Treat Your Next Big Move Like Three Small Ones

The expensive mistake isn't committing. It's committing at full size before the small version is proven. By the time the second fact arrives — usually dressed as good news — the first is a five-year lease and a board story to walk back.

A team sizing a new bet at full scale collapses four decision points into one, knowing the least it will ever know. Sign the lease rather than run a pop-up. Hire twelve rather than staff the pilot with two. Announce to the board rather than wait for the data — together they eliminate the room left to learn.

The Phase II team inherits commitments the Phase I team did not make — a number repeated by people, a reputation tied to the outcome. It will be defending, not deciding. This is where capital allocation discipline breaks down: when the bet is funded before it's tested.

An option is the right without the obligation — capped downside, full upside, the asymmetry commitment inverts. Most operators treat uncertainty as a thing to eliminate, when it is the thing to price.

Netflix put streaming in front of its DVD subscribers in 2007 as a free add-on and watched what they did. The cost of being wrong that year was a rounding error against being wrong three years later. That was somebody refusing to pay retail for a fact available on sale.

It's not commitment. It's optionality.

Phase I — the smallest version that can produce an honest no. Cheap enough that quitting embarrasses nobody. The trap is the demo disguised as a pilot.

Phase II — the bet on the model, not the idea. Phase I asks whether anyone wants the thing. Phase II asks whether it survives being ordinary. Most failures trace to two causes — unit economics carried by the discount, or revenue carried by one heroic person.

Phase III — the commitment. The lease, the hires, the announcement. By then the remaining uncertainty is execution, not existence.

Writing the exit condition before entering Phase I — not discipline, but because the version of you inside Phase II is disqualified from judging it. The commitments, not the information, disqualify him. The earlier you, writing while the decision is still cold, is the only disinterested party you will ever have.

A usable kill criterion has three parts: a number, a date, and a name. The name matters more than the number — a criterion nobody else has read is a diary entry.

Most people confuse the kill criteria with the forecast.

A 90-day pilot at $40K that prevents a 3-year lease at $1.2M is a 30x return on the learning spend, before any revenue. Same math in acquisition strategy — phase the diligence.

Businesses that destroy value on big moves skipped the experiment, got the news three years later, spent the interval defending a number. (Defenders always lose the press release, not the bet.)

Phasing breaks in two cases. A regulatory filing collapses the pilot into the commitment — sequencing, not phasing. Or Phase III collapses if you do not move first — patent window, one-time target. Most bets that feel time-sensitive are not.

The framework is not "commit slowly." It is "commit in the right order, at the right size, with the exit pre-written."