Contact
12 programs
Explore the site

Zoho Sheet KURT Function: Understand Kurtosis and Outliers

Statistical Intermediate Zoho Sheet

Calculate and interpret kurtosis in Zoho Sheet, with a reproducible sales dataset, outlier comparison, native-validation cautions and troubleshooting.

Verification status: Official documentation referenced; example outcomes independently reasoned or arithmetically checked where stated. Not executed in Zoho Sheet. How formula examples are checked

Quick answer: Use =KURT(B2:B9) to request the kurtosis of numeric observations in a Zoho Sheet range. With the eight numbers 10, 12, 14, 14, 16, 18, 20 and 50, an independent calculation of the commonly used adjusted sample excess kurtosis is approximately 6.576687207. This indicates a distribution strongly influenced by the distant high value. This is a calculated expectation, not a result obtained by running the formula in Zoho Sheet.

Kurtosis is a statistic about distribution shape, especially the impact of unusually extreme observations. Zoho describes it in terms of being relatively peaked or flat compared with a normal distribution. That wording is a useful starting point, but kurtosis does not simply measure the height of a histogram’s tallest bar: tail behaviour and outliers are central to its interpretation. A high kurtosis does not automatically mean that the data are incorrect, nor does a low value prove the underlying process is safe.

Official Zoho Sheet syntax

Zoho’s current official KURT help page prints:

KURT(number; [number1]; ...)
Argument Required? Description Example
number Yes The first numeric input or a range of values to evaluate. B2:B9
[number1], additional arguments No Additional numerical inputs, references or ranges. D2:D5 or 35

A semicolon separates independent arguments in the official reference, while : identifies a continuous range. The examples in this guide follow that published notation. Check locale settings and formula language when transferring formulas from comma-delimited Excel or Google Sheets examples.

Zoho documents #DIV/0! when there are fewer than four numeric observations. Its KURT reference also provides generic #NAME!, #VALUE!, #REF! and #N/A! explanatory entries. Unlike the neighbouring Zoho SKEW reference, the KURT page does not expressly state its own behaviour for text, Boolean values and blank cells. Do not infer a fully verified KURT coercion policy from a different function’s remarks. For mixed data, clean the input to known numbers and test KURT in the native application before making definitive claims.

Availability: KURT is present in the current public Zoho Sheet statistical-function catalogue. The official function entry checked October 12, 2026 does not specify a first-supported version, feature tier, or full historical platform compatibility matrix. Zoho has announced a Windows/macOS desktop beta with many core web functions, but this announcement alone does not establish exact KURT behaviour on that beta, on mobile clients, or in every historical release.

Example 1: calculate kurtosis across eight weeks

Enter the following data.

Row A: Week B: Orders
1 Week Orders
2 Week 1 10
3 Week 2 12
4 Week 3 14
5 Week 4 14
6 Week 5 16
7 Week 6 18
8 Week 7 20
9 Week 8 50
  1. Check that B2:B9 contains eight genuine numeric cells, not strings containing commas or quantity units.
  2. Choose an empty output cell, such as D2.
  3. Enter =KURT(B2:B9).
  4. Compare the application result against the independent adjusted sample excess-kurtosis expectation of 6.576687207. That number is included for later validation, not reported as native Zoho output.
  5. Record the count using =COUNT(B2:B9) (eight numeric observations expected) and inspect the unusually large 50 before drawing conclusions about variability.

This is a useful practical pattern for checking promotional order spikes, delivery lead times or unusual inventory movements. It is not evidence that most future weeks will look like Week 8. For only eight observations, kurtosis can change dramatically when just one observation is added or removed; avoid overconfident decisions from a small dataset.

Where the numerical expectation comes from

A common adjusted excess kurtosis estimator uses the sample mean m, sample standard deviation s and number of observations n:

adjusted excess kurtosis =
 [n(n+1) / ((n-1)(n-2)(n-3))] * SUM(((x-m)/s)^4)
 - [3(n-1)^2 / ((n-2)(n-3))]
s = SQRT(SUM((x-m)^2) / (n-1))

For our data: n = 8; mean m = 19.25; SUM((x-m)^2) = 1151.5; s ≈ 12.825755338; and SUM((x-m)^4) = 905805.90625. The independent calculation yields 6.576687207 when rounded to nine decimals. The subtraction term makes this an excess kurtosis convention (approximately zero for a normal distribution under an idealised reference model), rather than an unadjusted fourth-moment measure commonly referenced to 3.

Important: Zoho’s KURT help page does not publish an explicit mathematical estimator equation. We therefore label the above formula an independently chosen comparison estimator and do not claim it is an exhaustive specification of Zoho’s engine. The vendor’s printed examples provide some corroboration, but the exact blank-argument case below remains unresolved without native testing.

Example 2: investigate the outlier rather than silently delete it

The following parallel views use independent arithmetic, applying the same adjusted sample excess-kurtosis estimator consistently:

Input set Proposed Zoho formula Independent expected excess kurtosis
All eight observations =KURT(B2:B9) 6.576687207
Seven routine weeks only =KURT(B2:B8) −0.638085328
Five symmetric illustrative values =KURT(1;2;3;4;5) −1.2

The comparison isolates the unusually high Week 8 value. Removing it is only a labelled sensitivity analysis, not a recommendation to erase actual transactions. If the 50 came from a real wholesale sale, retain it and explain its impact. If it came from a duplicated import, correct the underlying records with an audit trail before recalculating.

A seemingly symmetrical dataset can have a strongly negative excess-kurtosis result, and a right-skewed distribution can have high positive kurtosis. KURT and SKEW are different measurements. The sample is also too small to use either number as strong evidence of a particular probability distribution.

Official example ambiguity: an empty final argument

Zoho’s public KURT reference includes this literal example:

=KURT(2;4;7;1;)

It displays approximately 0.12987013. Notice the trailing semicolon, which leaves an apparent empty argument after the value 1. When a standard adjusted excess-kurtosis calculation is performed independently on the four explicit numbers [2, 4, 7, 1], the result is −0.285714286. Running the same independent estimator on five numbers [2, 4, 7, 1, 0] instead gives 0.129870130, matching Zoho’s displayed answer to the shown precision.

That agreement is an interesting documentation-level ambiguity, not proof of how Zoho actually parses empty positional arguments. The official page does not explain whether the trailing empty argument is treated as zero, whether the example relies on another setting, or whether the displayed result was calculated from different data. A practical safeguard is to avoid a trailing separator in example formulas, and test all three alternatives in the native application: =KURT(2;4;7;1), =KURT(2;4;7;1;), and =KURT(2;4;7;1;0). Do not resolve the ambiguity by speculation or present either expectation as a confirmed native result.

This is distinct from the usual case of an actual blank cell within a range, which would require its own native test. A blank argument and a blank cell need not be handled identically.

Advanced application: safe reporting with sample-size checks

For a shared dashboard, surface missing data before requesting a kurtosis from an undersized sample:

=IF(COUNT(B2:B9)<4;"Need at least 4 numeric observations";KURT(B2:B9))

Because kurtosis calculations also require nonzero dispersion, a proposed stronger guard is:

=IF(COUNT(B2:B9)<4;"Need 4+ numeric observations";IF(STDEV(B2:B9)=0;"No variation";KURT(B2:B9)))

These are untested composed formulas built from separately documented IF, COUNT, STDEV and KURT concepts. They should be executed and checked in Zoho Sheet before relying on native behavior. The vendor documents the below-four-numbers error; a constant series has zero standard deviation and is mathematically undefined for standardised kurtosis, but the precise native error returned for constant input has not been verified here.

A useful operational dashboard pairs KURT with COUNT, AVERAGE, MEDIAN, STDEV, SKEW and a chart. If kurtosis changes abruptly, check sample size and data import quality first. Compare similar periods with similar collection rules; avoid treating a large positive statistic as a forecast of the next week’s result.

Errors, edge cases and troubleshooting

Symptom Check and correction
#DIV/0! with three or fewer numbers Zoho explicitly says KURT needs at least four numeric observations. Use COUNT(range) to check what the sheet considers numeric.
Every numeric value is identical The standardised fourth moment is undefined when sample standard deviation is zero. Native display not tested; verify STDEV before relying on KURT.
Numbers imported as text Convert quantities to numeric values and check for stray characters. Zoho does not document KURT’s full mixed-cell coercion policy on this specific reference page.
A trailing separator changes the result Use no trailing separator; compare Zoho’s unusual printed example with explicitly adding zero and log the native result.
An unexpectedly high statistic Check for a legitimate single very large/small value, duplicate records, wrong unit conversions or a range extending into totals.
#NAME! Verify KURT spelling, defined names and any surrounding formula.
#REF! Look for broken or deleted cells/ranges, then reselect the correct data.
#VALUE! Investigate incompatible direct arguments, invalid nested expressions and coercion of nonnumeric inputs. Avoid pretending the generic help text covers every case.
Google Sheets or Excel example cannot be pasted as-is Check formula delimiter conventions and vendor-specific handling. The appearance of the same function name is not proof of identical implementation.

Zoho’s generic list also mentions #N/A! with lookup examples. That does not establish lookup-specific errors as a normal KURT outcome.

Cross-software behaviour and useful alternatives

Microsoft Excel: Microsoft’s KURT reference documents at least four values, warns about zero sample standard deviation, and describes its kurtosis estimator. Its published example uses comma-separated function arguments in the English reference. Excel’s description distinguishes values typed directly from text and logical values inside referenced arrays. Those details should not automatically be imported into Zoho’s more limited KURT documentation.

Google Sheets: Google’s official KURT page documents the same function name but uses comma-separated examples and describes text handling separately. Its note about the insufficient-values boundary is worded differently from Zoho’s under-four threshold. Treat such differences as matters for program-specific verification, not assumptions of cross-platform interchangeability.

Other Zoho functions: SKEW measures asymmetry; STDEV measures spread; VAR measures spread in squared units; DEVSQ returns a sum of squared deviations. If your real goal is to find unusual orders, first inspect the actual rows, consider a chart and report both the median and sample count. KURT does not identify which specific record caused the shape or why it occurred.

Related topics: SKEW, STDEV, VAR, DEVSQ, COUNT, AVERAGE and MEDIAN.

References and verification (checked 2026-10-12)

Verification status: Official vendor references checked on October 12, 2026; sample statistics independently calculated and cross-checked outside Zoho. Not executed in native Zoho Sheet, and the trailing-empty-argument interpretation remains unverified. No genuine Zoho Sheet screenshots are included; no mock application images are used. Native edge-case validation remains pending.