Donella Meadows Was Not a Management Consultant Either

Donella Meadows was a systems scientist at Dartmouth College who spent her career making systems thinking accessible to people who weren't systems scientists. She was lead author of The Limits to Growth (1972), the Club of Rome report that modeled planetary resource depletion decades before it became mainstream concern. She died in 2001, a year before her masterwork Thinking in Systems was published posthumously.

Her 1999 paper "Leverage Points: Places to Intervene in a System" was written for sustainability activists but immediately recognized as applicable to anyone trying to change complex systems — organizations, communities, educational programs, supply chains. It was one of the most forwarded documents on early internet sustainability lists. Its core argument is blunt: most interventions target the wrong places, and the right places are counter-intuitive.

The list she developed — 12 leverage points ranked from least to most effective — is not a management framework. It is a diagnostic tool. It asks: where in this system are you actually intervening, and why do you think that will produce the change you want?

Where Practitioners Almost Always Intervene

The lowest leverage points are constants and parameters — the numbers in the system. How many hours of training. How many modules. How many competencies in the framework. How much budget. These are the things that are easiest to measure and therefore easiest to report on. They are also, Meadows argues, almost never the thing that is actually constraining the system.

L&D interventions at levels 1–5 look like: changing the number of training days, increasing practice time, adding a module on a missing topic, reducing class size, increasing manager briefing time. All of these can make marginal differences. None of them change the system's behavior because none of them touch the feedback loops, the information flows, or the goals that are actually driving behavior.

Peter Senge made the same point in The Fifth Discipline (1990): organizations repeat their worst patterns not because the people in them are irrational but because the structure of the system produces those patterns regardless of who occupies the roles. Changing the people — the "parameters" — without changing the structure produces the same behavior with different faces.

Information Flows, Rules, and Who Gets to Change What

Mid-leverage interventions change how information moves through the system (level 7), which rules govern behavior (level 8), or who has the power to add or modify system structure (level 9). These are substantially more powerful than parameter changes, and substantially harder to implement.

In an organizational L&D context, level 7 looks like: ensuring that managers receive real-time performance data that shows whether training is transferring, rather than end-of-year aggregates that arrive too late to act on. Level 8 looks like: changing the rule that says learning objectives are set by the training team rather than by the learners and their managers together. Level 9 looks like: giving L&D professionals the authority to redesign job structures that make transfer impossible, rather than only designing programs that try to work within those structures.

Jay Forrester, the founder of system dynamics at MIT, wrote in 1971 that "the human mind is not adapted to interpreting how social systems behave." His point was that the interventions that seem most powerful — the ones that feel most direct — are often the lowest leverage. The mid- and high-leverage interventions are not visible in the way a training day or a competency framework is visible. They require intervening in things that don't look like training.

Goals, Paradigms, and the Power to Change the Game

The highest leverage points are also the most counter-intuitive and the most resistant to deliberate change. Level 10 is the goal of the system: if the goal of the training function is to deliver a certain number of training hours, and not to produce a certain level of performance, everything in the system will align to deliver hours regardless of what it costs in actual development.

Level 11 is paradigm: the shared assumptions from which the system arises. The paradigm that "training is what you do when performance is inadequate" produces a very different system than the paradigm that "development is the primary mechanism through which we pursue our goals." Changing this level — actually shifting the underlying belief — is the most powerful intervention available. It is also the one that most practitioners have no tools for and no mandate to attempt.

Level 12 — the ability to transcend paradigms, to hold multiple framings simultaneously — is what Meadows calls the "meta" leverage point. It is not something you implement. It is a capacity you develop. Practitioners who operate at this level don't argue for their framework; they see the framework as one possible framing and work flexibly within and between multiple paradigms depending on what the system needs.

High-Leverage Practitioners Charge More. Here's Why.

The practitioner who works at levels 1–5 is selling a deliverable: training hours, modules, assessments, certifications. These are commodities. Other practitioners can produce them. Technology is reducing the price of the commodity version every year. Working at levels 1–5, regardless of how well you do it, puts you in a market that is becoming more competitive and less differentiated.

The practitioner who works at levels 7–10 — who can diagnose where in the system the actual constraint lives, who can design information flows and challenge incentive structures and work with leadership on system goals — is not selling a deliverable. They are selling a capability that most organizations cannot develop internally and cannot easily replace. The market for this is less crowded and the price is structurally higher.

This is not about positioning or personal branding. It is about what you can actually do. Meadows' leverage hierarchy maps directly to practitioner economic value: the higher the leverage point you can work at, the rarer your capability and the more valuable your outcomes. The path from the commodity layer to the high-leverage layer runs through diagnosis — the ability to see where in a system the actual intervention point is and to articulate it clearly enough that a client can authorize the work.

Why Most Training Doesn't Stick — From a Systems View

Training transfer research consistently finds that approximately 10–20% of what is learned in training transfers to job performance. This has been documented repeatedly since Mary Broad and John Newstrom's landmark review in 1992. The response has typically been to improve the training — to add practice, to improve alignment, to strengthen the learning design. These are level 1–5 interventions. The 10–20% figure has not moved significantly in thirty years.

From a systems view, this is exactly what you'd expect. The training system and the work system are two separate systems with different feedback loops, different incentives, and different information flows. Improving the training system doesn't change the work system that participants return to. The work system's feedback loops — what gets rewarded, what gets punished, what managers actually model — are substantially stronger than the training system's feedback loops. The training signal is overwritten by the work system signal within weeks.

The implication is not that training should be better. It is that practitioner interventions at levels 7 and above — in the work system, not just the training system — are required for the 10–20% figure to move. This is a structural diagnosis, not a performance critique. The system is working exactly as designed.

Diagnosing Where You Actually Work

The Leverage Audit: List the five most significant things you did in your last major engagement. For each, identify the leverage level (1–12). Where did most of your work land? What does that tell you about your current market position?

The System Diagnosis: For a current client or program, map the system: what are the feedback loops that are reinforcing the behavior you're trying to change? What level would you need to work at to actually move those loops?

The Proposal Test: When you write a proposal for new work, what leverage level is the work implicitly promising? If it's levels 1–5, what would change if you proposed working at level 7 or 8 instead? What would you need to be able to demonstrate to make that case?

The Income Question: Meadows noted that high-leverage interventions are often the cheapest to implement and the hardest to see. If you could charge for the diagnosis — identifying which leverage point actually needs to move — what would that be worth? How does that compare to what you currently charge for the training itself?

Sources: Meadows (1999 paper) · Meadows (2008 book) · Senge (1990) · Sterman (2000) · Forrester (1971) · Senge (1994 Fieldbook) · Broad & Newstrom (1992) · Saks & Belcourt (2006) · Club of Rome / Meadows et al., Limits to Growth (1972)