Why Your Percentage Is Wrong: Points, Per Cent and the Missing Denominator

The Missing Denominator

Almost every misleading percentage in circulation has the same defect. It states a change without stating what it is a change in, and the reader supplies a baseline of their own — usually the wrong one.

"Conversions up 32.5%" invites you to imagine a meaningful improvement. It is equally consistent with going from 4 sales to 5.3, which is noise, and with going from 240 to 318, which is a genuine result. The fix costs four words: state the base. "Up 32.5%, from 240 to 318" cannot be misread, and it lets a reader check the arithmetic rather than trust it.

Small bases are where this does the most damage. A campaign that improves from 2 conversions to 3 has improved by 50%, and that figure will look impressive in a slide deck for exactly as long as nobody asks for the counts. The percentage calculator shows the before and after alongside the change for this reason, and the conversion rate calculator reports the raw counts next to the rate.

The habit worth building is simple: whenever you write a percentage, write what it is a percentage of in the same breath. It resolves most of the problems in this article at once, because nearly all of them are the same mistake — a ratio detached from its denominator.

Points Are Not Per Cent

If a conversion rate moves from 2.0% to 2.5%, it has risen by 0.5 percentage points and by 25 per cent. Both statements are correct and they describe the same event. Writing "up 0.5%" describes something entirely different — a move from 2.0% to 2.01% — and it is wrong by a factor of fifty.

This is not pedantry. Interest rates, tax rates, margins, unemployment and conversion rates are all quantities that are themselves percentages, and any change in them can be expressed either way. A mortgage rate rising from 6% to 6.5% is half a point, or an 8.3% increase in the rate — and a very much larger increase in the monthly payment, which the mortgage calculator will show you is not proportional to either.

The convention that avoids all ambiguity is to use "percentage points" for absolute changes and "per cent" for relative ones, and to say which you mean rather than assuming context will carry it. Where a document mixes both, the reader has no way to tell them apart.

Margin and Markup Are Different Numbers

Buy for $100, sell for $150. Is that a 50% margin or a 50% markup?

It is a 50% markup and a 33.3% margin. Markup is measured against cost: $50 of profit on $100 of cost is 50%. Margin is measured against the selling price: $50 of profit on $150 of revenue is 33.3%. The same transaction, two denominators, two very different figures.

The gap widens as the numbers grow. A 100% markup is a 50% margin. A 300% markup is a 75% margin. A business that believes it is running at a 40% margin when it is actually applying a 40% markup — a 28.6% margin — has overstated its gross profit by nearly a third, and every downstream plan built on that figure inherits the error.

The markup calculator and the profit margin calculator deliberately show both conversions on the same page, because the useful thing is rarely one number in isolation. It is knowing which one your supplier means, which one your accountant means, and whether those are the same.

Percentages Do Not Average

Two campaigns, one converting at 2% on 10,000 visitors and one at 6% on 500. The average of 2 and 6 is 4, and the actual combined rate is 2.19% — because the first campaign has twenty times the traffic. Averaging percentages without weighting them by their bases is a reliable way to produce a number that is not true of anything.

The correct method is always the same: add the numerators, add the denominators, divide once. Two hundred plus thirty conversions is 230; ten thousand plus five hundred visitors is 10,500; 230 ÷ 10,500 is 2.19%.

The same trap sits inside grade calculations, where it is at least explicit. A course grade is a weighted average of component scores, and the grade calculator asks for the weights precisely because averaging the scores without them gives the wrong answer. Business reporting rarely makes the weights that visible, which is why the error survives there and not in a syllabus.

One final case worth knowing: percentage changes do not reverse. A price that falls 20% and then rises 20% does not return to where it started. It ends at 96% of the original, because the second percentage is applied to a smaller base. Over several steps this compounds, and it is the reason a sequence of small gains and losses can leave a portfolio down while the average change reads as zero.

Percentages Do Not Reverse

A price falls 20% and then rises 20%. Where does it end up?

Not where it started. A $100 item falling 20% is $80, and $80 rising 20% is $96. The second percentage is applied to a smaller base, so it moves a smaller amount in absolute terms. To undo a 20% fall you need a 25% rise — because the $20 you have to recover is now 25% of the $80 you are recovering it from.

The asymmetry grows sharply at the extremes and this is where it stops being a curiosity. Recovering from a 50% loss takes a 100% gain. Recovering from a 90% loss takes a 900% gain. That is why a portfolio, a business or a body of traffic that suffers one catastrophic drop is far harder to restore than an equivalent run of good years suggests.

It also explains a result that reads as a paradox. A series of returns averaging zero per cent leaves you down. Gain 50%, then lose 50%: $100 becomes $150 becomes $75. The arithmetic mean of +50 and −50 is zero, and the outcome is a 25% loss. The measure that describes what actually happened is the geometric mean, which here is −13.4% a year — and the gap between the two grows with volatility.

The practical version is that arithmetic averages of percentage changes systematically flatter. Anyone quoting an average annual return without saying which mean they used is quoting a number you cannot check, and the difference is not academic on a thirty-year horizon.

Three Questions Before Believing a Percentage

Everything above reduces to three questions, and asking them takes about five seconds.

Per cent of what? Identify the denominator explicitly. If the figure does not state it and you cannot infer it, you do not have enough information to interpret the number, and the honest response is to say so rather than to guess. This one question catches the missing-denominator problem, the margin-versus-markup confusion and most misleading marketing claims in a single move.

How big is the base? A 50% improvement on a base of two is one additional event. Small bases produce large percentages from noise, which is why any percentage change should travel with its counts. The CTR calculator and the email open rate calculator both display the underlying numbers next to the rate, because a rate on 40 sends is not a measurement.

Points or per cent? If the quantity being described is itself a percentage, establish which kind of change is meant. "Margin improved 5%" is genuinely ambiguous and the two readings can differ by an order of magnitude.

None of this requires statistical training. It requires treating a percentage as an incomplete sentence — a ratio missing its second half — and refusing to act on it until the other half arrives. Most of the time it arrives easily, and the times it does not are exactly the times you should be most careful.