Lord of the Flywheels—SERD's Denholm Review Arrival

Jim Cooper*, 31 March, 2026

This article draws on a longer piece published as a blog on the author’s LinkedIn

The flywheel, when properly conceived and executed, creates both continuity and change. On the one hand, you need to stay with a flywheel long enough to get its full compounding effect. On the other hand, to keep the flywheel spinning, you need to continually renew and improve each and every component.

In Jim Collins’ work, particularly Good to Great and Turning the Flywheel, he describes how enduring organisations build momentum through reinforcing activities that feed back into themselves over time.

The insight was not just that successful companies had flywheels, but that they understood them deeply—how each component connected, how momentum was built, and how discipline over long periods of iteration created compounding advantage.

The flywheel is not something you declare. It is something you discover, test, and refine through rigor and method. It assumes direction is correct, and that the challenge is persistence. And it only works if you’re playing the right game to begin with. That is the standard Collins sets. It is also the standard the Denholm Review does not meet.

If I asked you to name the world’s leading innovation ecosystems—Silicon Valley, Kendall Square, Shenzhen, DARPA, Boston’s biotech corridor, even Lockheed’s Skunkworks—what makes them enduringly great? This is not about imitation. It is about understanding what they actually have—structurally, culturally, financially—that creates momentum and reinforces it over time.

Now ask a harder question: what parts of the Australian innovation ecosystem would you take with you to a desert island? Strip away legacy systems, inherited structures, and polite fictions. What survives first contact with reality? When you combine the best of what exists globally with what we genuinely have, and remove what does not function, what do you actually see for Australia?

The Denholm Review leans heavily on the flywheel metaphor to describe Australia’s innovation system: research feeds innovation, innovation feeds firms, firms feed growth, growth feeds reinvestment. It is neat, intuitive, and reassuringly controllable. But if the direction is wrong, or if the system is not connected to the forces that actually generate momentum, then all we are doing is perfecting a loop that does not scale.

Momentum does not come from internal consistency alone. It comes from interaction with markets, capital, and competition. The world’s leading ecosystems are not closed systems refining their own motion. They are open systems embedded in global flows of talent, capital, and demand. The flywheel, as used here, risks turning that outward dynamic into an inward one.

The SERD report diagnoses problems that are broadly correct: low business R&D intensity, declining public research investment, weak economic complexity, and fragmented policy architecture. These observations are not controversial. But the report assumes the primary failure mode is coordination—that if the right councils sit in the right rooms with the right reporting lines, the system will produce startups and industries.

That assumption is wrong. Innovation systems do not fail primarily because of coordination problems. They fail because of market formation problems. If markets do not exist, if capital does not flow into risk, and if talent cannot scale ventures into global markets, no amount of governance reform will create industries.

The emphasis on “coherence” reflects this misunderstanding. Coherence is not created through proximity or conversation, but through shared incentives, overlapping problems, and aligned pathways to market. Forcing adjacency without purpose creates noise. Institutions in high-performing ecosystems do not orchestrate behaviour; they remove friction and allow markets to form. Silicon Valley is not centrally coordinated. It is a globally integrated system, distributed across talent, capital, and execution, moving faster than any institution can manage.

The report’s “pillars” illustrate the same issue. Health, agriculture, defence, resources, and technology are not pillars; they are sectors. They reflect how government is organised, not how innovation systems function. There is no shared logic binding them together, no mechanism through which activity in one reinforces momentum in another. Real coherence is created through shared incentives and the ability to move across boundaries in pursuit of opportunity. The best ecosystems organise around problems, capabilities, and capital—not categories.

The report is widely described as “bold” and “ambitious.” It is neither. It refines and consolidates but does not fundamentally reimagine. Translation funding, while important, does not solve the problem of scaling companies from promising to global. More funding at the front end does not produce better outcomes at the back end without capital, markets, and networks.

The proposed National Innovation Council risks becoming another layer in an already bureaucratic system. Coordination should enable, not direct. The moment it becomes central authority, it risks becoming a constraint.

There is also a persistent belief that government can “pick winners.” Governments are not well positioned to determine what markets will value. Their role is to remove barriers and create conditions for winners to emerge. The R&D Tax Incentive continues to favour incumbents, doing little for small, emerging teams. The proliferation of RD&I programs has created a labyrinthine funding landscape, yet the solution still assumes government funding is the central lever. Innovation happens when markets pull technologies into existence. Government funding cannot substitute for demand.

On capital, the report gestures toward reform but stops short. Australia does not need one improved mechanism; it needs many competing, iterating, and failing. Capital formation remains constrained, limiting who can participate and how quickly funds can emerge. Global capital flows are underemphasised, despite being essential for learning and scale. At the same time, the system continues to privilege institutional research pathways, even as much R&D now occurs within companies and startups operating under different constraints.

The report’s workforce recommendations are sensible but incomplete. The issue is not simply talent supply, but the absence of large-scale firms capable of absorbing and deploying that talent. Without such firms, talent will continue to leave. Founders, meanwhile, are already ambitious—operating in uncertainty, absorbing failure, and pushing against system constraints. The role of policy is not to manufacture ambition, but to ensure it is not systematically exhausted.

Taken together, the Denholm Review is not a radical departure. It is a careful extension of the current system. It improves mechanisms and clarifies some structures but does not change how capital is formed, how companies scale, or how Australia connects to global markets. It is incremental reform presented in the language of ambition. That distinction matters, because it shapes expectations. We should not confuse movement with progress, or optimisation with transformation.

Australia is making the innovation bed it will have to lie in. The system reflects accumulated choices: coordination over competition, grants over markets, committees over capital formation, and domestic alignment over global integration. These choices are defensible in isolation, but together they produce a system that is internally coherent and externally uncompetitive.

The question is whether we are building an ecosystem we can comfortably manage, or one that can win. Because systems that produce globally competitive companies are not neat or centrally directed. They are messy, capital-intensive, globally entangled, and uncomfortable.

What is missing is not effort or intent, but an “Imagination of the Possible.” Not ambition expressed through funding envelopes and coordination bodies, but a willingness to conceive of an ecosystem that does not yet exist, and design for it. Until that shift occurs, we will continue to refine the system we are comfortable with, rather than create the one we actually need. And right now, we are still circling, not imagining the possible.

References

Collins, J. (2001). Good to great: Why some companies make the leap and others don’t. HarperBusiness.

Collins, J. (2019). Turning the flywheel: A monograph to accompany Good to Great. HarperBusiness.

Australian Government. (2025). Strategic Examination of Research and Development (SERD): Denholm Review. Canberra: Commonwealth of Australia.

*Jim Cooper works at the intersection of startups, innovation ecosystems, policy, and capital, with a career shaped by practical engagement across these domains. An Australian who divides his time between California and Sydney, he operates across venture activity, commercialisation pathways, and ecosystem design, bringing an applied perspective to how innovation systems function in practice.

He teaches innovation at the University of California and works with US-based accelerators, including LARTA Institute and StartBlue. His advisory roles extend to Australia, where he serves as Senior Board Advisor to Software Australia and as External Industry Advisor to the California State University Biotech program.

He is also Cofounder and Director of Unfinished Ventures, an advisory firm focused on supporting innovation and venture development. Over a career spanning more than three decades, he has advised more than 300 startups, mentored across a wide range of accelerator programs, and led one himself.

Alongside his professional work, he has been a senior football coach with the Orange County Giants AFC. His work is guided by a practical orientation, with an emphasis on being useful, and a continuing interest in understanding why seemingly sensible ideas often fail to materialise, while less effective alternatives persist.