Percentage Calculator

Percentages: percentage points, compound changes, and the four mistakes people make

The four modes this calculator handles

Percentages look simple until you have to decide which question you are actually asking. This tool is split into four independent, live-updating blocks, each answering a different question. Getting the question right is most of the battle, because the same two numbers can mean four different things.

The fourth mode is the one people get wrong most often: the denominator is the sum of the ratios, not the count of items. Split 600 by 1:2:3 and you get 100 / 200 / 300 — the denominator there is 6, not 3. Only when the ratios happen to sum to the same number as the items (for example 1:1:1) does the calculation collapse into a plain average. That coincidence is exactly why "split evenly" and "split proportionally" get mentally merged, and why a proportional split is silently read as an even one.

Mistake 1 — confusing percentage points with percent

This is the single most common error in news and reports, and the two numbers can be worlds apart. A percentage point is the arithmetic difference between two percentages; a percent is the relative change of one percentage against another.

Both statements about the same event are true, but they feel completely different to a reader. Any time you see "grew by X%", you must first establish what the percentage is measured against. Interest rates, tax rates, approval ratings and conversion rates are all already expressed as percentages, so mixing these two readings is a frequent, sometimes deliberate, source of misleading claims. When a figure is a percentage to begin with, a two-point move is not a two-percent move.

Mistake 2 — adding percentages instead of multiplying

Consecutive changes must be multiplied, never added. Adding them assumes each change acts on the same base, which is false once the first change has already shifted the number.

• Three years of 10% growth is not 30% total but 1.13 − 1 = 33.1%.
• A 20% discount followed by a 10% discount is not 30% off but 0.8 × 0.9 = 0.72, which is 28% off, not 30% off.
• A price up 20% then down 20% does not return to the start: 1.2 × 0.8 = 0.96, i.e. 4% below the original. 100 becomes 120, then 96.

The third item is the most counter-intuitive, and the most useful to internalise: a rise and a fall of the same size never return to the origin, because the fall acts on the now-larger base. A 50% drop needs a 100% gain to break even; an 80% drop needs a 400% gain. This is the mathematics behind why investment losses are so hard to recover from — the base keeps shrinking, so the required percentage keeps growing.

Mistake 3 — averaging percentages without weights

The average of a set of percentages must be weighted by the base each one sits on; a plain arithmetic mean over the percentages themselves is wrong whenever the groups differ in size.

• Department A has 10 people and a 10% conversion rate (1 conversion).
• Department B has 1000 people and a 1% conversion rate (10 conversions).
• The arithmetic mean gives 5.5%, which is completely wrong.
• The correct figure is 11 / 1010 = 1.09%.

Whenever the bases differ, you cannot average the percentages directly. This bites when you roll up monthly growth rates, per-store completion rates, or per-platform conversion rates, and the error can be large enough to reverse a decision. The right denominator is always the total of the underlying counts, not the count of the percentages you are averaging.

Mistake 4 — rounding leaves leftover cents

When you split an amount proportionally, the results are usually not whole numbers, and rounding each part independently leaves the parts summing to a few cents off the total. Every accounting system that divides money has to deal with this — it is why "after splitting, the total does not add up" happens so often in practice.

The thing to avoid is rounding each line and then ignoring the total — that mismatch is the most common source of reconciliation errors at month end. If the parts must sum exactly to the whole, someone has to absorb the rounding remainder deliberately, not accidentally.

Where the basis matters in real workflows

The second row deserves a closer look: to go from a tax-inclusive price back to the exclusive price you divide by (1 + rate), not subtract the rate. 113 ÷ 1.13 = 100, whereas 113 × 0.87 = 98.31. The gap is not small, and the latter is the classic hand estimate gone wrong. The same inversion trips people up with markups, margins and any "price with X added" figure.

How the tool works

The four blocks are independent and compute live as you type — there is no button to press. Fill A and B to see the share, fill A and a percent to see the value, fill original and new values to see the rate of change, fill a total and comma-separated ratios to see the allocation. A positive change means growth, a negative one means decline. The change-rate block reports the sign so you can tell a drop from a rise at a glance.

All calculation happens locally in your browser; amounts and business data are not uploaded, which matters when the numbers are sensitive.