Iterate or Exit

Definition

At the end of Customer Discovery, Customer Validation, Customer Creation, and Company Building, the model inserts an explicit "Iterate, Return, or Exit" phase: founders write down what their tested hypotheses actually showed and answer a fixed checklist (have we found a problem customers want solved, does our product solve it, is there a viable business model) before deciding whether to advance to the next step, loop back to an earlier one with revised assumptions, or abandon the current direction. It is a designed decision gate, not an informal gut check.

In the Book

Chapter 3 spells out the Customer Discovery version: if the market and customers don't check out, "take everything you learned in phases 1 through 3, modify your presentations based on feedback, go back to phase 1 and do it again." If only the product packaging is wrong, the team recycles to phase 3 (product presentation) instead of starting over completely — the checkpoint routes you to the specific phase that needs redoing, not blindly back to zero. The same structure recurs at the close of Customer Validation ("Iterate, Return, or Exit") and Customer Creation, and again in Company Building's "Iterate and Grow" phase, each time paired with a written checklist so the decision is documented rather than assumed. The book frames each of these steps as recursive loops in the diagram, not a straight line, explicitly because "the Customer Development model assumes it will take several iterations of each of the four steps."

Why It Matters

It turns "should we pivot" from a crisis-driven, emotionally loaded call made after the money runs low into a scheduled, structured, low-stakes checkpoint built into the process from day one. Because the gate specifies which earlier phase to return to rather than treating every setback as total failure, it makes course-correction cheap and routine instead of catastrophic — a discipline transferable to any multi-stage process where sunk cost and momentum otherwise push people past the point where they should stop and re-test an assumption.