Percentage Change: Formula, Bases and Comparisons

Definition and direction

Percentage change expresses a difference relative to a designated starting value. Its direction matters: changing from A to B is generally not the same percentage as changing from B to A. The conventional formula is 100 × (B−A)/A, where A is the original value and B is the new value. A must not be zero. Positive results indicate an increase when the base is positive; negative results indicate a decrease.

The absolute change B−A retains the original unit, such as dollars or units sold. The relative change is a dimensionless ratio expressed per hundred. Reporting both often communicates more than either alone. An increase of two items can be a large percentage for a small base and a tiny percentage for a large base. The percentage does not replace the underlying scale.

Variables and a worked calculation

Let A = 160 and B = 184. The absolute change is 24. Dividing by A gives 24/160 = 0.15, and multiplying by 100 gives a 15% increase. Substitution back into A(1+0.15) gives 184. This reverse calculation checks the denominator and percentage conversion in one step. The same process with B = 136 gives −15%, meaning a reduction of 24 from the original 160.

The percentage increase calculator applies a known rate to a starting amount. If both endpoint values are known and the rate is missing, use the ratio above first. The percentage decrease tool similarly applies a chosen reduction. Tool names alone do not establish the input direction; field labels tell you which quantity is being solved.

Percentage points compare rates directly

When the values being compared are already percentages, a direct subtraction produces percentage points. Moving from 8% to 10% is an increase of two percentage points. Its relative percentage change is 100 × (10−8)/8 = 25%. Neither result is inherently wrong, but they describe different quantities. Label them explicitly instead of writing an ambiguous “2% increase.”

For a rate based on observations, retain the counts if possible. Eight successes out of 100 and ten out of 100 produce the stated rates, but changing the number of opportunities also changes the denominator. A comparison of counts alone can disagree with a comparison of rates. See the percentage basics reference for combining groups and defining the whole consistently.

Repeated changes use factors

Represent each signed decimal change r as a multiplier 1+r. Two increases of 10% produce 1.1 × 1.1 = 1.21, or a 21% total increase. An increase of 20% followed by a decrease of 20% produces 1.2 × 0.8 = 0.96, or a 4% total decrease. Adding the rates misses the changing base at each step.

For several changes, multiply every factor, subtract one, and multiply by 100. This method assumes each change applies to the current value. If multiple percentages instead apply independently to one fixed original base, the problem has a different structure and must say so. The practical increase article shows how this distinction appears in budgets and repeated price changes.

Reverse a change rather than negating it

To recover an original positive amount after a known change r, divide the final amount by 1+r. If 120 is the result of a 20% increase, 120/1.2 = 100. Taking 20% off 120 gives 96, not 100, because that reduction acts on the new amount. Undoing an increase requires division by its multiplier, not multiplication by an opposite-signed rate.

The relative decrease from 120 back to 100 is 20/120, approximately 16.67%. More generally, reversing a positive increase r requires a decrease r/(1+r). A complete 100% decrease has no unique inverse because every original amount becomes zero. The algebra's undefined denominator reflects lost information, not a calculator limitation that can be solved by rounding.

Zero and negative bases

If A is zero, the standard relative-change ratio is undefined. Report “from 0 to 12” or “an increase of 12” instead of inventing a percentage. If both values are zero, the absolute change is zero but division by zero remains undefined. A display convention such as zero percent must be clearly identified as a convention, not the result of this equation.

Negative bases can make the sign counterintuitive. A loss improving from −100 to −40 yields −60% under the signed formula. Describing the loss as shrinking by 60 units is often clearer. Do not silently replace the denominator with its absolute value unless you explicitly define that alternative measure. Consistency matters more than finding a formula that produces a favorable-looking sign.

Different comparison questions need different measures

Percentage difference often uses an average of two values as the denominator when neither is naturally the starting point. That is not interchangeable with time-ordered percentage change. The percentage difference calculator addresses that separate comparison. ROI also uses a relative ratio, but its numerator is a defined gain and its denominator is an investment cost; see the ROI reference.

Before reporting a result, state the baseline date, unit, scope and rounding. Check that a price comparison includes equivalent quantities and that a sales comparison covers comparable periods. The worked examples here are direct algebraic derivations, not empirical claims. For a gentler introduction to choosing the denominator, read how to calculate a percentage.

Combine changes across unequal starting amounts

If one line item rises from 40 to 60 and another rises from 160 to 176, their individual increases are 50% and 10%. The combined original amount is 200 and the combined new amount is 236, so aggregate growth is 18%. The simple average of the two percentage rates is 30%, which gives the small line item as much influence as the large one.

An equivalent aggregate calculation weights the rates by their original amounts: (40×0.5 + 160×0.1)/200 = 0.18. Original-value weights are appropriate here because the question is growth in the combined amount. This does not mean every average of percentage rates should use the same weights; the desired statistic determines the denominator and weighting scheme.

Check that the line items share a compatible unit and do not overlap. Adding a subtotal to its own component would double-count part of the starting amount and distort the aggregate. Similarly, adding a monthly figure to a yearly figure without conversion creates a meaningless total. When a result differs from an average shown elsewhere, inspect the aggregation definition before treating it as an arithmetic error. Reporting original total, new total and the difference makes the calculation reproducible and reduces the chance that a reader mistakes an average item-level change for the overall change in the combined quantity.