Using a number in an argument without misleading anyone

Numbers end discussions they should not end

A figure introduced into a meeting does something to the conversation that a well-constructed argument does not. It reads as settled, because it appears to come from measurement rather than from opinion, and the people in the room usually have no way of checking it. Discussion tends to stop.

That effect is worth understanding whichever side of it you are on. It means a weak number beats a strong argument surprisingly often, and it means that when you present one you are being trusted more than you may have earned. The corrective is not to distrust all figures, which is its own kind of laziness, but to know which questions restore the discussion.

Ask what was counted and what was left out

Almost every misleading figure in ordinary working life is misleading because of its boundaries rather than its arithmetic. A count of something includes certain cases and excludes others, and those decisions were made by a person, often for a reason that made sense in a different context and was never revisited.

So the first question is always what exactly is in this number. Which period, which cases, which definition of the thing being counted. It sounds pedantic and it is the single most productive question available, because in a large proportion of cases nobody in the room can answer it, which is itself the answer to whether the figure should be deciding anything.

A comparison needs a baseline you can see

Change is where the greatest amount of accidental misdirection happens. A figure that has risen means very little without knowing what it normally does, how much it moves week to week, and whether anything else changed at the same time. A rise that sits comfortably inside ordinary variation is not a finding, though it is very frequently presented as one.

The other common trap is a comparison between two things that were not measured the same way. Two periods with different definitions, two teams with different recording practices, a before and after that also spans a change in how the data was collected. These are not deceptions; they are the normal state of information inside an organisation, which is why they survive so easily.

Say where it came from and what it cannot tell you

When you present a figure, name the source and the limits in the same breath. It feels like undermining your own case and it does the opposite: it establishes that you understand the number rather than merely holding it, and it prevents the far worse moment when somebody else identifies a limitation you had not mentioned.

Stating what a figure does not show is particularly valuable. This tells us how many, not why. This covers one team and may not generalise. That is the difference between analysis and advocacy, and senior people can tell them apart quickly, which is why the habit is worth forming early even when the room would have accepted less.

Precision is not accuracy

A number carried to two decimal places looks more authoritative than a rounded one, and often it is less trustworthy, because the precision comes from arithmetic performed on an estimate. If the inputs were approximate, the output is approximate regardless of how many digits the tool produced.

Round to what the underlying information can support, and say plainly when something is an estimate. The people who lose credibility over numbers are rarely the ones who said roughly; they are the ones who gave a precise figure that turned out to have been a guess with a calculation performed on it.

The limits of doing this well

It is worth conceding that careful use of evidence does not always win. Some decisions are made for reasons that have nothing to do with the figures, and a well-qualified analysis can lose to a confident assertion, particularly where the person asserting outranks the person analysing. That is genuinely how a lot of organisations work.

What the habit buys is narrower and still worth having. Over a couple of years, being the person whose numbers hold up under questioning changes how much your input is sought and how early. It also protects you from the specific and common career damage of having presented something that turned out to be wrong in a room where it mattered.