Estimates are forecasts, not promises

Somewhere in every project’s paper trail there’s a number that got written down before anyone knew anything, and that number now runs the relationship.

It started life as an estimate. Everyone said the word, everyone nodded at the word, and then the number landed in a budget spreadsheet and quietly changed species. A forecast became a promise. Promises are things you break, and now the whole engagement is a countdown to the breaking.

We’ve watched this transformation enough times to know where it happens. Never in the room where the estimate is given, with all its caveats attached. It happens later, in the retelling, when the caveats get trimmed for brevity by people who weren’t in the room. The defense isn’t repeating the caveats louder. It’s changing how the number gets built in the first place.

Four myths do most of the damage.

Myth: an estimate is a promise with decimal places

A promise is a commitment about behavior. We’ll show up, we’ll tell you the truth, we won’t go quiet in week six. Those we make flat-out, and you should hold us to every one. An estimate is a different animal entirely: a forecast about the future, made at the moment of maximum ignorance. Day one is when we know the least we will ever know about your project. The estimate is the best map we can draw of territory nobody has walked yet.

Meteorologists solved this grammar problem ages ago. Nobody calls the forecast a lie when Saturday turns out sunny; it was a probability, stated honestly, and everybody planned with their eyes open. Project estimates deserve the same treatment and almost never get it.

Myth: when an estimate drifts, somebody failed

Estimates drift because building a thing is how you find out what the thing actually is. The integration that looked routine turns out to hinge on an API nobody has touched since flip phones were cool. An allegedly simple content migration uncovers eleven years of undocumented exceptions. That isn’t incompetence, it’s information arriving on its own schedule, and no amount of experience makes it arrive earlier, because experience just teaches you to stop being surprised that it arrives at all.

Bad estimating exists, and it looks different: no assumptions written down, no ranges, round numbers pulled from ambient optimism. But most drift isn’t a skill failure. It’s the ordinary cost of learning, billed at the moment the learning happens.

Myth: a single number sounds more professional

A single number sounds confident. The range is honest. And the single number was always secretly a range anyway; someone just deleted the honesty before hitting send. When we say eight to eleven weeks, we’re not hedging, we’re showing our work. Eight if the assumptions hold. Eleven if the usual number of them don’t.

The range also starts a conversation the single number can’t: what moves the needle. Which assumptions carry the most weight, which unknowns are expensive, what the client can do to help us land near the low end. A lone number invites exactly one follow-up question, and it’s “can you do it for less.”

Myth: clients want certainty

Clients want to not be surprised. Those are different desires, and confusing them is how relationships die. Certainty theater has a script: the crisp number, the confident nod, the long silence while the wheels come off. It produces the worst surprise there is, delivered late, when nothing can be done but apologize. A range plus honest updates produces no surprises at all. The number moves in plain sight, gradually, with reasons attached.

Here’s the part we didn’t fully believe until we’d watched it repeat for years: estimating this way makes trust go up, not down. The client who hears “here’s the range, here’s what would push us toward the high end, and here’s how fast we’ll tell you if it starts happening” doesn’t hear weakness. They hear people who’ve done this before. Because everyone who’s done this before knows the crisp single number is fiction, including the client, who has been quietly braced for the overage since the day they signed.

None of this is a license to shrug. A forecast obligates us harder than a promise does: name the assumptions in writing, narrow the range as we learn, say so out loud the moment the weather shifts instead of hoping it shifts back. Uncertainty handled honestly is a discipline. Hidden behind a confident number, it’s just a delayed apology.

It’s boring machinery in practice, and that’s the compliment: assumptions listed in the proposal and revisited at every milestone, a re-estimate delivered the week an assumption breaks rather than bundled into a surprise at the end.

Boring is what honesty looks like on a schedule.

The forecast says rain, and we said so up front. Bring a jacket. You’ll hear from us the second the sky changes.