Percentage Calculator

Take 10% first: move the decimal point one place to the left, so 10% of 340 is 34. Everything else is built from that — 5% is half of it, 20% is double, 1% is a tenth of it. For 15% of 340, add 34 and 17 to reach 51. Percentages are also reversible: X% of Y always equals Y% of X, so 8% of 25 is easier read as 25% of 8, which is 2. Round awkward figures first — a rough answer you can sanity-check beats an exact one you cannot.

A percentage point is the plain arithmetic gap between two percentages; a percentage is that gap measured against where you started. A mortgage rate going from 3% to 4.5% has moved 1.5 percentage points and fifty per cent, and a headline may use either. The distinction bites hardest where the underlying figure is itself a rate — interest, unemployment, conversion, poll share, profit margin — because the two readings differ enough to change the story. Enter 3 and 4.5 in the change panel and you get the 50%; the 1.5 you get by subtracting.

Because the two percentages are taken of different numbers. Start at 100, add 20% and you are at 120; 20% of 120 is 24, so you land at 96. Reversing the order does not help — 100 less 20% is 80, then up 20% is 96 again. Multiplying makes it plain: 1.2 × 0.8 is 0.96. To undo a 20% increase you need a decrease of 16.67%, and to undo a 20% decrease you need a rise of 25%. Successive discounts compound the same way: 20% off then 10% off is 28% off, not 30%.

Divide rather than add back. If 60 is the price after 20% off, that 60 is 80% of the original, so divide by 0.8 to get 75 — adding 20% to 60 gives 72 and is the common wrong answer. VAT behaves the same way: a gross figure carrying 20% VAT divides by 1.2, so 120 gross is 100 net and 20 tax. The rule is to divide by one plus the rate for an increase, or one minus the rate for a decrease. No panel here does that division for you.

No, and it catches people out. Change has a direction: it needs a before and an after, and the earlier figure is the denominator. So 40 to 50 is a 25% increase while 50 to 40 is a 20% decrease, and one shop being 25% dearer is not the claim that the other is 25% cheaper. Where neither came first — two products, two regions, two teams — percentage difference is the usual alternative, dividing the gap by the average of the pair. State which figure is the baseline; a reader cannot infer it.

Nothing leaves the page. The six figures live as ordinary component state and the arithmetic happens on your own machine, so no request carries a number anywhere, there is no account, and nothing is written to storage. Refresh and the boxes are empty; there is no history to clear because none was kept. That matters when the numbers are a salary, a client's revenue or a medical result, since a percentage often gives away the figure you avoided naming. There is no copy button either: what leaves is what you select yourself.

An increase can be any size: a figure trebling is a 200% increase, because the change of 200 is measured against the starting 100. A decrease cannot pass 100% while the quantity stays positive, since 100% off is all of it — anything reported as a 120% fall has crossed into negative territory, possible for a profit, not for a headcount. Negative starting values stop percentages informing: the sum still returns a number, but a loss narrowing from -50 to -25 described as a 50% increase is read the wrong way round.

Divide the mark by the total and multiply by a hundred: 43 out of 60 is 71.666..., which becomes 71.67% or 72% depending on the precision your reader needs. Recurring results are the norm rather than the exception, so settle the rounding before you publish rather than after. Two habits cause trouble. Rounding each share and then adding them gives a set that does not total 100, which is why published tables so often carry a footnote. And rounding an already rounded figure drifts — go back to the original division each time.