How to lead through complexity using experimentation and humility

Executive overview

Most leadership challenges are complex, not complicated. Complicated problems have knowable solutions that stay solved; complex ones are emergent, context-dependent, and uncontrollable.

The instinct to control, predict, and avoid failure actively undermines leadership in complex environments. The alternative is learning to experiment: creating conditions rather than chasing outcomes, tolerating failure as data, and treating humility as a practical skill rather than a virtue.

The shift from controlling outcomes to creating conditions for good outcomes is the core move in leading through complexity.

Complicated vs. complex: a critical distinction

• Complicated problems are predictable but require expertise — solved once, they stay solved.
• Complex challenges are emergent and entangled — context changes what works; prior experience can mislead.
• Humans prefer certainty so strongly that knowing a disease is permanent produces higher life satisfaction than uncertainty about recovery.
• Leaders default to treating complex challenges as complicated — this is the core error.

Why experimentation is hard even when you know it's right

• Intellectually accepting failure as data and actually tolerating it in the moment are different things.
• The instinct to feel like a failure when an experiment doesn't go perfectly is near-universal, even among people who teach experimentation.
• Attachment to a particular outcome makes it hard to notice early when something isn't working.
• A CEO whose $80M product launch failed publicly found that no one would even discuss it — until he reframed it as "an investment in collective learning" that had to be talked about to yield any return.

Creating conditions vs. chasing outcomes

• Complex goods — psychological safety, a happy child, a successful product launch — cannot be purchased or guaranteed; they arise.
• Treating them as targets to hit forces you to substitute a measurable proxy for the actual thing you want.
• Creating conditions asks: what makes this more likely to arise, who else has influence over those conditions, and how do we invite them in?
• This reframe opens up multiple possibilities rather than a single answer you're diligently trying to force.

The gift basket manager: what good conditions look like

• Same cold basement, same long hours, same products — two managers produced entirely different experiences.
• The first treated workers as artisans: brought sandwiches, hosted product experts, ran wacky contests, created joy.
• The second tracked basket count on a wall, removed everything else, and made the job miserable.
• Outcomes followed conditions, not instructions.

Releasing attachment to outcomes

• Strong commitment to a specific outcome makes failure invisible until it's catastrophic.
• The $80M product was perfect for the market two years before launch — by launch, everything had changed.
• People avoided discussing it until the CEO reframed failure as a learning investment.
• Once talkable, the team spent twelve months trying new things and succeeding, because they could name what had gone wrong before.

Humility as a practical leadership tool

• The belief that experiments should always go well is a form of arrogance — a subtle claim to god-like control.
• Leaders who do this well notice whether they default to blaming others or blaming themselves, then interrupt that habit.
• The move is: catch the blame reflex, recognise it as a way of discharging pain, and shift to a learning question instead.
• "Stumbling forward together" — knowing failure will happen and practising recovery and reconnection through it — is more useful than trying to avoid stumbling.

Putting an end date on experiments

• Humans are additive; open-ended commitments pile up and collapse under their own weight.
• A time-boxed experiment ("I'll do this for two weeks, then stop and assess") lowers the commitment barrier.
• Ending something is itself an experiment — many meetings, habits, and practices outlive their usefulness.
• Leaders scaling to larger roles often find the most valuable move is not adding a new skill but stopping something that made them successful earlier.

What actually blocks experimentation — and the evolution of the answer

• Early hypothesis: people don't know how to experiment (knowledge gap).
• Second hypothesis: people lack contexts that enable experimentation (structural gap).
• Third hypothesis: people lack bosses who encourage it (permission gap).
• Current view: identity gets in the way — who you are being as you experiment, not just what you know or what's allowed.
• The real answer is probably all three, not one silver bullet.