Barnaby Robson

Mental Model · Analysis & Framing · Innovation & Design · Strategy & Competition

The Idea Maze

Before building, map the space: the key forks, dead ends and dependencies—so you can choose a promising path and run smarter tests.

After Chris Dixon (2013; building on Balaji S. Srinivasan’s “Startup Engineering” lectures)

Chris Dixon’s Idea Maze frames startup/product creation as navigating a maze of choices — technology, business model, go-to-market, timing, regulation, and more. Great founders study prior attempts, imagine how the maze might shift, and plan sequence and wedge entry points before running experiments. Use it as a thinking tool to reduce naive detours, not as a crystal ball.

The Idea Maze" title above a multi-line graphic resembling a maze.

How it works

Map dimensions – tech approach, target segment, price model, distribution, data/standards, complements, regulation. Each creates forks

Sequence matters – decide build order (what must precede what), entry wedge, and triggers that open new corridors (platform shifts, cost curves). 

Study prior art – list who tried what, why they failed/succeeded, and what has changed (devices, bandwidth, APIs, costs). 

Choose corridors, not points – prepare 2–3 plausible routes with pre-defined switch rules if facts contradict assumptions.

Probe, don’t pontificate – run cheap tests to validate corridor assumptions; update the map as reality pushes back.

Use-cases

Zero-to-one product/venture ideation.

Entering regulated or multi-sided markets (payments, health, marketplaces).

Major pivots where distribution or complements change the viable path.

Technical strategy when platform or cost curves are in flux (AI, edge, new standards).

Step-by-step

  1. Define the mission & constraints – problem, target user, must-haves, guardrails (runway, compliance).

  2. List forks per dimension – e.g., self-serve vs sales-led, SDK vs SaaS, centralised vs on-device, regulated vs unregulated niches.

  3. Study the maze – 10+ historical attempts; note why now (what walls moved: tech, policy, behaviour). Add likely incumbent responses

  4. Draw 2–3 corridors – each with a wedge (the smallest use case you can dominate), build order, and distribution plan.

  5. Write falsifiable assumptions per corridor – critical uncertainties, success metrics, kill/switch thresholds.

  6. Run probes – smallest tests that hit the riskiest assumptions (design partnerships, fake doors, concierge pilots).

  7. Update the map – double down on corridors that clear thresholds; prune dead ends; log learnings for new entrants.

  8. Revisit on regime shifts – new platforms, cost drops, or rule changes can open hidden doors; remap quarterly.

Pitfalls & Cautions

Retrospective coherence – the maze can look obvious only after success; treat it as a planning aid, not prophecy. (Critiques: execution and effectual, improvisational learning often dominate.) 

Over-planning – months mapping, little learning; bias to short probes.

Ignoring distribution – great tech in a corridor with no repeatable channel.

False “why now” – hype without a real wall moving (cost/UX/regulation).

Single-path dogma – lock-in to one corridor; keep alternates with pre-set switch rules.