Coordination creates demand for intermediaries. Intermediaries make coordination manageable. Neither stall requires an exclusionary decision. Present in 117 of 221 diagnosed ecosystems across 44 countries — the most structurally prevalent stack in the dataset. Together they form a system that coheres beautifully and never has to change.
Every meeting creates demand for a broker. Every broker justifies the next meeting. Neither behaviour is irrational — both are the cheapest available response to complexity. The system absorbs uncertainty, complexity, and pressure without structural adaptation. Understanding why this is the most common stack requires following the money, the meetings, and the actor incentives.
Coordination (S2) and intermediation (S5) are the two most common structural behaviours in regional innovation ecosystems. Separately, each is a rational response to complexity. Together, they form a self-sustaining loop: coordination structures — working groups, MoU frameworks, partnership boards, alignment processes — create transaction costs that intermediaries absorb. Intermediaries — accelerators, cluster organisations, technology transfer offices, innovation agencies — provide the brokerage that makes coordination manageable. Remove one and the other regenerates it.
The mechanism is not subtle. When a steward creates a multi-stakeholder working group, the group produces coordination overhead — agendas, minutes, follow-up actions, relationship management. That overhead creates demand for an intermediary to manage it. When an intermediary brokers connections between actors, those connections create coordination requirements — alignment meetings, partnership frameworks, reporting obligations. Each stall feeds the other's demand.
This is not a failure of leadership. It is a structural property of ecosystems under complexity pressure. The actors performing the coordination are responding rationally to a real problem: multi-stakeholder environments are genuinely hard to navigate. The intermediaries absorbing the coordination overhead are providing a real service: they reduce transaction costs for individual actors. The problem is that the combination is self-sustaining. It never needs to simplify.
117 of 221 diagnosed ecosystems — 53% of the dataset. 44 countries. Present in growing ecosystems (52), mature ecosystems (34), and emerging ones (14). Corporate-anchored ecosystems show it most frequently (62), followed by university-anchored (37) and government-anchored (11). It is not a phase-specific pattern. It is not a sector-specific pattern. It is structural.
The corporate dominance is telling. Corporate procurement processes are inherently complex — multiple approval layers, compliance requirements, vendor management frameworks. That complexity creates genuine demand for intermediation. The intermediation then generates coordination requirements that perpetuate both stalls. University-anchored ecosystems show the same pattern for a different reason: research partnership frameworks create coordination overhead that technology transfer offices absorb.
The government-anchored finding is equally instructive: only 11 of 38 government-anchored ecosystems show this stack. Government ecosystems mediate heavily (S5 appears in 55% of them) but their coordination structures are more directive — they decide rather than align. The coordination-intermediation loop requires the system to maintain alignment as its primary output. Government anchors are more likely to skip alignment and mandate.
Boston Robotics & AI (87 evidence items, medium confidence). Coordination through multi-stakeholder structures — MassRobotics, working groups, industry partnerships — sustains mediation through those same intermediary bodies rather than enabling direct coupling. The coordination structures create demand for the intermediary infrastructure. The intermediary infrastructure makes coordination appear necessary. The system absorbs uncertainty about partner selection, pressure to maintain network access, complexity of multi-party technical integration, and disruption risk from exclusion. The reinforcing logic is visible in how new entrants navigate the ecosystem: they join MassRobotics (intermediary) to access the partnership network (coordination), which generates the coordination activity that justifies MassRobotics' existence.
Detroit Mobility & Auto Tech (95 evidence items). Partnership formation activity sustains intermediary organisations by generating coordination demand. Intermediary presence enables continued partnership formation by reducing transaction costs. The system absorbs uncertainty about technology evolution, complexity of multi-domain integration, and pressure around capital deployment expectations. Detroit's leverage hypothesis converges on transparency: if joint venture agreements were required to disclose decision authority and exit conditions, the system's ability to absorb uncertainty through partnership formation would be reduced. The cost of coordination would become visible.
Tel Aviv AI & Deep Tech (87 evidence items, medium confidence). Extensive network infrastructure without exclusionary decisions co-occurs with accelerators and platforms mediating connections. The system maintains multiple pathways simultaneously — university support units, investment networks, accelerator programmes — each intermediating connections that could form directly. Complexity is absorbed by intermediaries who reduce transaction costs. The question the system never answers: which connections would form without them?
Uncertainty — about partner selection and commitment. The intermediary absorbs the risk of choosing wrong by keeping all options open. The coordination structure absorbs the pressure to commit by maintaining alignment as the primary output.
Complexity — of multi-stakeholder environments. Intermediaries reduce transaction costs. Coordination structures manage relationships. Neither requires the system to simplify by making exclusionary choices.
Pressure — to maintain stakeholder buy-in. Coordination demonstrates inclusion. Intermediation demonstrates activity. Together they produce visible ecosystem health signals that satisfy funders and boards without requiring throughput evidence.
Disruption — from new actors or technologies. The coordination-intermediation loop absorbs new entrants by routing them through existing structures. Disruption is managed rather than leveraged.
The system is not broken. It is working exactly as designed. It absorbs the signals that would otherwise force structural adaptation — and it does so at a cost that no individual actor has an incentive to measure.
Across 117 ecosystems, the leverage hypotheses converge on two structural interventions:
1. Making intermediation costs visible. The intermediary layer has a cost — in time, in conversion rate, in opportunity. But that cost is invisible because the alternative (direct coupling) is structurally unavailable. The leverage is not to remove intermediaries. It is to make their cost measurable. If outcomes from direct partnerships were made observable alongside intermediated ones, the system's ability to absorb complexity through intermediation without exposing comparative performance would be reduced. This is what Tel Aviv's leverage hypothesis targets: making direct university-corporate partnership outcomes visible to intermediary organisations.
2. Creating one direct coupling pathway. One direct actor-to-actor connection that succeeds without routing through the intermediary layer demonstrates that the dependency is optional. Not all connections — one. The demonstration changes the risk calculus for subsequent actors considering whether to use the intermediary pathway. Boston's data points here: if one robotics startup secured a corporate partnership without MassRobotics making the introduction, the information that direct coupling is possible would propagate.
What to watch for: the leverage test takes 6-12 months. Resistance is diagnostic — if the intermediary layer actively prevents direct coupling from being attempted, that is the strongest possible evidence that the stack is structural, not incidental. If direct coupling is attempted and fails, the intermediary layer was providing genuine value. Both outcomes are informative.
"How many connections in the last quarter formed without an intermediary making the introduction?" If the answer is close to zero, the intermediation dependency is structural, not optional. The ecosystem has made direct coupling structurally unavailable.
"Could any two actors in your ecosystem have found each other without the cluster organisation?" If the honest answer is no, the system has been designed — intentionally or not — to route all connections through a broker.
"What would happen if the intermediary layer went on holiday for three months?" If the answer is "nothing would connect" — that is the diagnostic finding. If the answer is "some things would connect directly" — those connections are the leverage point. They prove the dependency is optional for at least some actors.
Stack confidence is higher than individual stall confidence because the reinforcement mechanism can be directly observed: coordination structures generating intermediary demand and intermediary activity generating coordination requirements are visible in programme reporting, governance minutes, and the temporal relationship between partnership announcements and intermediary activity. What requires more care is distinguishing between intermediation that genuinely reduces transaction costs (a structural service) and intermediation that perpetuates coordination overhead (a structural stall). The test is comparative: outcomes from intermediated vs direct connections.
Is this stack operating in your ecosystem? The interactive self-diagnostic identifies your stalls, detects reinforcing stacks, and generates leverage hypotheses — calibrated against 221 comparable ecosystems.
Diagnose your ecosystem →