The arithmetic mean is the sum of the observations divided by their count. Use it when every observation should have equal weight. Before calculating, decide what one observation represents: one item, one person, one day or one measurement.
Worked example: five processing times
Suppose five jobs took 12, 15, 18, 20 and 25 minutes. Their sum is 90 minutes and there are five jobs, so the mean is 90 ÷ 5 = 18 minutes per job. Enter 12, 15, 18, 20, 25 in the Average Calculator. Check that the count is 5 and the displayed average is 18.000000.
Zero is different from a missing observation
For 0, 10, 20, the mean is 10. If the first measurement was never recorded, the known values are 10 and 20 and their mean is 15. Do not replace missing readings with zero merely to fill a spreadsheet. Conversely, do not remove a genuine zero because it lowers the result.
Why averaging averages can be wrong
Team A completed two jobs averaging 10 minutes; Team B completed eight averaging 20 minutes. The combined average is (2 × 10 + 8 × 20) ÷ 10 = 18 minutes, not (10 + 20) ÷ 2 = 15. The simple calculator treats each entered number equally; calculate the weighted sum separately when groups have different sizes.
Check units, outliers and rounding
Convert all measurements to the same unit with the Unit Converter before combining them. A list containing both metres and centimetres has no useful mean until converted. For 10, 11, 12, 13 and 104, the mean is 30 while the median is 12. Neither is a typo: the high value pulls the mean upward. Report which statistic you used and investigate unusual readings rather than deleting them automatically.
Keep the full values during calculation and round the final answer to a precision appropriate for the original measurements. Six displayed decimal places do not make imprecise input more accurate.
Input checks in ToolVero
Separate plain numbers with commas, spaces or newlines. Use a decimal point; enter one thousand as 1000, not 1,000, because a comma separates observations. Remove currency symbols and unit labels. Blank input and invalid tokens produce a message instead of a partial average. The result shows a count so you can compare it with your source.