The geometric mean calculator above multiplies your values together and takes the nth root of the product, where n is how many values you entered. That sounds like a curiosity until you know when to reach for it, which is the point of this page. The geometric mean is the correct average whenever your numbers multiply rather than add — growth rates, investment returns, compounding percentages, index numbers, ratios of any kind.
Our mean, median and mode calculator already covers the ordinary arithmetic mean, and for most data that is the right choice. This page exists because for one specific and very common family of data the arithmetic mean is not merely different but systematically wrong, always in the same direction: it overstates. Arb Digital publishes both so you can pick deliberately rather than by habit.
What This Geometric Mean Calculator Does
Enter a series of positive values in any format — commas, spaces, line breaks — and the calculator returns the geometric mean along with the product of all the values, the count, the arithmetic mean for comparison, and the equivalent growth rate per period expressed as a percentage. The working panel prints the substituted nth-root expression using your actual numbers.
The mode selector saves you a conversion step. In plain mode, values are used exactly as typed, so growth is entered as a factor: 1.10 for a 10% increase, 0.90 for a 10% decrease. In percentage mode you type the rates themselves — 10, 25, −10 — and the calculator converts each to a factor before multiplying. Both routes give the same answer; the second just removes an easy place to make a mistake.
Values of zero or below are rejected rather than silently handled, and that is a mathematical necessity rather than a limitation of the tool. A single zero makes the entire product zero, which drags the geometric mean to zero regardless of every other value. A negative value makes the root of the product undefined for many values of n — the square root of a negative number is not a real number. The calculator says so explicitly and tells you which entries caused the problem.
How to Use It
- Decide what your values represent. If they are growth rates, use percentage mode. If they are already factors, ratios or index values, use plain mode.
- Enter the series in chronological or any other order — the geometric mean is unaffected by ordering.
- Check every value is above zero. Convert declines to factors below 1 rather than entering negative numbers.
- Read the equivalent growth per period if your data is growth. It is the constant rate that would produce the same total change.
- Compare against the arithmetic mean in the grid to see how much the wrong average would have overstated the result.
The Formula and How It's Calculated
The formula is GM = (x₁ × x₂ × … × xₙ)^(1/n). Multiply all n values together, then take the nth root of the product.
Take a fund that gains 10% in year one, gains 25% in year two, and loses 10% in year three. As factors those are 1.10, 1.25 and 0.90. The product is 1.10 × 1.25 × 0.90 = 1.2375, meaning the investment finished 23.75% up overall. There are three values, so the geometric mean is 1.2375^(1/3) = 1.0736. The equivalent constant growth rate is therefore 7.36% per year.
Check it by compounding: 1.0736 × 1.0736 × 1.0736 = 1.2375, which returns the true total. Now try the arithmetic mean of the same three factors: (1.10 + 1.25 + 0.90) ÷ 3 = 1.0833, or 8.33% per year. Compound that instead and you get 1.0833³ = 1.2714, implying the investment grew 27.14% when it actually grew 23.75%. The arithmetic mean has invented nearly three and a half percentage points of return that never existed. That gap is not a rounding artefact — it is the structural bias described by the inequality of arithmetic and geometric means, set out in MathWorld's entry on the arithmetic-geometric mean inequality.
When to Use the Geometric Mean Instead of the Arithmetic One
The test is simple. Ask what happens when you combine two periods: do the values add, or do they multiply? If a shop sells 40 units on Monday and 60 on Tuesday, the total is 100 — those add, so the arithmetic mean of 50 is correct and meaningful. If an investment grows 10% then 25%, the total is not 35%; it is 1.10 × 1.25 = 1.375, or 37.5%, because the second year's growth applies to the larger balance. Those multiply, so the geometric mean is the correct average.
The families of data that multiply are worth memorising: investment returns, revenue or traffic growth rates, inflation rates, population growth, compound interest, price indices, and any ratio-scaled measure such as engagement rate relative to a baseline. Anything where the phrase "percent per period" appears is almost always a case for the geometric mean.
There is a second, less obvious use, noted in MathWorld's definition of the geometric mean. The geometric mean is the right way to average quantities measured on completely different scales, because it is insensitive to the units chosen. Averaging a score of 200 on one scale with 0.5 on another produces an arithmetic mean dominated entirely by the larger scale, whereas the geometric mean treats a doubling on either scale as equally significant. That property is why composite indices built from unlike indicators are often constructed geometrically.
Geometric Mean and CAGR Are the Same Idea
Compound annual growth rate is the geometric mean wearing a finance label. CAGR takes an ending value divided by a beginning value, raises that ratio to the power of one over the number of periods, and subtracts 1. If you feed the yearly growth factors into a geometric mean, you get precisely the same number, because multiplying the factors reconstructs exactly that ending-to-beginning ratio.
The practical difference is what you have to hand. Use our CAGR calculator when you know only the start value, the end value and the number of periods. Use this page when you have the individual period returns and want the average of them, or when you want to see the working. Both answer the question "what constant rate would have produced this outcome" and both will disagree with the arithmetic mean whenever the periods vary — which is always.
Why Zero and Negative Values Break It
A geometric mean multiplies. A single zero anywhere in the series makes the product zero, and the nth root of zero is zero, so one zero value collapses the entire result regardless of how large everything else is. This is not a flaw to be patched: it correctly reflects that a compounding process which hits zero can never recover, since every subsequent multiplication leaves it at zero.
Negative values are worse, because they can make the root undefined. The geometric mean of −4 and 9 would require the square root of −36, which is not a real number. In practice this rarely bites, because the data types that call for a geometric mean are usually positive by nature. The mistake to avoid is entering a decline as a negative number: a 10% fall is the factor 0.90, not −10 and not −0.10. Percentage mode in this calculator makes that conversion for you, and our percentage change calculator converts between rates and factors if you need to prepare data first.
Reading the Gap Between the Two Means
The difference between the arithmetic and geometric means is itself informative: it grows with the variability of the data. Identical values give identical means. A series of 1.05, 1.05, 1.05 has both means at exactly 1.05. A series of 1.50, 1.10, 0.60 has far more scatter and a correspondingly wide gap between the two averages.
This is why volatile investments look so much better under arithmetic averaging than they perform in reality: the more the returns swing, the more the arithmetic mean overstates the outcome. A portfolio that gains 50% then loses 50% has an arithmetic mean return of 0% and an actual outcome of −25%, since 1.50 × 0.50 = 0.75. The geometric mean correctly reports about −13.4% per period. When you see an average return quoted without specifying which mean was used, the gap between the two is worth calculating before drawing conclusions. Our standard deviation calculator quantifies the underlying variability that drives that gap. This page describes how these averages are calculated and is not investment advice.
Arb Digital maintains a free library of maths, statistics and finance calculators. Browse the collection, or tell us what would be useful to you.
Browse All Free Tools Suggest a ToolCommon Mistakes to Avoid
- Averaging growth rates arithmetically. This overstates the result every time the rates differ, and the overstatement grows with volatility.
- Entering declines as negative numbers. A 10 percent fall is the factor 0.90. Negative values make the root undefined.
- Including a zero. One zero drives the whole product, and therefore the whole geometric mean, to zero.
- Using it on additive data. Units sold, headcount and revenue in dollars add rather than multiply, so the arithmetic mean is correct for them.
- Confusing it with the harmonic mean. Growth needs the geometric mean; rates such as speed over a fixed distance need the harmonic mean.
Related Free Tools From Arb Digital
For ordinary averages of additive data, use the mean, median and mode calculator. For rates and speeds, use the harmonic mean calculator. When you know only a start and end value, the CAGR calculator gives the same answer more directly, and the compound interest calculator projects forward from a rate. Use the weighted average calculator when some values deserve more influence, and the standard deviation calculator for spread. Browse the full free online tools hub for more.
Frequently Asked Questions
The geometric mean is the nth root of the product of n values. Multiply every value together, then take the root matching how many values there are. For 1.10, 1.25 and 0.90 the product is 1.2375, and the cube root of that is 1.0736.
Use it whenever your values multiply rather than add, which covers growth rates, investment returns, inflation, compound interest and index numbers. Use the arithmetic mean for quantities that add together, such as units sold or hours worked.
Because of the arithmetic-geometric mean inequality, which holds for any set of positive numbers. The two are equal only when every value is identical, and the gap widens as the values become more spread out.
No. The product of the values would be negative for an odd count, and taking an even root of a negative number does not give a real result. Enter declines as factors below 1, such as 0.90 for a 10 percent fall, rather than as negatives.
The product becomes zero, so the geometric mean is zero no matter what the other values are. That correctly reflects a compounding process that has hit zero, but it means a single zero makes the average uninformative about the rest of the series.
They give the same answer from different inputs. CAGR works from a start value, an end value and a period count, while the geometric mean works from the individual period growth factors. Multiplying those factors reconstructs the same overall ratio.
Subtract 1 and multiply by 100. A geometric mean of 1.0736 corresponds to an equivalent growth rate of 7.36 percent per period, which is the constant rate that would produce the same total change.
This page explains how different averages are calculated. It is general information about arithmetic, not investment advice, and past growth figures do not indicate future results.
