MROUND in Google Sheets: Round to a Nearest Multiple with Signed Controls

• Introduced: first launch date or introduction version not established by the official history reviewed.
• Availability: documented in the rolling Google Sheets service; this guide uses ordinary cells in the web editor, without a claim about identical mobile authoring workflows.
• Checked: October 14, 2026 against the official MROUND reference. Examples were independently calculated; no native Sheets execution was performed.

=MROUND(value, factor)

Both arguments are required and may be fractional. Google requires matching signs for nonzero value and factor, and explicitly documents 0 if either argument is zero. Its dedicated page states the tie rule in terms of absolute magnitude. Official MROUND reference

Factor is a step between permitted results, not a number of decimal places. Nearest rounding may select a lower result, so it is unsuitable by itself for a policy that must always cover a minimum required quantity.

Enter 25, 26, 27, 28, 29 in A2:A6. In B2 enter and fill down:

=MROUND(A2,4)

Expected results: 24, 28, 28, 28, 28. Input 25 is closer to 24 than 28. Input 26 is halfway between them, so the greater-magnitude result 28 is selected. Input 28 already lies on the grid.

For a requirement of 25 pieces, rounding to 24 would leave a shortfall. A correct nearest-rounding formula can therefore be the wrong capacity-planning rule.

=MROUND(26,4)
=MROUND(-26,-4)

Expected results: 28 and −28. The negative midpoint selects the larger magnitude, −28, rather than the numerically larger value −24.

Do not keep the factor positive when passing a negative value. Google’s page specifies matching signs but does not identify a complete error-code matrix, so this guide does not import an exact opposite-sign error classification from Excel.

For a numeric cell A2 and a four-unit rule, an explicit branch is:

=IF(A2<0,MROUND(A2,-4),MROUND(A2,4))

For −27, −25, 0, 25, 27, expected results are −28, −24, 0, 24, 28. The zero case uses Google’s documented zero behavior.

=MROUND(27,0)
=MROUND(0,4)
=MROUND(27,1)

Expected results: 0, 0, 27. A zero factor does not mean round to the nearest integer; a factor of one supplies the integer grid.

If zero is not meaningful for an editable step parameter, validate it explicitly. A numerical zero result may otherwise conceal an accidental missing or zero parameter.

=MROUND(10.625,0.25)
=MROUND(-10.625,-0.25)
=MROUND(10.6,0.25)

Expected results: 10.75, −10.75, 10.5. The first two inputs are exact midpoint controls for quarter-unit spacing. For the third, 10.5 is closer than 10.75.

Google’s published rule is the source for the tie direction here. These independently calculated examples are not a certification of every decimal floating-point boundary or a native test of all imported measurements. Keep actual boundary cases in the target workbook’s acceptance checks.

For three fictional quantities 10, 10, 10 in D2:D4, round each with =MROUND(D2,4) filled down. Each becomes 12, so their rounded total is 36.

Compare that with:

=MROUND(SUM(D2:D4),4)

Expected result: 32. The unrounded total is 30, halfway between 28 and 32, so the greater magnitude is selected. The two workflows give different answers because rounding is applied at different stages.

Choose the stage required by your process, and retain unrounded records if both detailed and aggregate reports must be reconciled. Do not alter a formula merely to force these two intentionally different calculations to agree.

For positive translated inputs, grid points beginning at 1 and repeating every 4 can use =1+MROUND(F2-1,4). Inputs 2, 3, 4, 7 produce 1, 5, 5, 9.

The example is deliberately restricted to F2 at least one so the translated value has a compatible sign with the positive factor. Inputs below that origin require a signed-factor branch after translation. Subtracting an origin can change the sign even when the original input was positive.

ROUND uses decimal places; ROUNDUP and ROUNDDOWN choose directions rather than the nearest arbitrary multiple. Check the quantity being specified before substituting one function for another.

Test exact multiples, both sides of a midpoint, negative values and zero. A display format does not change the source number, and a step that looks rounded on screen can still contain an unwanted fractional value. These checks are more informative than a single positive example.

Verification: Expected results use fictional data and independent calculations rather than native Google Sheets observations. See the editorial testing standards.