← Patterns

Iteration Speed Is the Strategy

Pattern · design-philosophy

Core Concept

When facing deep uncertainty, the team that iterates fastest wins — not the team with the best first guess. Iteration speed is not a tactic; it is the primary strategic variable. Everything that accelerates the rebuild-retest-relearn cycle is strategic infrastructure; everything that slows it is strategic debt.

The Pattern

MacCready didn’t build a better airplane. He built a faster rebuild process (Mylar + aluminum tubing + wire → repair in hours). The airplane quality followed from the iteration speed, not the other way around.

This generalizes: in any domain with high uncertainty, the cost of a single iteration determines how many iterations you can afford, which determines how much you learn, which determines whether you win.

Iteration cost = Time to rebuild + Time to test + Time to learn
Total learning = (Available time / Iteration cost) * Learning per iteration

The only variables you control are iteration cost and learning per iteration. Most teams try to maximize learning-per-iteration (build the perfect experiment). MacCready minimized iteration cost instead.

Where It Applies

Software engineering:

• CI/CD pipelines, feature flags, vertical slices, trunk-based development — all reduce iteration cost
• Test suites are iteration infrastructure: they tell you what broke in seconds, not days
• Microservices (when warranted) reduce the blast radius of a change, which reduces iteration cost for each service
• AI-Ready-Engineering is the modern version: code health makes AI-assisted iteration faster

Product development (WCP):

• Ship → measure (PostHog) → learn (user research) → ship. The product process loop’s cadence IS the iteration speed.
• The non-consumption playbook’s “discovery over broadcast” is about learning speed: find the pre-qualified users fast, skip the education step, get to the signal.
• Per-release process (update job map → write before/after → notify users) is an iteration-speed investment — it closes the loop faster.

Startup strategy:

• “Find a faster way to fail, recover, and try again” (Raskin). This is why lean startup works when practiced honestly.
• Validate-Before-Building: a conversation is a faster iteration than a prototype. A prototype is a faster iteration than a product.
• Distribution-Over-Building: when you can build fast, distribution is the bottleneck — same principle, different constraint.

The Anti-Pattern: Magnum Opus Syndrome

“If the problem you are trying to solve involves creating a magnum opus, you are solving the wrong problem.” — Aza Raskin

Signs you’re in magnum opus territory:

• The project has been in planning/design for weeks without a test
• “We need to get this right” is used to justify delay
• Failure is expensive or embarrassing (so you avoid testing)
• The team is building on conjecture and theory without empirical grounding

The fix is always the same: what’s the cheapest version of this that teaches us something?

Related Patterns

Cross-References

• You-Are-Solving-The-Wrong-Problem — Source story (MacCready / Gossamer Condor)
• Solve-The-Meta-Problem — Parent pattern