TL;DR: Most transformations fail not from bad execution but from bad design — too much order, imposed on systems too complex to control that tightly. The fix isn’t less structure. It’s structure in the right place: lock down outcomes and non-negotiables, release control over how people get there. Amazon, Netflix, Microsoft, and Haier all do a version of this, even without the language. It’s not a case against control — it’s a case for putting it where it actually helps.
It was month fourteen of an eighteen-month transformation programme.
The programme office had a dashboard. Forty-seven workstreams. A RAG status report that went to the steering committee every fortnight. Everything was Green.
The adoption rate was 23%.
Not because the people were incompetent. Not because the technology failed. But because somewhere in the design of the programme, someone made a decision — probably unconscious, probably well-intentioned, but deeply wrong — to impose order on a system that needed something else entirely.
Déjà vu, anyone?
In an earlier piece, I wrote about being “in command but out of control.” The response told me that this resonates with something leaders feel but rarely name. Writing it surfaced a harder question that I couldn’t shake: Why does it work?
Because “trust your team” is easy to say and surprisingly hard to believe — especially when your name is on the outcome. Intuition isn’t enough. When leaders are under pressure, they tend to fall back on what feels safe And what feels safe is control.
The answer I found doesn’t come from management consulting or leadership theory. It comes from the science of complex systems — from physicists and biologists studying how order emerges in nature. And it turns out they figured out something about organisations that most boardrooms still haven’t caught up with.
It’s called the edge of chaos. (If this world interests you, James Gleick’s Chaos: Making a New Science is the classic entry point, and Mitchell Waldrop’s Complexity tells the Santa Fe story.) And it explains a lot.
The Uncomfortable Statistic
A figure you’ll hear constantly: roughly 70% of transformations fall short of their objectives. The exact number is debated, but the direction isn’t.
And the needle hasn’t moved much in two decades. That’s the part that should really bother us.
Think of it this way – if this were simply an execution problem — the wrong technology, the wrong vendor, the wrong project manager —should’ve been off the table by now. We’ve had twenty years. Better tools. Smarter methodologies. Entire consulting practices dedicated to “change management.”
So maybe it isn’t an execution problem. Maybe it’s a design problem. Maybe there’s something fundamentally wrong with how we conceive of transformation itself.
The Framework: Edge of Chaos
Starting in the late 1960s, and gathering pace at the Santa Fe Institute in the 80s and 90s, a quiet revolution was happening in physics, biology, and mathematics. Scientists like Stuart Kauffman were studying a deceptively simple question:
How does order arise in nature?
Not imposed order. Nobody tells a flock of starlings how to murmur. Nobody tells an embryo how to differentiate into organs. They were studying spontaneous order — structure that emerges from interaction rather than instruction.
What they found was a spectrum. At one end, you have rigid order. Systems locked into fixed patterns. Stable, but unable to adapt. At the other end, you have chaos. Systems with so many degrees of freedom that nothing coherent can form.
And somewhere in between — in a narrow, unstable, immensely productive band — they found what became known as the edge of chaos. This is the zone where systems are structured enough to maintain coherence, but loose enough to adapt, learn and reorganise. It’s where ice becomes water. Where evolution happens. Where complex systems do some of their most interesting work.
Now apply that idea to organisations. An over-ordered organisation is stable, but brittle. Rules for everything. Approval chains for every decision. Clear hierarchy. Predictable outputs. Until the environment shifts. And then the whole structure starts to crack because it was never designed to flex. If you have been in corporate for long enough you may have experienced this.
A chaotic organisation has the opposite problem. It can be creative, but incoherent. Ideas everywhere. Execution nowhere. Energy without direction. The brilliant startup that implodes at fifty people because nobody knows who decides what.
The edge of chaos is the narrow band between these two failure modes. And here’s the uncomfortable insight for anyone designing a transformation programme:
Most of them are engineered at the wrong end of the spectrum.
What Does Over-Ordering Actually Looks Like?
Here’s what an over-ordered transformation programme looks like from the inside. You’ve probably seen this in play –
A blueprint gets created in Month One and then treated as sacred ( reminds one of the classical waterfall model) — revisited only when something breaks badly enough to force the issue.
Change management runs as a separate workstream, bolted on rather than built in. It’s often staffed by people who joined the conversation after the real decisions were already made. Success metrics get defined upfront (vaguely mostly) and then never revisited, even when the business context underneath them has shifted completely.
Middle managers become implementers of a plan they didn’t shape or inform — rather than sense-makers who could tell you, three months earlier than the dashboard, that something isn’t landing.
And every deviation from plan gets treated as a risk to be corrected, rather than a signal to be understood. None of this comes from incompetence. It comes from a reasonable, well-intentioned instinct:
Complexity is frightening, so we control what we can. We write the detailed plan because detailed plans feel like safety.
But a transformation programme isn’t a bridge. It’s not a fixed engineering problem with a knowable answer. It’s a complex adaptive system — thousands of individual humans, each recalibrating their behaviour every day based on incentives, trust, fear, and what they saw happen to the last person who spoke up.
Plan it like a bridge, and the plan can become the very thing standing between you and the reality on the ground.
The “ticket-taker” problem I described in my last piece is a symptom of exactly this. Not a leadership failing. A systems design failing. People stop bringing judgment because the system was built to run on compliance instead.
What Does the Edge Look Like in Practice?
Organisations that transform well don’t necessarily have less structure. They have different structure. They set attractors, not blueprints.
A small number of non-negotiables — the outcome, the timeline, the values that can’t be traded away — and then they release the “how” to the people closest to the work. They treat resistance as information.
When adoption stalls, the instinct isn’t, “How do we push harder?” It’s:
“What is this telling us that the plan didn’t anticipate?”
They build feedback loops that actually change the plan. Not quarterly steering committees that simply rubber-stamp the original blueprint, but short cycles where what’s learned this week reshapes what happens next week.
And they give middle managers interpretive authority, not just execution authority. They trust them to translate intent into local reality rather than treating them as a relay for instructions.
This isn’t a claim that every successful company does this deliberately. Nor is it a claim that structure itself is the enemy. It’s a claim about matching the level of order to the level of genuine uncertainty. And when you look closely at organisations known for both scale and adaptability, you find a recurring architecture — even when they never use this language for it.
Satya Nadella’s early moves at Microsoft are a useful public example. He didn’t hand down a fixed transformation blueprint.
He set a small number of clear intents — growth mindset, cloud-first, customer obsession — alongside some very decisive moves on direction, and then spent years letting thousands of individual teams figure out what that meant for their own work.
The structure was in the intent. The freedom was in the execution.
Amazon separates reversible decisions from irreversible ones — pushing the former down to whoever is closest to the problem, while reserving the latter for senior judgment.
Netflix built an entire culture around “context, not control” — replacing approval chains with high-trust information sharing.
Haier, the Chinese appliance giant, went furthest of all — breaking itself into thousands of micro-enterprises, each with real autonomy over hiring and pricing, while operating inside a shared strategic architecture.
Different industries. Different language. Same underlying logic.
This is the edge of chaos, operationalised. Not the absence of control. A different relationship with it.
The Leader’s Reframe
If there’s one sentence to take out of this, it’s this:
Your job is not to impose order on a complex system. Your job is to set the conditions under which order can emerge.
That’s a different kind of leadership from what most of us were trained for. We were taught to plan, control and correct deviation. The edge of chaos asks something harder:
Define the intent precisely. And then tolerate the discomfort of not controlling exactly how it unfolds.
The Question to Ask
So here’s the question worth sitting with before your next transformation kicks off:
Where on the order-chaos spectrum is your programme actually designed? And is that where your organisation needs to be — or is it simply where it feels safest to start? One honest caveat, because this framework deserves scrutiny, not applause:
This isn’t a case against structure. And it isn’t a guarantee.
Some safety-critical, physically irreversible operations genuinely need rigid order. You don’t want a “self-organising” team improvising on a factory floor where the physics is unforgiving.
And releasing control without real accountability isn’t the edge of chaos. It’s just chaos with better branding. The leader remains accountable for the outcome and the boundaries. What moves is where the judgment sits within them.
Somewhere between the rigid blueprint and the free-for-all is a narrow, uncomfortable, productive space. Look around and you’ll find that many of the more successful transformations don’t avoid that discomfort.
They are designed for it.
