Mental models for building a startup

The mental models

• First-principles thinking

  Instead of copying what competitors do, decompose the customer's problem to its irreducible truths and reason up to a solution. This is the source of genuinely new products rather than incremental clones of what already exists.

• Bottleneck

  At any moment a single constraint — a leaky funnel, a hiring gap, a slow onboarding step — caps the whole company's growth. A founder's leverage comes from finding that one narrow point and widening it, not optimizing everything at once.

• Gall’s law

  Ship the simplest version that actually works and let it evolve, because a complex product designed from scratch almost never works. This is the whole case for an MVP over a grand, feature-complete v1.

• Flywheel

  Design the growth loop where each part feeds the next — new users create content or referrals that attract more users — so early effort compounds. The first pushes are hard, but momentum eventually carries the business.

• Opportunity cost

  A founder's scarcest resource is time and focus, so every feature, market, or hire you pursue is one you can't. The true cost of any bet is the best thing you gave up to make it, not the cash it consumed.

• Inversion

  Run a premortem: instead of asking how the startup succeeds, ask what would guarantee it fails — no distribution, running out of cash, a product nobody needs — then work backwards to avoid each failure mode.

• Economic moat

  Growth without defensibility gets copied away once you've proven the market exists. Ask early what durable advantage — switching costs, network effects, brand, or cost structure — will keep better-funded rivals out.

• Network effects

  For the right product, each new user makes it more valuable to existing users, so growth feeds on itself. Where it holds, it can produce a winner-take-most position and a natural, widening moat.

• Optionality

  Structure the venture as small experiments with capped downside and uncapped upside — cheap bets you can walk away from. You profit from the rare experiment that works without betting the whole company on any single one.

• Expected value

  Under deep uncertainty, weigh each bet by its payoff times its probability rather than its most likely case. A low-odds path with an enormous payoff can be the right move — which is why a startup is a portfolio of asymmetric bets.

• The OODA loop

  Startups win by running the observe–orient–decide–act cycle faster than the market: shipping, watching real usage, and re-orienting quicker than larger, slower incumbents can respond.

• Incentives

  Team, investors, and customers do what they're rewarded for, not what you hope. Align equity, compensation, and pricing with the behaviour you actually want — and expect people to game any metric you reward.

• Margin of safety

  Keep a runway buffer so a slow month, a failed hire, or a wrong assumption doesn't kill you. Because building takes longer and costs more than planned, plan explicitly to survive the gap between expectation and reality.

Biases that trip up building a startup

• Sunk-cost fallacy

  The time, money, and identity you've poured into a failing product pull you to keep going when the honest move is to pivot or kill it. What's already spent should be irrelevant to whether the next month is worth it.

• Confirmation bias

  Founders in love with their idea seek validating feedback and hear polite encouragement as real demand, discounting the customers who won't actually pay — so they build something the market never wanted.

• Survivorship bias

  Copying the visible winners — 'do what this unicorn did' — draws lessons only from startups that survived, while the far larger graveyard that tried the same thing is invisible. The 'proven playbook' is often just luck.

• The planning fallacy

  Founders underestimate how long building, hiring, and shipping will take and how much it will cost, so roadmaps slip and runway runs out before the plan is even half done.

• Optimism bias

  The optimism that lets founders start at all also makes them overrate their odds and underprice the risks — treating a fragile plan as a safe one and skipping the buffers that would let them survive being wrong.

• The curse of knowledge

  Once you deeply understand your own product you can't imagine not understanding it, so onboarding, pitches, and messaging assume knowledge newcomers lack — and confused prospects bounce.

• False-consensus effect

  Assuming everyone shares the problem you personally feel leads to building for yourself and overestimating the market — the pain may be real for you and a handful of others, but not a business.

Mental models for other work