Systems Thinking Is Not a Metaphor. It Is the Method

Strategic partnerships fail most often not because of bad intentions, but because they were designed without understanding the system they were entering. When you map the interdependencies first, everything else follows more cleanly than you expected.

The Recurring Pattern

Between 2013 and 2024, Canada’s Public Health Agency funded 30 multisectoral partnerships for chronic disease prevention. By the fourth year, agency staff had documented a troubling pattern: partnerships with identical mandates, similar funding levels, and comparable technical capacity were producing wildly different results. Some transformed community health outcomes. Others collapsed into recrimination before their second year concluded.

The difference was not commitment. It was structure.

Researchers examining 12 country case studies of multisectoral collaboration in 2021 identified what practitioners had long suspected but rarely articulated: partnership failure is overwhelmingly structural, not interpersonal. The absence of coordination mechanisms, the lack of environmental intelligence functions, and the failure to define shared purpose operationally create the conditions for breakdown regardless of goodwill.

This is where systems thinking stops being a metaphor and becomes a method.

The Viable System Model: Architecture for Partnerships That Last

In the 1970s, British cybernetician Stafford Beer developed the Viable System Model whilst working with the Chilean government. His question was precise: what structural functions must any system possess to survive and adapt in a changing environment? His answer became one of the most rigorous frameworks for organisational design ever constructed.

The model identifies five essential functions. If any are absent, the system becomes fragile. It may function during stable conditions, but it breaks under stress.

System 1: Operations. These are the doing partners, each delivering a distinct function. In a water and sanitation partnership across Sub-Saharan Africa, System 1 comprised Oxfam (community trust-building), UNICEF (supply chain logistics), state government (regulation and epidemiological surveillance), and women’s groups (last-mile distribution). Each unit was viable enough to operate independently if others exited.

System 2: Coordination. This prevents operational units from working against each other. It is not meetings or communication channels. It is anti-oscillation protocols: stakeholder maps defining lead responsibility per geographic zone, shared data reporting formats, escalation procedures before funders become involved. Without System 2, partners inadvertently undermine one another’s work with the same communities.

System 3: Operational Control. Somebody must optimise resource allocation across the whole system, not for individual institutional mandates. In the Sub-Saharan partnership, UNICEF held operational control not because it was the largest funder, but because it possessed supply chain infrastructure the others lacked. Authority to make resource decisions was pre-agreed by all partners.

System 3: The Algedonic Loop*. This is the emergency bypass. Critical information, particularly bad news, is filtered and softened as it moves through normal reporting channels. The algedonic loop allows urgent signals to reach governance directly. When laboratories in the water partnership detected unsafe contamination results, an automatic alert reached operational leadership within 24 hours, not three months later via quarterly reports.

System 4: Intelligence and Learning. The partnership must sense changes in its environment before they become crises. This combines epidemiological data, climate modelling, community perception research, and policy monitoring. Intelligence without a defined pathway from signal to decision is merely information.

System 5: Identity and Purpose. This is not a mission statement. It is an active governance function that resolves conflicts between competing institutional interests. When operational efficiency conflicts with strategic adaptation, System 5 adjudicates. It defines what the partnership will never do, regardless of operational pressure.

What Mapping Interdependencies Actually Looks Like

The Sub-Saharan water partnership faced annual cholera outbreaks that had become normalised. Four types of actor were present: development agencies, government, technology partners, and communities. Each viewed the problem through incompatible lenses, and none had structural authority to change that.

The Viable System Model was applied as design architecture. Before writing terms of reference or signing memoranda, the partnership diagnosed which functions existed, which were absent, and which were contested.

System 5 was negotiated first. The state government needed emergency response framed prominently. International partners needed community resilience foregrounded. The final purpose statement held both: end preventable waterborne disease, build community resilience, reduce emergency dependence. Neither actor was required to abandon their institutional mandate.

System 2 protocols were written explicitly: monthly coordination meetings with rotating chair, a stakeholder map defining lead partner per community zone, a shared data reporting format with agreed definitions, an escalation protocol before funders were involved. These were not aspirational. They were structural requirements with enforcement mechanisms.

The algedonic loop was designed with triggers and timelines: what qualified as an emergency (unsafe water lab results), who could activate it (independent laboratories), what authority it carried (immediate activation of emergency protocols with parallel notification to governance).

Within months, outbreak frequency dropped. Critically, response time when outbreaks occurred reduced from weeks to days. Community belief shifted from fatalism to agency. The state government moved from emergency response framing to prevention investment, not through advocacy but because System 4 intelligence made the systemic pattern undeniable.

When UNICEF underwent an internal restructure that reduced its operational presence, the partnership held. Each System 1 unit had been designed for autonomy from inception. No single partner was a systemic point of failure.

Why Traditional Project Management Tools Break Down

Project management excels at handling complicated problems. A construction project has many moving parts, but it follows fixed rules. The same inputs produce predictable outputs. Expertise solves it.

Partnerships are not complicated. They are complex. Actors adapt and respond to one another. No single expert holds the complete picture. The same inputs produce different outputs depending on context, timing, and the adaptive behaviour of other partners. Best practice rarely transfers cleanly.

This is why partnership coordination meetings often feel like displacement activity. They address symptoms (communication gaps, misaligned expectations, duplicated effort) without resolving the structural absence that created those symptoms. Partners communicate constantly but still work against each other. Information flows but decisions do not change. Coordination becomes performative rather than functional.

The Viable System Model does not require consensus to function. It creates structure that allows disagreement without dysfunction. Partners in the Sub-Saharan case still disagreed, but within a system designed to make that conflict structurally productive.

The Diagnostic Questions

Before designing a new partnership or repairing an existing one, six questions reveal structural fragility:

Can every partner write down the partnership’s purpose in one sentence, and do they match? If not, you have competing mandates pretending to be a partnership.

What prevents Partner A from inadvertently undermining Partner B’s work with the same community? If the answer is goodwill or effective communication, System 2 is absent.

Who has authority to allocate resources across partners, and is that pre-agreed? If no one or if everyone claims it, System 3 is absent or contested.

How would a critical failure reach governance in under 48 hours? If it would not, the algedonic loop does not exist.

What external intelligence changes partnership decisions? If intelligence does not change decisions, you have a reporting system, not System 4.

If your largest partner exited today, which functions would collapse? If the answer is most of them, operational units are not viable. They are dependent sub-contractors.

From Metaphor to Method

Systems thinking has been co-opted into development discourse as a framing device. Actors are encouraged to think systemically, to see connections, to appreciate complexity. This is useful as far as it goes, but it stops short of actionable design.

The Viable System Model is systems thinking operationalised. It does not ask you to appreciate interdependence. It requires you to map it structurally, assign functions explicitly, and design protocols that prevent predictable failures.

Research from the World Bank and OECD Development Assistance Committee shows that multisectoral collaborations mobilising $210 billion annually in official development assistance still struggle with basic coordination failures. Private Sector Instruments, introduced in 2019 to encourage private investment mobilisation through partnerships, required new rules by 2024 precisely because structural design had been neglected.

Strategic partnerships entered without mapping interdependencies first are not partnerships. They are parallel programmes with coordination costs. When you map the five systems before negotiation begins, you design for viability from the start.

The interdependencies do not care whether you acknowledge them. They exist. The question is whether you designed for them.

References

Beer, S. (1981). Brain of the Firm. John Wiley & Sons.

Convergence. (2024). State of Blended Finance 2024. Toronto: Convergence Finance.

Johnston, L. M., & Finegood, D. T. (2015). Cross-sector partnerships and public health: challenges and opportunities for addressing obesity and noncommunicable diseases through engagement with the private sector. Annual Review of Public Health, 36, 255-271.

Public Health Agency of Canada. (2020). Multi-sectoral Partnerships to Promote Healthy Living and Prevent Chronic Disease. Health Research Policy and Systems, 18(1), 1-14.

Reich, M. R. (2002). Public-private partnerships for public health. In Public-Private Partnerships for Public Health (pp. 1-18). Harvard Center for Population and Development Studies.

Woulfe, J., Oliver, T. R., Zahner, S. J., & Siemering, K. Q. (2010). Multisector partnerships in population health improvement. Preventing Chronic Disease, 7(6), A119.