Coefficient of Variation Calculator

What is a coefficient of variation calculator?

A coefficient of variation calculator measures how much a set of numbers varies relative to its own average. Enter your data points and the calculator reports the mean, the sample standard deviation, and the coefficient of variation (CV) — the standard deviation expressed as a percentage of the mean.

Unlike the standard deviation, which is measured in the same units as your data, the coefficient of variation is a pure, unit-free number. That makes it ideal for comparing the spread of two data sets that have different units or very different scales — for example, comparing the variability of monthly rainfall (in millimetres) with the variability of daily temperatures (in degrees), or comparing the volatility of a low-priced stock with that of a high-priced one.

A small CV means the values are tightly clustered around the mean; a large CV means they are widely scattered relative to the average.

How does it work?

The coefficient of variation is the ratio of the sample standard deviation $s$ to the mean $\mu$, multiplied by 100 to turn it into a percentage:

$$CV = \frac{\sigma}{\mu} \times 100\%$$

This calculator uses the sample standard deviation (with Bessel’s correction, dividing by $n - 1$), so at least two data points are required. The sample standard deviation is the square root of the average squared distance of each value from the mean:

$$s = \sqrt{\frac{1}{n-1}\sum_{i=1}^{n} (x_i - \bar{x})^2}$$

The calculation follows three steps:

1. Find the mean by adding all values and dividing by how many there are.
2. Find the sample standard deviation by summing the squared deviations from the mean, dividing by $n - 1$, and taking the square root.
3. Divide the standard deviation by the mean and multiply by 100 to express the result as a percentage.

The coefficient of variation is only meaningful for data measured on a ratio scale with a non-zero, positive mean. If the mean is zero the CV is undefined, and the measure becomes unreliable when the mean is close to zero or the data contains negative values.

Worked example

Consider the data set $1, 2, 3, 4, 5$, which has $n = 5$ values.

First, the mean:

$$\mu = \frac{1 + 2 + 3 + 4 + 5}{5} = \frac{15}{5} = 3$$

The squared deviations from the mean of $3$ are $4, 1, 0, 1, 4$, which sum to $10$. Dividing by $n - 1 = 4$ and taking the square root gives the sample standard deviation:

$$s = \sqrt{\frac{10}{4}} = \sqrt{2.5} \approx 1.5811$$

The coefficient of variation is then:

$$CV = \frac{1.5811}{3} \times 100\% \approx 52.7046\%$$

For the data set $2, 4, 4, 4, 5, 5, 7, 9$, the mean is $5$ and the sum of squared deviations is $32$. Dividing by $n - 1 = 7$ gives a sample standard deviation of $\sqrt{32/7} \approx 2.1381$, so:

$$CV = \frac{2.1381}{5} \times 100\% \approx 42.7618\%$$

The second data set has a higher mean but a lower relative spread, so its CV is smaller even though its raw standard deviation is larger.

Practical notes

The coefficient of variation shines whenever you need to compare variability across data sets that you could not compare with standard deviation alone — different units, different magnitudes, or different averages. In finance it is used to judge risk per unit of return; in laboratory science it quantifies the precision of a measurement method; in quality control it tracks the consistency of a process over time.

The CV is built directly on top of the average and the standard deviation, so it pairs naturally with both. For a broader summary of a data set’s centre, you may also want the mean, median and mode.

Frequently asked questions

What is a good coefficient of variation?

There is no universal threshold — it depends on the field. As a rough guide, a CV below 10% is often considered low variability, 10–30% moderate, and above 30% high. Always interpret the CV against the norms of your own domain.

Why use the coefficient of variation instead of the standard deviation?

Because the CV is unit-free, it lets you compare the relative spread of data sets that have different units or very different means. Standard deviation alone can be misleading: a standard deviation of 10 is large for data averaging 20 but tiny for data averaging 10,000.

When is the coefficient of variation not appropriate?

Avoid the CV when the mean is zero, negative, or close to zero, and when your data is on an interval scale (such as temperatures in Celsius) where the zero point is arbitrary. In those cases the ratio to the mean is unstable or meaningless.