Why reversing a percentage means dividing, not subtracting
A percentage change is a multiplication, so undoing it is a division. That single sentence covers almost every reverse-percentage question you will meet, and almost every wrong answer comes from subtracting instead.
Applying a 20% discount multiplies the price by 0.80. Applying 20% sales tax multiplies by 1.20. Adding a 3% pay rise multiplies by 1.03. In each case the operation that produced the number you are holding is original × multiplier, so the operation that recovers the original is known value ÷ multiplier. Subtracting 20% from a discounted price does not get you back, because the 20% you would be subtracting is 20% of the wrong number.
The error is easy to quantify. Take a £100 jacket reduced by 20% to £80. Subtract 20% of £80 — that is £16 — and you get £64, which is £36 away from the true original. Dividing instead gives 80 ÷ 0.80 = £100 exactly. The correct add-back is 25% of the sale price, not 20%, because 20 ÷ 80 = 0.25. That asymmetry is the whole reason this calculator reports the change as a percentage of the value you entered as well as of the original.
The third mode handles the other family of reverse questions, the ones phrased as 40 is 25% of what number? Here nothing was applied to anything; you have a part and the share it represents, and you want the whole. The arithmetic is the same division: 40 ÷ 0.25 = 160. If you have the whole and want the part instead, that is a forward calculation and the percentage calculator handles it.
The formula and its one failure point
Write the forward operation first, then rearrange. If a value x0 was increased by p percent to give x1:
x₁ = x₀ × (1 + p/100) → x₀ = x₁ ÷ (1 + p/100)
For a decrease the multiplier is 1 − p/100. For a part-and-percent question the multiplier is simply p/100, because the part is that fraction of the whole.
The quantity 1 ± p/100 is the only thing you need to get right, and it has exactly one failure point: it must not be zero. A 100% decrease makes it zero, and dividing by zero is not a numerical inconvenience, it is a genuine loss of information. If everything was taken away, every possible starting value produces the same result of nothing, so the original is unrecoverable. This calculator returns a dash and says so rather than showing an infinity.
Two regimes surprise people the first time they meet them. A decrease of more than 100% — a 130% reduction, say — makes the multiplier negative, and dividing by a negative number returns an original whose sign is opposite to the value you entered. That is arithmetically consistent, and it is a signal that the percentage you were given is not the one that was actually applied. Similarly, a part quoted at more than 100% of the whole gives a whole smaller than the part; the numbers have usually been entered the wrong way round.
One more relationship is worth memorising because it converts between the two directions. If a decrease of p percent took you from the original to the final value, then getting back requires an increase of 100p / (100 − p) percent applied to the final value. At p = 20 that is 2000/80 = 25%. At p = 50 it is 5000/50 = 100%, which is why a half-price item needs a doubling to return to list. The reference table below tabulates this for the common percentages.
Worked example: a sale price and a tax-inclusive total
Part one — the sale price. A bicycle is advertised at £340 after a 15% reduction. What was the list price?
- Percent as a decimal. 15 ÷ 100 = 0.15.
- Multiplier that was applied. 1 − 0.15 = 0.85.
- Divide. 340 ÷ 0.85 = £400. Check it forward: 400 × 0.85 = 340. Correct.
- Amount saved. 340 − 400 = −£60, so the reduction is £60.
- The saving as a share of what you pay. 60 ÷ 340 = 0.17647, so the discount is 17.65% of the sale price even though it is 15% of the list price.
Part two — stripping tax out of a total. An invoice shows £1,200 including VAT at 20%. The net figure is not £1,200 − £240.
- Multiplier. 1 + 20/100 = 1.20.
- Divide. 1200 ÷ 1.20 = £1,000 net.
- Tax. 1200 − 1000 = £200, which is 20% of the net figure and 16.67% of the gross figure, because 200 ÷ 1200 = 0.16667.
The wrong method — taking 20% of £1,200 — gives £240 of tax and a net of £960, overstating the tax by £40 on a £1,200 invoice. On a VAT return that is a 20% overstatement of the reclaim, which is exactly the kind of arithmetic that gets a bookkeeper a letter.
Part three — part of a whole. A charity reports that 40 of its grants, representing 25% of the total, went to rural projects. 40 ÷ 0.25 = 160 grants in total, leaving 120 elsewhere.
Reading the four outputs
Original value is the number you came for. Always sanity-check it by running the calculation forward: multiply the original by the multiplier and confirm you land back on your known value. That takes five seconds and catches every mode-selection mistake.
Amount of the change is the known value minus the original, so it is negative after a decrease and positive after an increase. In part-of-whole mode it is the remainder — the whole minus the part — which is the figure you want when a report gives you one segment and its share and you need the rest.
Known value as a percent of the original is the multiplier expressed as a percentage. It is 80% after a 20% discount and 120% after 20% tax. This is the number to quote when someone asks “what proportion of the original is this?”, and it is what a spreadsheet stores when you write a growth factor rather than a growth rate.
Change as a percent of the known value is the one people get wrong. It is the size of the change measured against the number you are holding rather than against the original, and the two differ by the multiplier. After a 20% discount it is 25%; after a 20% markup it is 16.67%. If you are trying to advertise a discount, quote it against the original. If you are trying to work out how much a shopper actually saved relative to what they paid, quote it against the sale price. Both are true statements about the same transaction, and conflating them is how a “50% off” sign and a “save 100%” sign end up describing the same shoe.
Multipliers and the return trip for common percentages
| p | Increase multiplier (1 + p/100) | Decrease multiplier (1 − p/100) | After a p% increase, the rise is this % of the total | After a p% decrease, the cut is this % of what remains |
|---|---|---|---|---|
| 5% | 1.05 | 0.95 | 4.7619% | 5.2632% |
| 10% | 1.10 | 0.90 | 9.0909% | 11.1111% |
| 15% | 1.15 | 0.85 | 13.0435% | 17.6471% |
| 20% | 1.20 | 0.80 | 16.6667% | 25.0000% |
| 25% | 1.25 | 0.75 | 20.0000% | 33.3333% |
| 30% | 1.30 | 0.70 | 23.0769% | 42.8571% |
| 40% | 1.40 | 0.60 | 28.5714% | 66.6667% |
| 50% | 1.50 | 0.50 | 33.3333% | 100.0000% |
Column four is 100p/(100 + p); column five is 100p/(100 − p). The 50% row is the memorable one: half off needs a doubling to get back.
Mistakes that produce a wrong original
- Subtracting the percentage from the final value. Taking 20% off £80 gives £64, not the £100 you started from. Always divide.
- Choosing the wrong mode. “20% off” is a decrease; “including 20% tax” is an increase; “40 is 20% of” is part-of-whole. All three divide, but by 0.80, 1.20 and 0.20 respectively, which give £100, £66.67 and £400 from the same £80.
- Chaining two percentages as if they added. A 20% markdown followed by a further 10% is not 30% off. The combined multiplier is 0.80 × 0.90 = 0.72, so it is 28% off. Reverse it by dividing by 0.72, not by 0.70.
- Assuming tax was applied last. If a discount was taken after tax rather than before, the multipliers commute arithmetically but the tax base does not, and the tax figure on the invoice will not match. Reverse the operations in the order the document says they happened.
- Rounding the intermediate. Dividing by a multiplier you have already rounded to two decimals introduces error that grows with the size of the number. Keep full precision until the final display.
- Reversing a 100% decrease. There is no answer. If a value went to zero, the original left no trace in the result.
Where this fits among the percentage tools
Reverse percentage is one of four operations that all use the same three quantities — a base, a percentage and a result — and differ only in which one is unknown.
When you know the base and the percentage and want the result, that is a forward calculation: the percentage calculator. When you know two values and want the percentage that connects them, that is percent change if one is chronologically earlier, or percentage difference if neither is a designated baseline. When you know the base and the result and want the percentage applied, that is percentage increase and decrease. This calculator covers the fourth case: you know the percentage and the result, and the base is missing.
Several everyday problems are reverse percentage in disguise. Extracting a net figure from a tax-inclusive total is the one accountants meet daily, and tax authorities publish the divisor directly for this reason — HMRC's VAT guidance quotes the 20% fraction as 1/6 of the gross, which is the same statement as dividing by 1.2, since 1 − 1/1.2 = 1/6. Working a cost price back from a retail price and a known margin is another, though watch the definition: a margin is quoted on the selling price and a markup on the cost, so a 25% margin corresponds to a 33.33% markup and the two need different divisors.
For questions of the form a is to b as c is to what, where the relationship is a ratio rather than a percentage, the proportion solver is the direct tool. And where you are comparing a measurement against a reference rather than undoing a change, use percent error — it shares the arithmetic but answers a different question.