The short answer: percentage difference compares two values by dividing their absolute gap by their average: |A − B| ÷ ((A + B) ÷ 2) × 100. Comparing 80 and 100 gives 20 ÷ 90 ≈ 22.2 %. The result is symmetric, so it does not matter which value you enter first, and neither value is treated as a baseline.

Key takeaways

  • The average is the base. The formula divides the gap by the midpoint of the two values.
  • Order does not matter. Swapping A and B leaves the gap and the average unchanged.
  • The result is unsigned. It says how far apart the values are, not which one is larger.
  • Use it without a baseline. Two quotes, two instruments, or two people are typical cases.
  • Use percentage change for before and after. When one value is clearly the starting point, the original base is the right denominator.
  • Watch small averages. As the average approaches zero, the percentage can become meaningless or undefined.

When neither value is a baseline

Percentage change needs a starting point. When you compare the price of the same item at two stores, the height of two people, or two labs measuring the same sample, neither number came first. Treating one as the base would make the comparison depend on an arbitrary choice and would give different answers depending on the order of entry.

Percentage difference solves that by using the average of the two values as the reference. The comparison becomes a statement about how far apart the two numbers are relative to their shared scale. It is the right measure for a symmetric comparison and the wrong measure for tracking growth over time.

The formula and what the average does

percentage difference = |A − B| ÷ ((A + B) ÷ 2) × 100

The numerator is the absolute gap, so a negative sign never appears. The denominator is the arithmetic mean of the two values. Dividing by the average rather than by one endpoint is what makes the measure fair to both numbers: the average sits at the same distance from each value, so neither is privileged.

If the two values are equal, the gap is zero and the result is 0 %, provided the average is not zero. If both values are zero, the average is zero and the expression is undefined. If one value is zero and the other is not, the result is 200 %, which is an artefact of measuring against a midpoint rather than a meaningful level of disagreement.

Worked example: 80 and 100

Compare 80 and 100. The gap is |80 − 100| = 20. The average is (80 + 100) ÷ 2 = 90. Divide 20 by 90 to get 0.2222, then multiply by 100 for 22.2 %. Reversing the inputs gives |100 − 80| = 20 and the same average of 90, so the answer is identical.

Note how this differs from percentage change. Treating 80 as the starting point gives a 25 % increase, and treating 100 as the starting point gives a 20 % decrease. Percentage difference reports a single number, 22.2 %, that is true in both directions because it never chooses a direction.

Why the order does not matter

Both operations in the formula are symmetric. The absolute value discards the sign of the gap, and the average is unchanged when the two inputs are swapped. The measure therefore has a useful property: whoever computes it, and in whatever order, gets the same result. That is not true of percentage change, where swapping the values changes both the base and the sign.

The symmetry also means the measure describes distance, not direction. A result of 22.2 % does not say whether A or B is larger. If direction matters, say which value is higher in the surrounding sentence or report the signed gap as well.

Pair Gap Average Percentage difference
80 and 100 20 90 22.2 %
100 and 80 20 90 22.2 %
2,400 and 2,600 200 2,500 8.0 %
12.4 and 12.6 0.2 12.5 1.6 %

Difference, change, and error compared

The three measures look similar and answer different questions. Percentage change compares a new value with an original value and is signed. Percentage difference compares two values with their average and is unsigned. Percentage error compares a measurement with an accepted or true value, using the true value as the denominator.

Measure Base Signed Typical use
Percentage change Original value Yes Before and after
Percentage difference Average of the two No Two peer values
Percentage error Accepted value Usually yes Measurement accuracy

Choosing the wrong row is the most common conceptual error. A price that rose from 80 to 100 is a 25 % increase, not a 22.2 % difference; two quotes of 80 and 100 differ by 22.2 %, not by 25 %. The numbers are close enough that a reader may not notice the substitution, which is what makes the mistake worth naming.

When the average is small or zero

Because the average is the denominator, small averages produce large percentages. Comparing 1 and 100 gives a gap of 99 and an average of 50.5, so the difference is 196 %. The absolute gap is 99, but the percentage suggests a near-total disagreement, which may overstate the practical significance if both values are tiny.

When the two values are equal in magnitude but opposite in sign, the average is zero and the result is undefined. Comparing −100 and 100 gives a gap of 200 and an average of 0. Even when the average is merely close to zero, the percentage can swing wildly with a small change in either input. In those cases report the absolute gap, or choose a different reference value that reflects the scale of the problem.

Worked examples: quotes and measurements

Two suppliers quote 2,400 and 2,600 for the same job. The gap is 200 and the average is 2,500, so the percentage difference is 8 %. That is a fair summary because the quotes are peers. Saying the second quote is 8.3 % higher than the first, or that the first is 7.7 % lower than the second, are both percentage changes, and both are correct but asymmetric.

Two instruments measure the same sample as 12.4 cm and 12.6 cm. The gap is 0.2 cm, the average is 12.5 cm, and the difference is 1.6 %. With percentages as inputs, 40 % and 50 % differ by 10 percentage points, but their percentage difference is 10 ÷ 45 ≈ 22.2 %. The two statements use different scales and should not be used interchangeably.

Common mistakes

  • Dividing by one value instead of the average. That computes percentage change from an arbitrary baseline.
  • Reporting a negative difference. The absolute value makes the result non-negative for real inputs.
  • Mixing units. Convert both values to the same unit before subtracting.
  • Comparing a value with itself near zero. The formula is undefined when both values are zero.
  • Treating percentage points as percent. A 10-point gap between 40 % and 50 % is not a 10 % difference.
  • Using difference for before and after. When one value is the starting point, percentage change is the honest measure.

Practical caveats

Rounding deserves care. Round the percentage once, at the end, and keep full precision in the inputs; rounding 12.4 and 12.6 to whole numbers before subtracting would turn a 1.6 % difference into a 0 % or 100 % result. State how many digits you are reporting and do not add precision that the inputs cannot support.

Percentage difference is a descriptive statistic, not a test. A small difference can be statistically meaningful with precise instruments, and a large difference can be noise when the values are single measurements. Pair the percentage with the absolute gap and the measurement uncertainty. And remember that the measure is only defined for two comparable quantities: comparing a price with a weight produces a number but not a meaning. When the values are on different scales, normalize them first or use a different comparison.

Frequently asked questions

How is percentage difference different from percentage change?

Percentage difference divides the gap by the average of the two values and is symmetric. Percentage change divides by the original value and is directional.

Does the order of the values matter?

No. The absolute gap and the average are both unchanged when the inputs are swapped, so the result is the same.

Can percentage difference be negative?

Not in the standard formula, because the numerator is an absolute value. A negative result signals an arithmetic or sign error.

Why are 0 and 0 undefined?

The average of 0 and 0 is 0, so the formula divides by zero. Two equal values normally give 0 %, but the zero case has no scale.

When should I use percentage difference instead of percentage change?

Use difference when the two values are peers and neither is a starting point, such as two quotes or two measurements. Use change for before-and-after comparisons.

What does a 22.2 % difference between 80 and 100 mean?

The gap of 20 is 22.2 % of the average value of 90. It says the values are about a fifth of their shared midpoint apart, without saying which is larger.

Further reading

Keep exploring

Read next Logarithm: Base, Product Rule, and Scale A logarithm is an exponent in disguise. Learn the definition, the product, quotient and power rules, change of… Read next Order of Operations: PEMDAS and Ambiguous Expressions PEMDAS is a shared convention, not a law of mathematics. Learn the real precedence rules, the exponent and… Read next Significant Figures: Counting, Rounding, and Precision Significant figures record how precisely a value was measured. Learn the counting rules, the addition and multiplication rules,…

Decision path

Use this guide in a decision

Open Math Decision Center Open companion solver