PERCENTILE.EXC in Google Sheets: Exclusive Percentiles and Small-Sample Limits

• Introduced: first introduction date is not established by the reviewed official history; no desktop release number is inferred.
• Availability: documented in the rolling Google Sheets service; this guide uses the web worksheet workflow.
• Checked: October 14, 2026; official PERCENTILE.EXC reference. Documentation and independent calculations, with no native Sheets execution or wider client certification.

=PERCENTILE.EXC(data,percentile)

Both arguments are required. Google documents interpolation and shows #NUM! when a requested position cannot be interpolated. Its five-observation example returns errors at 0.15 and 0.9 despite both fractions lying inside the nominal domain.

The independent numeric model used here places the requested value at (n+1)*percentile in a sorted list. A position outside the available first-to-last observations is not interpolated by this model. Google’s worked example corroborates this convention; no native execution is claimed.

Use a fraction such as 0.25 or 25%, not the number 25. The page does not provide every invalid-type error display, so those cases are not assigned by analogy with Excel.

Enter 10,20,40,80,160 in A2:A6:

=PERCENTILE.EXC(A2:A6,0.25)
=PERCENTILE.EXC(A2:A6,0.75)

Expected results: 15 and 120. There are five values, so the positions are 1.5 and 4.5. The lower result is halfway between 10 and 20; the upper result is halfway between 80 and 160.

For =PERCENTILE.EXC(A2:A6,0.5), position 3 gives expected 40, the middle value. Neither 15 nor 120 needs to be present in the input because interpolation is part of the documented function.

=PERCENTILE.EXC(A2:A6,0.1)
=PERCENTILE.EXC(A2:A6,0.9)

The independent positions are 0.6 and 5.4, outside the first and fifth observations. The expected documentation-based outcome is #NUM! rather than an extrapolated cutoff.

This is a sample-size restriction, not a display-formatting problem. Changing a result cell to Percentage does not add observations or make an unavailable tail percentile valid.

Do not automatically replace the error with zero. Zero could be a real measurement and would hide the fact that the requested cutoff was not available under the chosen convention.

Enter the integers 1 through 19 in C2:C20. There are nineteen observations, making the model position 20*percentile.

The 10th and 90th percentiles now correspond to actual interior observations. This does not justify padding a sample with copied or invented records merely to obtain a requested percentile.

For the original five values, the exclusive upper-quartile cutoff is 120. Only one value, 160, is at or above it. Under the inclusive PERCENTILE model, the upper-quartile cutoff is 80, and two values qualify.

=COUNTIF(A2:A6,">="&PERCENTILE.EXC(A2:A6,0.75))
=COUNTIF(A2:A6,">="&PERCENTILE(A2:A6,0.75))

Expected counts: 1 and 2. Naming the convention matters when a threshold determines which rows appear in a report. Neither convention promises that exactly 25% of a five-row dataset will be selected.

Google’s published five-value fixture 15,20,35,40,50 gives 26 at fraction 0.4 and 48 at fraction 0.8. The model positions are 2.4 and 4.8: 20+0.4*(35-20)=26 and 40+0.8*(50-40)=48. This independently checks the rule against documented interior outputs.

The exact treatment of text, Booleans, empty ranges, ties near numerical boundaries and fractional precision is not fully specified by the reviewed reference. Those cases remain unverified. The examples use sorted numeric inputs, though the mathematical percentile calculation depends on values rather than their original row order.

• Google PERCENTILE.EXC reference.
• Editorial testing standards.

Examples use fictional data, English function names and comma argument separators. Expected results are independent calculations or explicitly documented rules, not native Sheets captures. Check spreadsheet locale when reproducing punctuation-sensitive formulas.