Standard deviation can be invaluable for understanding the distribution of a data collection. But why should you care?

Well, one reason you might care is if you want to understand the relationship between the data points. You might care if you think the collection of data is mostly the same or if there are wild variations in the data.

If the data are mostly the same, you may have a higher confidence using the average (mean) value. Conversely, if there are wild swings in the data, you might be more cautious and spend some time studying the data further before you make a decision.

This is where standard deviation shines. This simple statistic will help you understand whether the data are mostly the same or divergent. A small standard deviation means the data are bunched together around the average (mean) while a large standard deviation suggests there are outliers in the data. Ignoring the outliers could result in poor decision making.

For example, suppose we analyzed a collection of ten projects where the average cost was $25M and the standard deviation was $1M. This suggests the ten projects were all pretty close to the $25M average and that a future similar project cost could reasonably be forecasted at $25M. Your confidence level is relatively high given the narrow distribution and low standard deviation of 4% (1/25).

Now suppose we analyzed a collection of ten projects where the average cost was $25M and the standard deviation was $12M. This suggests some projects were significantly less or more than $25M. The data has a broad distribution; the data are further apart compared to the previous example. Forecasting a future project around $25M suggests caution and further analysis given the significant standard deviation value of 48% (12/25).

Consider adding standard deviation to your arsenal if you plan to lean on your historical data to make better decisions, analyze data distributions, and gain insights into risk, uncertainty, and confidence.