Meaning
Mathematical estimation of spread calculates the average squared deviation of individual sample observations from their collective arithmetic mean. In startup metrics, sample variance quantifies the dispersion in customer acquisition costs or contract values across small experimental test groups. Calculating dispersion with a corrected denominator prevents underestimating true population volatility when sample sizes remain limited.
The application boundary stops when total population census data is available, rendering sample estimation obsolete.
Dispersion Metric
Squaring differences between individual points and the calculated sample mean penalizes extreme outliers within small testing groups. When evaluating early revenue metrics, sample variance reveals whether an average contract size reflects consistent deal flows or a skewed distribution driven by isolated large accounts. Understanding this internal dispersion allows finance operators to evaluate revenue stability accurately.
High dispersion indicates that central tendency metrics fail to represent typical performance outcomes across the underlying customer base.
Estimation Error
Small sample sizes increase the likelihood that calculated dispersion strays significantly from true population parameters. As cohort volume expands, sample variance converges on population variance, reducing uncertainty in unit economic forecasts.
Cohort Volatility
Unpredictable performance variations within customer groups expose early business models to cash flow unpredictability. Tracking sample variance across consecutive weekly cohorts signals whether operational delivery is becoming predictable or remaining volatile.
