the table.

Why One-Size-Fits-All Collaboration Doesn’t Work

Most “collaboration” in the development space still looks like one format stretched across every stakeholder. A workshop, a data-sharing platform, a multi-stakeholder convening, a WhatsApp group. These aren’t bad tools, but at times  assume everyone in the room wants the same thing from being in the room, and that’s where inclusion breaks down without anyone noticing.

A researcher wants to do research. An innovator wants a market. A funder wants a scaling pathway. A farmer wants to see, with their own eyes, that something works before changing how they farm. Put all four in the same convening format and you’ll get four different levels of engagement, and probably one or two people who quietly disengage, not because they don’t care, but because the format did not feel inclusive.

This is a communication problem before it’s a collaboration problem. Everett Rogers offers a useful distinction here: communication as convergence, not transfer. A one-way output such as  a report, a policy brief, or a recommendation  isn’t communication in this sense; it’s broadcast. Real communication requires two parties moving toward shared meaning together, which means the receiving party has to be able to push back, question, and negotiate. Applied here  we have to ask, ‘can the farmer question the advisory?’, ‘ Can the community push back on the model’s assumptions?’ If not, what we’re calling “engagement” is often distributed with a feedback and survey attached.

Rogers also points to something less comfortable: ideas move most easily between people who are already similar in belief, education, and social status ,homophily. It explains why innovation clusters diffuse quickly within a sector and stall badly across sectors. The flip side is heterophily, which is the harder and more important case here. The more technically sophisticated the source, the harder communication becomes, almost regardless of how accurate the message is. A brilliant model built by a team of climate scientists doesn’t automatically translate into something a smallholder farmer can act on. It is  not because the farmer can’t understand it, but because the distance between the two worlds was never bridged by anything other than the model’s output itself. So we went from a communication breakdown to a disconnect where maybe all the stakeholders are present in the room, some as props but the process pushes on because attendance becomes more important than participation. 

What Holds It Together: Collaboration as Exchange

We can push back a little bit and ask, if every stakeholder needs something different, what makes this a collaboration rather than five separate relationships running in parallel? A researcher getting a publication pathway, a funder getting risk data, a farmer getting a demonstration plot. None of that, on its own, adds up to people working together. It could just as easily describe five vendors serving five customers.

The answer is closer to how a society holds together than how a transaction does. No functioning society is made of people who all want the same thing, believe the same things, or come from the same place.  It’s a mix of different ethnic groups, cultures, and values, held together not by sameness but by a kind of elasticity that gives in the structure to stretch around difference without snapping, and enough shared stake in the same outcome to keep everyone pulling in roughly the same direction. 

Collaboration needs the same elasticity. It does not require everyone to want the same thing. It requires people with different interests to see enough value in moving toward a shared outcome together. If collaboration is only ever framed as “how do we meet everyone’s needs,” it quietly turns into customization: five bespoke experiences bolted onto one project name. What actually makes it collaboration is that it’s an exchange, each stakeholder is both a recipient and a contributor, and the shared outcome only becomes reachable because of what changes hands between them. 

A stakeholder map for a project might look like:

Stakeholder	What collaboration means to them	What they bring
Researcher	A pathway for knowledge to be used	Evidence and knowledge
Innovator	Access to knowledge and resources	Translation into a usable solution
Funder	Shared risk and a credible scaling pathway	Capital and risk tolerance
Farmer	Trust, relevance, and proof	Lived experience and ground-truth validation
Government	Coordination and governability	Legitimacy, policy, and institutional pathways

Seen this way, the shared purpose isn’t a slogan everyone signs onto — it’s the outcome that only becomes possible once these contributions are actually in circulation with each other. Designing collaboration isn’t about making everyone participate the same way. It’s about building the conditions under which different forms of contribution can reach each other.

It’s worth being honest about what this table hides. It’s an example based on assumptions, and it draws each exchange as a clean, roughly equal trade , evidence for pathway, capital for risk tolerance. Real exchanges are rarely that symmetrical. A funder can walk away from a stalled pilot and redeploy capital elsewhere; a farmer who has restructured a season around a demonstration plot usually can’t walk away from a bad harvest with the same ease. The farmer’s contribution such as  lived experience, ground-truth validation , is often the cheapest thing on the table to ask for and the most consequential thing to get wrong. Naming that imbalance doesn’t break the exchange framing; it’s a reason to weight the design of the exchange deliberately toward whichever stakeholder has the least power to renegotiate it later, rather than assuming a tidy five-cell table means a tidy five-way balance of power.

The Same Project, Five Different Reasons to Care

This is easier to see with something concrete. Take an AI-based climate advisory tool ,  the kind of thing that gets built, piloted, and written up as a “successful collaboration” fairly often in this sector.

For the researcher, it’s a chance to see years of modelling work finally applied outside a paper and  validation that the science holds up against real conditions. For the innovator it can be a product and a route to a market that didn’t exist before someone did the underlying research. For the funder it might be evidence and proof at small scale that an investment could unlock adoption at a scale that justifies the capital. For the farmer it’s none of those things ,  it’s a much more immediate question, ‘will this tool help me avoid losing a crop this season, and can I trust it enough to act on what it tells me?’ For the government, the tool might be  a possible building block,  but this one deserves more than a label. A ministry of agriculture weighing this tool is usually asking whether it fits inside an existing extension service, who is liable if the advice is wrong, whether it can be scaled without a foreign grant propping it up indefinitely, and whether it survives the next change in administration. “Coordination and governability” isn’t abstract; it’s the difference between a pilot that becomes policy and a pilot that becomes a slide in someone’s annual report.

Five stakeholders, one project, five different reasons to be in the room,  and five different things each is contributing to make it real. None of them are wrong about what the project means to them. What makes it collaborative is that the researcher’s evidence, the innovator’s translation, the funder’s capital, the farmer’s validation, and the government’s legitimacy are all things the others actually need  and none of the five could reach the outcome they want without the others’ contribution somewhere in the chain.

It’s also worth naming what it looks like when this goes wrong, because the idealized version above is the exception, not the rule. A common failure pattern: a funder and a research team co-design a tool, bring in an innovator to build the interface, and only then  often late, often through a single workshop  try to “onboard” farmers into a design that’s already fixed. The farmer’s contribution gets treated as validation of a finished product rather than an input that should have shaped it. The tool launches, adoption is thin, and the postmortem calls it a “user engagement problem,” when the actual problem was that four stakeholders exchanged with each other and invited the fifth to consume the result.

Listening Is How You Find the Exchange

There’s a temptation, once you accept that collaboration means different things to different people, to jump straight to mechanisms and solutions, give researchers a data-sharing agreement, give innovators an API, give farmers a demonstration plot. Reasonable ideas  but different meanings don’t hand you different formats automatically. Something has to sit in between, a way of finding out what a stakeholder actually needs, what they can contribute, and what’s stopping them from doing either.

That in-between step is listening not as the warm, values-driven conclusion of a collaboration process, but as its method, the mechanism by which the exchange gets discovered rather than assumed. Before designing anything and coming up with a stakeholder map as the example above, it’s worth asking, stakeholder by stakeholder, why would this person or institution actually collaborate? What do they hope to get versus what do they merely need? What can they realistically contribute? What constraints , time, trust, technical language, institutional risk ,stand between them and participating at all? A need and a want aren’t the same thing, and neither is the same as an incentive. A funder may need evidence but be incentivized by a story they can take to their own board. A farmer may want higher yields but face a constraint, distrust of an unfamiliar source, a season too tight to risk experimenting,  that has nothing to do with whether the advice is good. Skip that distinction and the mechanism you design solves a problem the stakeholder didn’t actually have.

Otto Scharmer’s Theory U is a useful frame for the quality of listening capable of surfacing this. He distinguishes four levels, from listening that only confirms what you already believe, to listening that notices contradicting data, to listening that steps far enough into someone else’s position to feel it rather than just record it. The level that matters most here is the fourth: generative listening, which listens from an emerging future rather than a fixed past and  holds  space for something genuinely new to take shape rather than defending what already  exists.

Most stakeholder engagement never gets past the first level. We convene, present, and collect feedback that confirms our existing model, and call voila! Successful collaboration. Getting to empathic or generative listening with a farming community means immersing in how the community actually works: what a “good yield” means beyond the number, why a practice persists even when it looks technically inefficient from the outside.

We build tools that require change of behaviour and have assumptions and wonder why they are not adopted.  Ask farmers where they get their information, and radio comes up before any app or dashboard, often enough,  not because radio is more advanced (it’s the opposite), but because it has folded itself into the existing rhythm of the day. Present at dawn, through the hours of tilling, ploughing, weeding, doing double duty as entertainment and information at once. No login, no data plan, no new habit to build. That’s the whole lesson in one artifact: the entry point that works is the one that asks nothing extra of the person entering.

This kind of listening isn’t free. Empathic and generative listening with a farming community takes time most project timelines don’t budget for, and it usually requires someone with standing in that community this could be  a trusted local voice, an extension worker, a long-term local partner,  not the funder or the research team parachuting in for a workshop. If nobody in the collaboration is positioned to do that listening, or resourced to do it slowly, the design will default back to downloading, no matter how good the intentions are.

Holding Different Futures in the Same Room

There’s one more thing in Scharmer’s idea of generative listening that the exchange framing above doesn’t fully capture on its own: he calls it listening from the future,  not from habit, and not even from present-day facts, but from something that hasn’t happened yet and is trying to take shape. Applied here, this points at something beyond needs and contributions: the five stakeholders in the advisory-tool example aren’t just holding different exchanges. They’re holding different futures , different pictures of what “this worked” would even mean, on different timelines. The exchange is what each stakeholder gives and gets; the time horizon is when they expect to know whether it was worth it, and that’s a dimension of the exchange too easy to leave out of the table above.

If we go back to our stakeholders  researcher’s future is a hypothesis validated at scale, a multi-year horizon, measured in publications and replication. The innovator’s future is a viable, defensible product, somewhere in between that has been adopted by mass. The funder’s future is a market mature enough that the next round of capital de-risks itself,  a three-to-five-year horizon, measured in adoption curves. The government’s future stretches longest of all,  an institution or extension system resilient enough to outlast the people currently in the room, measured in administrations rather than seasons. The farmer’s future is the shortest and, at the moment, the most consequential: will this crop survive the season that’s already underway.

None of these time horizons are wrong. But a collaboration process with the bandwidth to hold only one future at a time will almost always default to whichever stakeholder has the most resources or the loudest institutional voice. That’s usually the funder’s three-year horizon or the innovator’s product roadmap. The farmer’s one-season future quietly gets treated as a rounding error to be managed, rather than a future with equal claim to shape the design  and this is exactly where the power imbalance named earlier in the stakeholder table shows up again, now as a difference in whose clock the project runs on.

The Quintuple Helix’s theory is worth taking seriously here,if the environment itself belongs at the table as a stakeholder with its own future, a farming community’s much shorter, much more immediate future belongs there with at least as much weight. Designing collaboration well means respecting everyone in the room, not  picking one future and asking everyone to align behind it. It means  holding space for every future in the room without breaking.