How to Round Numbers in Excel: Functions and Rules

Rounding in Excel can change a calculated value, force it in a particular direction, move it to a convenient multiple, or simply make it look shorter on screen. Those are different jobs, and choosing the wrong method can change totals or leave hidden decimals in later calculations.

For ordinary decimal-place rounding, start with ROUND. Use ROUNDUP or ROUNDDOWN only when the rule calls for a fixed direction. Use formatting when you want a cleaner display without changing the value used by formulas.

Choose the rounding method by the required result

Use ROUND for the nearest value

The syntax has two required arguments:

=ROUND(number,num_digits)

number is the value or formula result to change. num_digits tells Excel where to round:

• A positive value rounds to that many decimal places.
• Zero rounds to the nearest integer.
• A negative value rounds to the left of the decimal point.

This Excel example shows how =ROUND(A2, num_digits) changes the rounding position. Positive values round decimals, 0 rounds to a whole number, and negative values round to tens, hundreds, or thousands.

Assume A2 contains 3487.6583. The same source number can be rounded at several positions:

The digit immediately after the requested position decides the direction. If that digit is 5 or greater, the retained digit moves away from zero; otherwise it stays in place. Consequently, =ROUND(2.5,0) returns 3, while =ROUND(-2.5,0) returns -3. Microsoft’s ROUND reference documents the same positive, zero, and negative num_digits rules.

The examples use commas between function arguments. Depending on regional settings, Excel may require semicolons instead. Formula AutoComplete shows the separator expected by the current installation.

Compare ROUND, ROUNDUP, and ROUNDDOWN

The three functions share the same two-argument structure, but their direction rules differ:

=ROUND(number,num_digits)
=ROUNDUP(number,num_digits)
=ROUNDDOWN(number,num_digits)

This example compares Excel’s ROUND, ROUNDUP, and ROUNDDOWN functions with positive and negative values. It shows that ROUND returns the nearest value, ROUNDUP rounds away from zero, and ROUNDDOWN rounds toward zero.

“Up” and “down” refer to magnitude for this function pair, not to movement on the number line. That distinction is easy to miss with negative values. The official references define ROUNDUP as away from zero and ROUNDDOWN as toward zero.

Use directional rounding only when a rule demands it. For example, a capacity model might require whole containers and therefore round a positive requirement away from zero. A general report that merely needs fewer decimals usually calls for ROUND or number formatting instead.

Round to a multiple instead of a decimal position

ROUND works in powers of ten. When the target increment is 5, 25, 0.5, or another specific multiple, use a multiple-based function.

This example shows how Excel rounds to specified multiples. MROUND returns the nearest multiple, CEILING.MATH rounds up to the next multiple, and FLOOR.MATH rounds down to the lower multiple.

MROUND requires the number and multiple to have the same sign; otherwise it returns #NUM!. Microsoft also documents a limitation around exact midpoint cases with decimal multiples, where binary-number representation can affect the direction. If midpoint behavior is contractually important, test the precise values and rule rather than assuming every decimal tie will behave identically. See the MROUND documentation for the limitation.

Negative values require extra care with CEILING.MATH and FLOOR.MATH because their optional mode argument can change whether negatives move toward or away from zero. Make the required direction explicit before using either function on mixed-sign data.

Remove decimals with TRUNC or INT

TRUNC removes digits without applying a nearest-value rule. Its optional second argument controls how many digits remain. With no second argument, it removes the fractional part:

=TRUNC(8.9)
=TRUNC(-8.9)

The results are 8 and -8. INT behaves differently for negative numbers because it returns the next lower integer:

=INT(8.9)
=INT(-8.9)

This Excel example shows the difference between TRUNC, INT, and ROUNDDOWN. With negative numbers, INT returns the next lower integer, while TRUNC and ROUNDDOWN remove decimals toward zero.

Those results are 8 and -9. For positive values, TRUNC(number), INT(number), and ROUNDDOWN(number,0) often look interchangeable. Negative inputs expose their different definitions. Microsoft’s TRUNC reference provides the same comparison with INT.

Know when formatting is enough

The Increase Decimal and Decrease Decimal controls on the Home tab change how many decimal places a cell displays. They do not normally change the stored value or the value used by another formula.

This example shows that number formatting only changes how a value appears on the worksheet. Excel still stores the full value for calculations unless you apply the ROUND function in the formula.

Suppose A2 contains 19.876 and is formatted to show two decimals. The cell displays 19.88, but this formula still uses the full stored value:

=A2*10

The result is 198.76. If the calculation must use the rounded amount, write the rule into the formula:

=ROUND(A2,2)*10

That result is 198.8. Formatting is right for visual consistency; a rounding function is right when later calculations must use the adjusted value. Microsoft’s guide to numbers that only appear rounded confirms that the full value remains visible in the formula bar and is used in calculations.

Avoid enabling Set precision as displayed as a routine fix. That workbook option replaces stored values with their displayed precision, and Microsoft’s rounding-precision guidance warns that the cumulative effect can make calculations less accurate. Prefer explicit rounding formulas at the points where a rule actually requires them.

Round percentages at the stored-value scale

Excel stores 12.345% as 0.12345. To round the displayed percentage to two percentage decimal places, round the stored number to four decimals:

=ROUND(A2,4)

This example shows why percentage rounding uses extra decimal places in the stored value. ROUND(A2,4) displays 12.35%, while ROUND(A2,2) displays 12.00%.

If A2 contains 12.345%, the formula returns 0.1235, which displays as 12.35% when formatted as a percentage. By contrast, =ROUND(A2,2) returns 0.12, displayed as 12.00%. The two extra decimal positions come from the factor of 100 used by percentage formatting.

Decide whether to round rows or the final total

Rounding at different stages can produce different totals. Consider three line items:

This example shows that rounding each line item can produce a different result from rounding only the final total. The sum of rounded rows is 91.55, while the rounded exact total is 91.54.

If quantities are in B2:B4 and unit amounts are in C2:C4, rounding each line in D2 and filling down gives a total of 91.55:

=ROUND(B2*C2,2)
=SUM(D2:D4)

Rounding only the exact grand total gives 91.54:

=ROUND(SUMPRODUCT(B2:B4,C2:C4),2)

Neither sequence is universally correct. Use the order required by the invoice, tax, accounting, measurement, or reporting rule, and document it. Do not mix row-level and total-level methods within the same model without a reason.

A reliable rounding workflow

1. Define the target. Decide whether you need decimal places, significant multiples, whole units, or display-only formatting.
2. Choose the direction. Specify nearest, away from zero, toward zero, toward a higher multiple, or toward a lower multiple.
3. Test positive and negative values. Directional functions can produce results that look surprising when signs change.
4. Place rounding deliberately. Decide whether it applies to each input, each row, an intermediate result, or only the final total.
5. Preserve source precision. Keep original values available when an audit or later recalculation may need them.
6. Check edge cases. Test values just below, exactly at, and just above a midpoint or boundary.

Common rounding mistakes

Frequently asked questions

Make the rounding rule visible

A dependable worksheet makes four choices explicit: the target precision or multiple, the direction, the handling of negative values, and the stage at which rounding occurs. Once those decisions are clear, the function usually becomes obvious.

Use ROUND for the nearest power-of-ten position, ROUNDUP or ROUNDDOWN for magnitude-based direction, and MROUND, CEILING.MATH, or FLOOR.MATH for custom increments. When appearance is the only concern, leave the underlying number intact and format the cell.