"The sales forecast for next quarter is 1.2 million." A sentence like this sounds precise, confident, definitive. And that is exactly why it is dangerous. Behind that exact number hides an uncertainty it does not reveal: is it 1.2 million exactly, or somewhere between 1.0 and 1.4 million? Or between 0.8 and 1.6? The difference is enormous for whoever will decide based on that forecast — and yet, the number alone does not show it. Communicating the uncertainty of an estimate, through confidence intervals, is one of the most important and most rarely practiced analytical skills, and its absence leads, every day, to decisions made with a false sense of precision.
The problem has a psychological root. An exact number conveys competence and confidence; an interval seems hesitant, as if whoever presents it does not quite know the answer. So there is constant pressure to give round, definitive numbers, even when reality is far more uncertain. But this preference for false precision has a real cost: it leads people to over-trust estimates that are, by nature, approximate, and to decide as if they knew the future with a certainty they do not have.
This article is about why uncertainty should be communicated, not hidden, and about how confidence intervals make decisions more honest and, paradoxically, more robust.
Every estimate has uncertainty
It is fundamental to start with a truth that is often ignored: practically any estimate about the future, or any value computed from a sample, has uncertainty built in. A sales forecast, the result of a survey, the estimate of the effect of a change — all are approximations, not exact truths. Uncertainty is not a sign of poorly done work; it is an inevitable property of estimating something you cannot measure perfectly. Pretending it does not exist, presenting an exact number as if it were certain, does not eliminate it — it just hides it from whoever will decide.
A confidence interval is the honest way to communicate this uncertainty. Instead of saying "the estimate is 1.2 million", you say "the estimate is, most likely, between 1.0 and 1.4 million". The interval is not an admission of ignorance; it is a more complete and truer description of what is known. It communicates not only the most likely value, but also the margin within which reality will almost certainly fall.
Why the size of the uncertainty changes the decision
The reason communicating uncertainty matters so much is that its size changes, often entirely, the right decision. An estimate of 1.2 million with a narrow interval — say, between 1.15 and 1.25 — lets you plan with confidence. The same estimate of 1.2 million with a wide interval — between 0.8 and 1.6 — demands a completely different, more cautious decision, with plans in case sales fall well below. The central number is the same, but the sensible decision is opposite. Whoever only sees the central number has no way of knowing which of the two situations they face.
This is where false precision becomes dangerous. Presenting only "1.2 million", without the interval, makes the two scenarios — the low-uncertainty and the high-uncertainty one — look identical, when they demand radically different decisions. The hidden uncertainty does not disappear; it simply stops being taken into account in the decision, which is the recipe for bad surprises when reality turns out further from the number than assumed.
What an interval communicates well
- The most likely value: the center of the interval, the best point estimate.
- The margin of error: how far from the center reality can reasonably be — the information the number alone hides.
- The degree of confidence: the probability of reality falling within the interval, making the strength of the claim explicit.
- The need for more data: a wide interval is often a sign that more data is needed for a firmer estimate.
The pedagogy of uncertainty
Communicating confidence intervals has a challenge that is not technical, but human: many people are not used to reasoning with uncertainty and may misinterpret an interval, or be uncomfortable with the ambiguity. So communicating uncertainty well is as much a matter of pedagogy as of statistics. It is not enough to present the interval; you have to help the decision-maker understand what it means for their concrete decision — what actions to take if reality falls at the low end, and which if it falls at the high end.
This pedagogy pays off. Organizations where decision-makers are used to thinking in terms of intervals, and not exact numbers, make more robust decisions, because they plan for a range of outcomes instead of betting everything on an exact forecast that is almost certainly slightly wrong. Cultivating this literacy of uncertainty is one of the most valuable ways to raise a company's analytical maturity — and one of the least practiced.
A concrete case
A company was preparing for an important investment decision based on a demand forecast for a new product. The analysis team presented, as usual, a single number: sales of a certain volume were expected in the first year, and the investment plan was sized for that number. Everything looked solid and precise. But one of the decision-makers, with a sensitivity for uncertainty, asked the right question: "and what is the margin of error of this forecast?". The team, on computing the interval, revealed something the single number hid — the uncertainty was enormous. The most likely value was the one they had presented, but the reasonable interval ranged from a scenario in which the product sold less than half the forecast to one in which it sold much more. This uncertainty completely changed the decision. Instead of sizing the investment for the central number, as they were going to, they opted for a phased approach: a more prudent initial investment, sized to withstand even the low scenario, with the possibility of scaling quickly if sales confirmed the higher scenarios. Months later, sales did fall well below the central forecast — at the low end of the interval. The company, having sized the investment taking that possibility into account, got through the period without serious losses. Had it invested for the central number, presented without the interval, it would have faced serious difficulties. The difference between withstanding and suffering was in a single question about uncertainty — and in the honesty of communicating it instead of hiding it behind a number that seemed precise.
The honesty that strengthens trust
There is a fear, when communicating uncertainty, that it will undermine the presenter's credibility — that an interval will seem less competent than a firm number. Experience shows the opposite. In the short term, an exact number may impress more; in the long term, whoever communicates uncertainty honestly builds far more credibility, because their estimates are not contradicted by reality in the same way. Whoever promises "1.2 million" and sees 0.9 loses trust; whoever says "probably between 0.8 and 1.4" and sees 0.9 was right. Honesty about uncertainty is, in the end, what makes an analyst trustworthy.
Seen this way, communicating uncertainty is not a weakness to hide, but a strength to cultivate. It is what separates analysis that gives a false sense of certainty from analysis that informs a truly robust decision, prepared for the range of possible futures instead of bet on a single one that almost certainly will not come true exactly.
In practice
Next time you present or receive an estimate — a forecast, a test result, a projection — resist the comfort of the exact number and ask: what is the margin of error? A number without uncertainty is a half-truth that invites false precision. Communicating the interval, and helping the decision-maker understand what it means, makes decisions more honest and more robust. Do the estimates your company decides with come with their uncertainty in plain sight, or hidden behind numbers that seem more precise than they really are?
