Error Bars

Error bars represent the range of uncertainty around a measured value. They show not just the estimate but how much that estimate could plausibly vary — encoding confidence, variability, and sample size into a single visual element. In statistics, they might represent standard error, confidence intervals, or standard deviation. In business, the principle extends to any estimate presented as a single number when it should be a range: revenue projections, conversion rates, NPS scores, market size estimates. A founder who says "our CAC is $47" without error bars is communicating false precision. The model teaches that every measurement is an estimate, every estimate has uncertainty, and decisions made on point estimates without understanding their spread are decisions made partially blind.

Attach ranges to every important estimate your company produces — revenue forecasts, growth projections, experiment results. When evaluating data, always ask for the spread, not just the center. Teach your team that presenting a range isn't weakness; it's analytical honesty. Use overlapping error bars as a decision filter: if two options' ranges overlap substantially, the data hasn't distinguished them yet.

Error bars show the uncertainty around any estimate. They prevent false precision, reveal when data hasn't yet distinguished between options, and force honest communication about what you know and don't know.