Less Than 10%

By the time a venture shows up as a line item worth arguing about, the decision has already been made—not in any meeting you could point to, but made nonetheless, through small commitments accumulating before anyone was watching. And this is exactly what a structured diagnostic is designed to prevent. Issue a clear verdict at the 10% mark—GO, PIVOT, WRONG COMPANY, or NO-GO. Four possible directions. Each one specific, each one a roadmap for where the venture actually goes next. GO if it's worth scaling. PIVOT if the core insight is sound but the model needs reworking. WRONG COMPANY if the idea belongs in a different organizational home. NO-GO if nothing about the venture, in its current form, is worth building at all. The venture lead gets that clarity—that dramatic shift in direction—while the spend is still nearly invisible and they still have the budget and authority to act on it. Wait until Stage 2, and you've spent the money and the options have narrowed to almost none.

The arithmetic explains why timing matters. It costs roughly a dollar to generate an idea, ten to build the product, and a hundred to commercialize it, so everything that happens before the build accounts for under 10% of what the venture will eventually consume. That early window is the one place where walking away is still cheap, and it's also the window nobody examines closely, because there's almost nothing there to examine. A few weeks of someone's time and a deck don't trigger a review.

Small spend slips past governance while it quietly accumulates commitment. By Stage 2 the team is staffed, the roadmap has been shown to people, and the venture lead is defending a position instead of testing one. The evidence bar goes up right as the willingness to act on evidence goes down, and those two lines cross somewhere nobody was watching.

So the answer isn't more review. It's review placed early, at the 10% mark, while an honest answer is still cheap enough to act on and the venture lead still has budget left to act with.

Ask the expensive question while it's still cheap.