The 70% Confidence Principle

The trap this fixes

A launch decision sits in analysis for six months. Forty-seven models. Twenty-three meetings about methodology. The team is still not ready, because there is always one more test that would make the answer a little more certain. Meanwhile a competitor ships in weeks, learns from real customers, iterates twice, and owns the category before the careful team has finished debating its confidence interval. All that extra rigor did not make the team smarter. It made them slower, and in business, slow is the expensive mistake — it just doesn't show up on the same line item as a wrong answer, so it gets a pass.

Why 70% is the number, not 95%

Most business decisions do not require statistical certainty. They require a directional call made in time to act on it, with a way to notice quickly if it was wrong. Ninety-five percent confidence sounds more responsible than seventy, right up until you calculate what the extra twenty-five points actually cost in calendar time, and compare that to the cost of shipping something imperfect and course-correcting in the real world instead of a conference room.

This is not an argument for recklessness. It is an argument for defining 'good enough' data quality up front, on purpose, instead of discovering it by accident after the deadline has already blown past twice. Set the decision deadline first and work backward. Decide what confidence threshold actually matters for this specific call — a pricing change and a compliance filing do not deserve the same bar — and stop the moment you clear it.

Ship with a kill switch, not a leap of faith

The version of this that actually works in practice is not 'ship it and hope.' It's ship at seventy percent confidence with instrumentation that tells you within weeks, not quarters, whether you were right. Launch to a slice of the audience. Watch a small number of metrics that would tell you fast if it's failing. If it tanks, you've lost two weeks, not two quarters, and you've learned something a spreadsheet never would have told you.

The team that adopts this ships more, not less carefully — they just spread that care across several small, fast bets instead of concentrating it all into one slow, over-analyzed one. Speed and learning compound. Perfection, chased in isolation from the market, does not.

Where this shows up in the book

This is the exact arc of Chapter 8, where the team learns to stop chasing perfection and ship at seventy percent — and it gets a full leadership essay of its own, built around a project that should have taken three weeks and instead took six months.