Coming next: The Second Factory →

Back to Library
Metrics & Evidence⚠️ Pitfall

The 95% Confidence Trap (Perfection is the Enemy of Progress)

The Problem

Your team has been "analyzing" a launch decision for 6 months. You've run 47 different models, debated methodology in 23 meetings, and you're still not ready because "we need more data to be sure." Meanwhile, your competitor shipped 3 months ago, learned from real customers, iterated twice, and now owns the market. You were chasing 95% confidence when 70% was good enough to learn. It's like a manufacturing plant that won't ship a product until it's 100% perfect—so nothing ever ships. Paralysis by analysis. The irony? All that extra analysis didn't make you smarter—it made you slower. And in business, slow is expensive.

The Principle

Ship at 70% confidence and learn fast. Set a decision deadline and work backward. Define "good enough" data quality upfront—not perfect, good enough. The world rewards speed + learning over perfection. Think manufacturing: you don't wait until you have perfect raw materials to start the line. You start production, monitor quality, and adjust. Same with decisions. Make the call, instrument the hell out of it, watch what happens, and iterate. Slow down just enough to not be reckless, then speed up. The magic happens when teams get comfortable with uncertainty and learn to course-correct in real-time instead of over-planning in conference rooms.

🎯

3 Steps to Take Action

  1. 1

    Set Decision Deadlines: For any analysis or launch decision, set a hard deadline. Work backward. Define "good enough" upfront - not perfect, good enough.

  2. 2

    Ship at 70% with Kill Switches: Launch to 10% of users with monitoring on 5 key metrics. If it tanks, kill it fast. If it works, scale it. Learn in weeks, not months.

  3. 3

    Track Decision Speed: Measure time from "we need to decide this" to "decision made." If it's >2 weeks for most decisions, you're overthinking. Speed up.

In Practice

A product team delayed a feature launch for 6 months seeking 95% confidence that it would drive engagement. Ran A/B test simulations, customer surveys, focus groups, competitive analysis. Meanwhile, competitor shipped a similar feature in 6 weeks. Customers loved it. By the time the overthinking team finally launched, the market had moved on. New VP came in, instituted a rule: "Ship at 70% confidence with kill switches." Next feature? Shipped in 3 weeks with monitoring to track 5 key metrics. Feature underperformed—but they knew in 2 weeks, not 6 months. Killed it, learned, shipped v2 in another 3 weeks. That one crushed it. Six months later they'd shipped 8 features vs the competitor's 2, learned what worked, and owned the category.

More on Metrics & Evidence

From the book

In the book this comes to a head in a board meeting, where growth theater meets its limit. The fix is honest metrics, not better slides.

This card is one of 57 drawn from The Auction Block by Brian Cariveau — the story of an analytics leader with 90 days to fix what the cards describe.

Each leadership pattern falls into a few pitfalls most often. Find your pattern.The Analytics Leadership Mirror: free, 5 minutes, no account