Paste a list of numbers and get the coefficient of variation instantly — with the mean, standard deviation and a plain-English verdict on how variable the data are.
The coefficient of variation (CV) is the standard deviation divided by the mean, shown as a percentage. Use this coefficient of variation calculator to find the CV for one dataset, or compare two datasets side by side to see which varies more relative to its mean — even when their units or scales differ.
Your data
Separate numbers with commas, spaces or line breaks. Non-numeric entries are ignored and counted below.
Clear Dataset B to calculate a single dataset only.
How the bands work
CV ≤ 5% is low, CV ≤ 15% is moderate, CV ≤ 30% is high and CV > 30% is very high relative variability. These are practical guide bands, not a statistical law.
COEFFICIENT OF VARIATION — DATASET A (SAMPLE)
Mean, standard deviation and CV appear here as you type.
Results
CV chart
The bars compare the sample CV of each dataset. A longer bar means more variability relative to the mean.
Calculation ledger — Dataset A
What is the coefficient of variation?
The coefficient of variation, written CV, is a relative measure of spread. It tells you how large the standard deviation is compared with the mean, expressed as a percentage. A CV of 10% means the typical distance from the mean is about one-tenth of the mean itself; a CV of 50% means the data are spread out by about half the size of the mean.
That ratio is what makes CV so useful. The standard deviation calculator on this site gives you the absolute spread in the original units — pounds, centimetres, test marks — but an SD of 15 means something very different for a dataset centred on 20 than for one centred on 2,000. Dividing by the mean removes the units, so CV lets you compare variability across datasets that live on completely different scales.
The vertical bars around the mean mean “absolute value”: a negative mean still gives a positive CV, because spread is a size, not a direction. The formula works with either a sample standard deviation (dividing by n − 1) or a population standard deviation (dividing by n). The tool above reports both, so you can see exactly how the choice changes the answer.
What is the difference between sample CV and population CV?
Sample CV uses the sample standard deviation, s, and is the right choice when your numbers are a sample drawn from a larger group — for example, 30 students from a school. Population CV uses the population standard deviation, σ, and applies when your numbers are the whole group. Sample CV is slightly larger than population CV for the same data, because dividing by n − 1 instead of n gives a slightly larger standard deviation. With large datasets the two are almost identical; with five values, as in the worked example below, the gap is visible: 22.59% against 20.20%.
How to use this coefficient of variation calculator
- Paste Dataset A. Enter your numbers separated by commas, spaces or line breaks. The example values 10, 12, 14, 16, 18 are already filled in.
- Add Dataset B (optional). Paste a second list to compare the two side by side. The tool then gives a verdict on which dataset is more variable relative to its mean.
- Choose sample or population SD for the headline answer. Both CVs are always shown in the results grid, whichever you choose.
- Read the answer. The headline shows the CV as a percentage, the band (low, moderate, high or very high), the mean and both standard deviations, a bar chart, and a step-by-step calculation ledger you can check by hand.
- Copy the link if you want to share the page or come back to it later.
Worked example: CV step by step
Take Dataset A: 10, 12, 14, 16, 18. First, the mean is (10 + 12 + 14 + 16 + 18) ÷ 5 = 70 ÷ 5 = 14. The squared deviations from 14 are 16, 4, 0, 4 and 16, which add up to 40. Dividing by n − 1 = 4 gives a sample variance of 10, so the sample SD is √10 = 3.16. The sample CV is therefore 3.16 ÷ 14 × 100% = 22.59%. Dividing by n = 5 instead gives a population variance of 8, a population SD of 2.83 and a population CV of 20.20%.
Now compare with Dataset B: 100, 110, 120, 130, 140. Its mean is 120 and its sample SD is 15.81 — five times the SD of Dataset A in absolute terms. Yet its sample CV is only 15.81 ÷ 120 × 100% = 13.18%. Dataset B is less variable relative to its mean than Dataset A, even though its absolute spread is far larger. That is the whole point of CV, and it is the comparison this calculator is built to make.
| Dataset | Mean | Sample SD | Sample CV | Population CV | Band |
|---|---|---|---|---|---|
| A: 10, 12, 14, 16, 18 | 14 | 3.16 | 22.59% | 20.20% | High |
| B: 100, 110, 120, 130, 140 | 120 | 15.81 | 13.18% | 11.79% | Moderate |
| 80, 90, 100, 110, 120 | 100 | 15.81 | 15.81% | 14.14% | High (sample) / Moderate (population) |
| −10, −12, −14 | −12 | 2.00 | 16.67% | 13.61% | High |
The third row shows how close the boundary can be: a sample CV of 15.81% sits just above the 15% line, so it is graded high, while its population CV of 14.14% is moderate. The fourth row shows why the absolute value matters: the mean is −12, the sample SD is 2, and the CV is 2 ÷ 12 × 100% = 16.67%, a positive number. If the mean were zero, as with −5, 0, 5, the CV would be undefined — the tool says so plainly instead of giving a number.
What is a good coefficient of variation?
A good coefficient of variation depends on the field, but as a practical guide, a CV under 15% usually signals fairly consistent data, and a CV under 5% signals very consistent data. A CV above 30% signals highly variable data that may need a closer look before you trust an average on its own. There is no universal pass mark: in laboratory measurements a CV of 5% might be poor, while in house prices a CV of 30% might be normal.
| CV band | Meaning | Typical reading |
|---|---|---|
| CV ≤ 5% | Low relative variability | Values cluster tightly around the mean; the mean is a good summary. |
| CV ≤ 15% | Moderate relative variability | Noticeable spread, but the mean is still a useful summary. |
| CV ≤ 30% | High relative variability | Spread is a large share of the mean; check for outliers or subgroups. |
| CV > 30% | Very high relative variability | Data vary widely relative to the mean; the mean alone may mislead. |
Use the bands to compare, not to judge in isolation. Two investment funds, two delivery routes or two exam groups can be ranked by CV even when their averages differ wildly — that is exactly the job CV was designed for. Always report the mean and the SD alongside the CV, so readers can see what the percentage is relative to.
When should you use CV instead of SD?
Use the coefficient of variation instead of the standard deviation when you need to compare variability between datasets with different means or different units. The SD is the right tool when the datasets share the same scale and you want the spread in real units. CV is the right tool when the scale itself is part of the problem.
- Comparing different units: the variability of heights in centimetres against weights in kilograms cannot be compared with SDs, because the units differ. CV is unit-free, so the comparison is fair.
- Comparing different scales: a machine that fills 100 ml bottles and one that fills 1 litre bottles may both be accurate, but only CV shows which is more consistent relative to its target size.
- Comparing investments: a fund returning 4% a year and one returning 12% a year have very different risk in pounds, but CV (often called relative volatility) shows which return is more variable relative to what it earns.
- Quality control: CV tracks consistency over time even as the average drifts, for example when a process mean moves from 50 to 500 units.
Do not use CV when the mean is close to zero, when the data contain negative values that cross zero, or when the scale is an interval scale such as temperature in Celsius, where zero is arbitrary rather than a true absence. In those cases the divisor is meaningless and the sample standard deviation is the honest measure. For data that follow a bell-shaped pattern, a normal distribution check and empirical rule probabilities are often more informative than CV alone.
Why CV is unit-free — and why it fails near a zero mean
CV is unit-free because the units cancel: the SD is in the same units as the data, and so is the mean, so their ratio has no units left. Multiply every value by 10 and the SD and the mean both grow by 10, leaving the CV unchanged. That is why a CV of 22.59% means the same thing for test marks, stock returns and delivery times.
The same division is CV’s weakness. When the mean is near zero, a tiny change in the mean makes the CV explode, and when the mean is exactly zero the CV is undefined — you cannot divide by zero. Values such as −5, 0 and 5 average to zero, so no CV exists for them at all; the calculator shows a clear message instead of a number. CV also assumes a ratio scale, where zero means “none”, so it is not meaningful for temperatures in Celsius or Fahrenheit, or for any data where zero is just a point on a scale. If your data are skewed, pair CV with a look at the shape: the skewness and kurtosis calculator shows whether a long tail is driving the spread.
Frequently asked questions
What is the coefficient of variation?
The coefficient of variation (CV) is the standard deviation divided by the mean, multiplied by 100 to give a percentage. It measures spread relative to the size of the mean, so it is unit-free and can be used to compare datasets with different units or scales.
What is a good coefficient of variation?
There is no single good value, but as a practical guide a CV of 5% or less is low, up to 15% is moderate, up to 30% is high and above 30% is very high. What counts as good depends on the field: laboratory work may demand a CV under 5%, while other data may be perfectly usable at 25%.
When should you use CV instead of SD?
Use CV instead of SD when you compare variability between datasets with different means or different units, because CV removes the scale. Use the SD when the datasets share the same scale and you want the spread in the original units.
How do you calculate the coefficient of variation?
First find the mean of your numbers, then find the standard deviation. Divide the standard deviation by the absolute value of the mean and multiply by 100. For 10, 12, 14, 16, 18 the mean is 14 and the sample SD is 3.16, so the sample CV is 3.16 ÷ 14 × 100% = 22.59%.
What is the difference between sample CV and population CV?
Sample CV is based on the sample standard deviation, which divides the squared deviations by n − 1, while population CV is based on the population standard deviation, which divides by n. Sample CV is therefore slightly larger. Use sample CV for a sample and population CV when you have the whole population.
Can the coefficient of variation be negative?
No. CV uses the absolute value of the mean, so it is always zero or positive, even when the mean itself is negative. For −10, −12, −14 the mean is −12 and the sample SD is 2, giving a CV of 2 ÷ 12 × 100% = 16.67%.
Why is the coefficient of variation undefined when the mean is zero?
Because CV is the standard deviation divided by the mean, and division by zero has no answer. For −5, 0, 5 the mean is exactly zero, so the CV is undefined. The mean being very close to zero is also a warning sign: the CV becomes unstable and can be huge even when the data barely change.
What does a coefficient of variation of 20% mean?
A CV of 20% means the standard deviation is one-fifth of the mean, so the data typically spread by about 20% of the average value. On the bands used here, 20% is high relative variability — the spread is a substantial share of the mean.
Reference
National Institute of Standards and Technology (NIST), “Coefficient of Variation”, NIST/SEMATECH e-Handbook of Statistical Methods companion reference: https://www.itl.nist.gov/div898/software/dataplot/refman2/auxillar/coefvari.htm — verified live on 12 November 2026. NIST defines CV as the ratio of the standard deviation to the sample mean, expressed as a percentage, and notes that it is used for ratio-scale data and is sensitive to a mean near zero.
More Tools to Explore
Carry on with the tools that pair naturally with a CV calculation, or browse all tools on the site.
Find the mean, variance and standard deviation behind any CV.
Check the shape and tails of your data before trusting the mean.
Probabilities and areas under the bell curve for your data.
Find 68-95-99.7 probabilities for a normal distribution.