The misconception
A rate rises from 4% to 6%. One report calls that a two-percent increase. Another calls it a 50% increase. The first phrase is usually intended to mean two percentage points, while the second is a relative percent change. Without the missing word and denominator, readers can reasonably think the reports conflict.
Rates are already expressed as percentages, so a change between them has two common scales. Percentage points preserve the absolute distance on the 0%-to-100% rate scale. Relative percent change compares that distance with the starting rate. Neither scale is automatically more honest. Each answers a different question and can be misleading when presented without context.
A synthetic experiment
Imagine a fictional newsletter sent to 10,000 generated recipients in each of two periods. In the first period, 400 synthetic recipients click, producing a 4% rate. In the second, 600 click, producing a 6% rate. There are 200 additional clicks, the rate is two percentage points higher, and the number of clicks is 50% higher relative to the first period.
No real newsletter was sent, no audience was observed, and the values do not estimate human behavior or marketing performance. The scenario simply gives the rate a visible numerator and denominator.
In the linked percentage-change experiment, readers set an integer Baseline rate from 0% to 90%, an integer New rate from 0% to 100%, and a Reference count of 1,000, 10,000, or 100,000. Two zero-based bars show the before and after rates. Separate metrics report percentage-point change, relative percent change, and the corresponding count difference; the value table exposes both rates and counts.
Try a second setting: 40% to 42%. The percentage-point change is still two points, but the relative increase is only 5%. Then compare 1% to 2%. The absolute change is one point, while the relative increase is 100%. The baseline determines the relative framing. Finally set the baseline to zero: the point and count differences still exist, but the interface correctly reports conventional relative change as undefined.
Derivation and reasoning
Let p1 be the starting proportion and p2 the ending proportion, represented as decimals. Percentage-point change is:
percentagePointChange = (p2 - p1) * 100
For 0.04 and 0.06, the result is (0.06 - 0.04) * 100 = 2 percentage points.
Relative percent change is:
relativePercentChange = ((p2 - p1) / p1) * 100
For the same values, ((0.06 - 0.04) / 0.04) * 100 = 50%. The relative calculation asks how large the change is compared with where the quantity started.
Absolute count change is simply endingCount - startingCount, provided both counts cover comparable populations and periods. In the synthetic example, it is 200. Showing all three measures prevents the small starting rate from making a large relative number look like a large population impact.
Direction matters. Moving from 6% down to 4% is a decline of two percentage points, but the relative decline is about 33.3%, not 50%, because the baseline is now 6%. Relative percent changes are not symmetric when their direction is reversed.
Assumptions
The numerator and denominator definitions must stay consistent. A click rate among delivered messages cannot be compared directly with a click rate among all messages sent if delivery failures differ. The observation window and counting rules must also match.
The baseline must be greater than zero for the usual relative percent-change formula. Division by zero is undefined. Near-zero baselines produce very large relative changes, which may be mathematically correct but require absolute values for perspective.
Rates estimated from samples have uncertainty. A visible difference does not automatically exceed sampling variation, and neither change formula supplies a confidence interval or causal explanation.
Failure cases
Do not say “increased by 50%” when you mean “reached 50%.” An increase describes movement relative to a baseline; a level describes the final proportion. Likewise, “two percent” should not be used as shorthand for two percentage points.
Do not add a series of monthly percentage changes to obtain an annual relative change. Compounding uses multiplication: successive factors such as 1.02 and 0.98 multiply rather than cancel exactly. Official index series can also have specific base and seasonal-adjustment rules.
Do not calculate conventional relative change across a negative or sign-changing baseline without explaining the meaning. The arithmetic can produce a number that conflicts with ordinary ideas of growth. Absolute differences or a domain-specific alternative may be clearer.
Finally, do not choose whichever form sounds more dramatic. A responsible sentence can state all relevant scales: “The synthetic rate rose from 4% to 6%, up 2 percentage points or 50% relative to the 4% baseline, equal to 200 additional outcomes per 10,000.”
Source notes
The U.S. Bureau of Labor Statistics explains percent change as the difference divided by the earlier value and stresses consistent index bases and periods. The UK Office for National Statistics notes that relative movement in an indicator already expressed as a percentage can exaggerate how a change sounds and uses percentage points in that context. The newsletter counts and presentation are original synthetic material.
Keep these distinctions
- Percentage-point change subtracts two rates; relative percent change divides by the starting rate.
- Always name the baseline, denominator, direction, and time interval.
- Absolute counts reveal practical scale that either percentage can hide.
Sources checked
These sources support the statistical definitions and boundaries. The scenario, prose, calculations, and figure on this page are original to Count & Chance.
- Calculating Percent Changes, U.S. Bureau of Labor Statistics. Checked 2026-08-19.
- Sustainable Development Indicators: Three Percentage Point Rule, Office for National Statistics. Checked 2026-08-19.