Find a percentage of a whole
To find 20% of 150, convert the percentage to a fraction and multiply: 20 ÷ 100 × 150 = 30. Keep the units of the whole with the result. If 150 represents centimeters, the answer is 30 centimeters; the percentage itself is dimensionless.
A percentage can be greater than 100 or negative when the context supports it. The calculator retains the sign instead of silently clipping values. Before choosing a formula, write the question in words: a portion of a whole, growth from a baseline and separation between two measurements are different tasks.
Use the old value for directional change
Percentage change = (new − old) ÷ old × 100. Going from 80 to 100 gives 20 ÷ 80 × 100 = 25% growth. Going from 100 back to 80 gives −20 ÷ 100 × 100 = −20%. The direction matters because the denominator changes.
This is why an increase of 25% followed by a decrease of 25% does not return to the starting value. Starting with 80, an increase to 100 followed by a 25% decrease ends at 75. When reporting a change, include the old value, the new value and the period or condition being compared.
Try it with your own input
Your result
Use symmetric difference for unordered values
When neither value is a natural baseline, a symmetric comparison can be useful. Difference = absolute(a − b) ÷ ((a + b) ÷ 2) × 100. For 80 and 100, that is 20 ÷ 90 × 100, approximately 22.22%. Swapping the inputs leaves the result unchanged.
The calculator restricts this mode to nonnegative values. For two zeroes it reports zero by an explicit convention. Other fields use different definitions of relative difference, so state the denominator in a report rather than presenting the percentage as if there were only one possible formula.
Subtract rates in percentage points
A rate moving from 20% to 25% changes by five percentage points. Relative to the old rate, it increases by 25% because five divided by twenty is one quarter. Both statements can be true, but they communicate different quantities.
Write “percentage points” when subtracting rates directly. A five-percent increase from a 20% rate would instead give 21%, because 20% × 1.05 = 21%. This distinction matters in dashboards, survey summaries and any comparison where the starting value is already a percentage.
Treat zero and negative baselines carefully
Ordinary percentage change is undefined when the old value is zero because the formula divides by zero. Moving from no events to ten events is an increase of ten events; it is not a finite percentage increase under that formula.
Negative baselines can produce signs that conflict with ordinary growth language. This calculator therefore requires a positive old value rather than presenting a mathematically computed number as an intuitive growth rate. Use an absolute change or a domain-specific measure when the ordinary relative comparison does not fit.
Make the result auditable
Round the displayed result only after doing the calculation. Keep enough input detail that another person can reproduce it. Distinguish a measured input from an estimate and make the unit visible. A precise percentage cannot rescue an uncertain measurement.
For image resizing, the aspect-ratio guide shows proportional scale. For power planning, the UPS guide and inverter guide use headroom as part of an explicit model. Check what the baseline represents before transferring one percentage rule into another task.
Common questions
Why are a 25% rise and a 25% fall not opposites?
They apply to different baselines. The second percentage is calculated from the increased value.
What is the difference between percent and percentage points?
Percentage points subtract rates directly. Relative percent change divides that subtraction by the old rate.
Can I calculate growth from zero?
The ordinary percentage-change formula divides by zero. Report the absolute increase or another suitable measure.
Why does symmetric difference use an average?
That denominator makes the comparison unchanged when the inputs are swapped. State the definition because conventions vary.
When should I round?
Keep intermediate precision and round the final reported value to a level appropriate for the inputs.