Home / Guides / ROUND, ROUNDUP and ROUNDDOWN in Excel

ROUND, ROUNDUP and ROUNDDOWN in Excel

Excel Functions · Updated June 2026

The ROUND function in Excel rounds a number to a set number of decimal places, changing the stored value, not just the display. ROUNDUP and ROUNDDOWN force the direction regardless of the digit, and MROUND rounds to the nearest multiple. The argument that confuses people is num_digits, which can be positive, zero, or negative to round to tens, hundreds, or thousands.

How the ROUND family works

=ROUND(number, num_digits) rounds number to num_digits decimal places using standard half up rounding. =ROUND(2.345, 2) returns 2.35, and =ROUND(2.344, 2) returns 2.34.

=ROUNDUP(number, num_digits) always rounds away from zero and =ROUNDDOWN(number, num_digits) always rounds toward zero, ignoring which way the digit would normally go. So =ROUNDUP(2.341, 2) is 2.35 and =ROUNDDOWN(2.349, 2) is 2.34.

The num_digits argument also accepts zero and negative values. Zero rounds to a whole number. A negative value rounds to the left of the decimal point, so =ROUND(12345, -2) returns 12300 and =ROUND(12345, -3) returns 12000. MROUND is different again: =MROUND(127, 25) returns 125, the nearest multiple of 25.

Worked example: rounding a price and a total

Suppose a unit cost of 2.346 sits in B1 and a large total of 12345 sits in B2.

  1. In C1 write =ROUND(B1, 2) to get a clean price of 2.35.
  2. In C2 write =ROUNDUP(B1, 2), which forces it up to 2.35 as well, while =ROUNDDOWN(B1, 2) would give 2.34.
  3. In C3 write =ROUND(B2, -2) to round the total to the nearest hundred, returning 12300.
  4. In C4 write =MROUND(B2, 50) to snap the total to the nearest multiple of 50, returning 12350.
  5. Use these rounded cells in downstream math so the model totals match what readers see.
Formulanum_digitsResult
=ROUND(2.346, 2)22.35
=ROUND(12345, 0)012345
=ROUND(12345, -2)-212300

Negative num_digits rounds to the left of the decimal: -2 means nearest hundred.

How analysts use them in models

Explicit rounding keeps a model's printed totals and its underlying math in agreement.

Pitfalls to watch

Display rounding and value rounding are not the same. Reducing the decimals shown by a cell format leaves the full precision underneath, so a column can display 2.35 plus 2.35 yet total 4.69 because the real values were 2.346 and 2.347. Wrapping the inputs in ROUND fixes the mismatch by rounding the actual values.

ROUNDUP and ROUNDDOWN ignore the usual half up rule entirely. =ROUNDUP(2.341, 2) is 2.35 even though a normal ROUND would give 2.34. Reach for them only when you truly want a forced direction.

Hardcoding the precision inside the formula, like =ROUND(B1, 2) scattered everywhere, buries a policy decision. If your rounding standard changes, you must edit every formula. Keep the digit count in a named input where practical.

Do it in one click

Find Hardcodes

Find Hardcodes flags a precision value typed straight into a ROUND formula so your rounding policy lives in one place instead of being scattered through the sheet.

Get ModelMint See how it works

FAQ

What does num_digits mean in the ROUND function?

num_digits sets how many decimal places to keep. A positive value rounds after the decimal, 0 rounds to a whole number, and a negative value rounds to the left, so -2 rounds to the nearest hundred.

What is the difference between ROUND and ROUNDUP?

ROUND uses standard half up rounding based on the next digit, while ROUNDUP always rounds away from zero regardless of that digit. =ROUND(2.341, 2) is 2.34 but =ROUNDUP(2.341, 2) is 2.35.

Why do my rounded numbers not add up?

You are likely rounding the display with a number format while the cells still hold full precision. Wrap the actual values in =ROUND(value, digits) so the stored numbers match what you see, and the totals will reconcile.