Rounding replaces a number with a simpler, nearby value by looking at one specific digit: the one immediately to the right of the place you’re rounding to. If that digit is 5 or higher, you round up. If it’s 4 or lower, you round down. That single rule covers the vast majority of rounding you’ll ever need to do, whether you’re estimating a grocery bill, filling out a tax form, or working in a spreadsheet.
The Basic Rule
Every rounding problem follows the same three steps. First, find the digit in the place value you’re rounding to. Second, look at the very next digit to its right. Third, decide: if that next digit is 0, 1, 2, 3, or 4, keep your target digit the same (round down). If it’s 5, 6, 7, 8, or 9, increase your target digit by one (round up). Ignore every other digit in the number. Only the one directly to the right of your target matters.
After rounding, drop all digits to the right of the place you rounded to. If you’re rounding a whole number, replace those dropped digits with zeros so the number keeps its size.
Rounding Whole Numbers
Suppose you need to round 4,372 to the nearest hundred. The hundreds digit is 3. Look one place to the right: the tens digit is 7. Since 7 is 5 or more, the 3 rounds up to 4, giving you 4,400. The tens and ones places become zeros because you no longer need that precision.
To round 4,372 to the nearest thousand, the thousands digit is 4 and the next digit to the right is 3. Since 3 is less than 5, the 4 stays the same, and you get 4,000.
A quick way to think about rounding to the nearest ten: any number ending in 1 through 4 rounds down, and any number ending in 5 through 9 rounds up. So 83 becomes 80, while 87 becomes 90.
Rounding Decimals
Decimal places are counted from the decimal point moving right. The first decimal place is the tenths, the second is the hundredths, the third is the thousandths, and so on. The process is identical to rounding whole numbers: find the target digit, check the next one, and decide.
Round 8.736 to two decimal places. Your target is the hundredths digit, which is 3. The next digit is 6. Because 6 is 5 or more, the 3 rounds up to 4, and the result is 8.74.
Round 6.8279 to one decimal place. Your target is the tenths digit, which is 8. The next digit is 2. Because 2 is less than 5, the 8 stays, and the result is 6.8. You simply drop everything after that tenths place.
What Happens When the Digit Is Exactly 5
In everyday math, a 5 rounds up. That’s the rule taught in most schools, and it works perfectly for homework, quick estimates, and general use. But in finance and statistics, always rounding 5 upward introduces a small but consistent bias when you’re adding up thousands of rounded numbers. Over time, the totals creep higher than they should.
To fix this, some industries use a method called banker’s rounding (also known as “round half to even”). The rule: when the digit to the right is exactly 5 with nothing after it, round your target digit to the nearest even number. If the target digit is already even, leave it alone. If it’s odd, bump it up by one. For example, 1.125 rounds to 1.12 because 2 is already even, while 1.115 rounds to 1.12 because 1 is odd and gets pushed up to 2. J.P. Morgan uses this method for dividend payments and trade settlements.
For digits above or below 5, banker’s rounding works exactly the same as standard rounding. The only difference is how it handles that exact midpoint.
Rounding on Tax Forms
The IRS allows you to round cents to the nearest whole dollar on tax forms. Drop amounts under 50 cents and increase amounts from 50 to 99 cents to the next dollar. So $2.30 becomes $2, and $2.50 becomes $3. If you choose to round, you need to do it consistently on every line of the form, not selectively on some entries. Employers using the IRS percentage method tables for income tax withholding follow the same approach.
Rounding in Spreadsheets
Excel and Google Sheets give you several built-in functions that handle rounding differently depending on what you need.
- ROUND(number, digits) follows the standard rule. ROUND(8.736, 2) returns 8.74. Use a positive number for decimal places and a negative number for whole-number places: ROUND(4372, -2) returns 4400.
- ROUNDUP(number, digits) always rounds away from zero, regardless of the next digit. Useful when you need a ceiling value, like calculating how many boxes you need to ship a certain quantity of items.
- ROUNDDOWN(number, digits) always rounds toward zero, effectively truncating the number at your chosen precision.
- MROUND(number, multiple) rounds to the nearest multiple you specify. MROUND(17, 5) returns 15, because 17 is closer to 15 than to 20. Both the number and the multiple must share the same sign, or the formula returns an error. One quirk to know: when you give MROUND a decimal multiple, its behavior at exact midpoints can be inconsistent, so double-check results when precision matters.
Choosing the Right Level of Precision
The place value you round to depends on how the number will be used. Money in everyday transactions rounds to two decimal places (cents). Prices at the gas pump use three decimal places (tenths of a cent). Scientific measurements might need four or more. Rough estimates for budgeting or mental math often round to the nearest ten, hundred, or thousand.
One practical guideline: don’t round intermediate steps in a calculation. If you’re multiplying several numbers together, keep the full precision until you reach the final answer, then round once. Rounding at each step compounds small errors and can push your final result noticeably off target, especially in multi-step financial calculations or unit conversions.