Starting a Company

Most of what determines a company's outcome is fixed before founding. What you can see about a market, who will take your call, what you can afford to be wrong about, and how many times you can be wrong before you must be right — these are the initial conditions, and companies are path dependent. Choosing a good starting position is far more tractable than executing your way out of a bad one. Founding advice mostly ignores this because initial conditions are unglamorous and largely inherited.

The inputs that actually vary

Strip out the noise and four things separate starting positions:

• Unfair information. Something you know about how a market operates that the market has not yet priced. Usually earned by working inside it, not by researching it.
• Warm distribution. The count of people who would answer your email and who can either buy or refer.
• Runway measured in attempts, not months. How many wrong guesses the business survives.
• An evaluation function. A way to tell a good answer from a bad one before a customer tells you.

"Solve your own problem" is popular for exactly one legitimate reason: it hands you the fourth item for free. You can judge the product without a customer meeting. That is the entire mechanism. If you have an evaluation function from somewhere else — a decade selling into the market, say — the rule does not apply to you and following it will send you somewhere worse.

Early decisions are cheap to make and expensive to reverse

The first customer sets the roadmap for two years, because their edge cases become your defaults. The first hire sets what the culture tolerates. The pricing model decides which customers you can afford to serve, and therefore who you are allowed to sell to later.

The most underweighted of these is revenue model. The same software sold as per-seat subscription, as usage-based, or as an engagement with people attached becomes three different companies: different hiring, different sales motions, different cash cycles, different ways of dying. Founders spend months on the product and an afternoon on the model that determines its economics.

Where the common understanding is wrong

Idea quality gets treated as the dominant variable. It is not. Ideas are graded by markets rather than juries, and nearly every founding idea is wrong in its details and roughly right in direction. The details get corrected by contact; the direction rarely does.

Timing gets more credit than it can carry. It is visible only in retrospect and cannot be acted on prospectively, except by standing in a market long enough to feel it move. "Right place, right time" is a description, not a strategy.

The variable nobody wants is capacity to stay wrong. Most companies that fail did not lose to a competitor. They ran out of attempts while still being wrong in an interesting direction.