The number they want to hear

Everything this course has asked you to produce so far has been written. A one-pager, a list of questions, a design doc, a range, a pre-mortem. Written work is forgiving — you get to think, edit, and choose your words with nobody watching.

This lesson is the room. The plan is on the table, everyone has read the estimate, and the VP of Sales says the sentence that undoes more good planning than any other:

“Understood. But can you just try for the 15th?”

A forecast and a commitment are different objects

This is the distinction the whole lesson rests on, and most trouble here comes from the two words being used interchangeably by people who mean very different things.

• A forecast is what you believe will happen, carrying its uncertainty with it. It is a description of the world.
• A commitment is what you will be held to. It is a promise, and its uncertainty has been stripped off.

The pressure in the room is to convert the first into the second without anybody saying that is what happened. You give a forecast; a date gets written down; three weeks later it is referred to as what you promised. Nothing dishonest occurred. Everyone remembers the number and nobody remembers the qualifier.

You are not the first to notice this. Scrum changed its own vocabulary over it: the 2011 revision of the Scrum Guide replaced commit with forecast for the work selected into a sprint, precisely because teams were being held to estimates as though they were promises, and the fix had to happen at the level of the word (Scrum.org, “Commitment vs. Forecast”). An entire framework found this confusion expensive enough to rename a core concept. You are allowed to find it hard in a meeting.

Why “just try for the 15th” works on you

You might think that knowing the date is arbitrary protects you from it. It does not, and the research on this is unusually blunt.

Tversky and Kahneman described anchoring and adjustment: faced with an uncertain quantity, people start from whatever initial value is available and adjust away from it — and the adjustment is systematically insufficient. The anchor keeps some of its grip no matter what you do afterwards (Tversky & Kahneman, 1974).

The demonstration is the part to remember. Participants watched a wheel of fortune spin and stop on a number, then estimated the percentage of African countries in the UN. The wheel was rigged, but more importantly it was visibly random — everyone could see the number had nothing to do with the question. Their estimates moved anyway, in the direction of whatever the wheel had shown.

Which means: say your number first

The uncomfortable practical consequence of anchoring is that the first number in the room does disproportionate work, and it is going to be said by someone. If you wait to be asked, you have handed that job to the person with the strongest incentive to pick a small one.

So open with it. And open with the range, not its midpoint — because a midpoint offered on its own is heard as the number, and the uncertainty you carefully worked out gets quietly discarded on the way into someone’s notes.

Give the range, name the branch that leads to its top end, and say when the range narrows. You already wrote all three when you built the estimate. The room is where they earn their keep.

The move when they push

They will push anyway. The instinct is either to hold the line — no, it is six to nine weeks — which sounds inflexible and makes the conversation adversarial, or to soften into well, we can try, which is the silent conversion.

There is a third move, and it is the one worth practising until it is automatic: name the conversion out loud, and turn it into a scope decision.

“I can commit to the 15th if we ship Okta only and add the other providers in the following month. Or I can forecast the 29th for the full scope, and I will know by the 22nd whether that is holding. Both are real options — which one do you want?”

Look at what that does. You have not refused, so you are not the obstacle. You have made the trade explicit rather than absorbing it privately. And you have handed the decision to the person whose decision it actually is — scope belongs to them; the relationship between scope and time belongs to you, and it is not negotiable by enthusiasm.

When someone has already committed on your behalf

Sooner or later you will find that the date was promised to a customer before anyone asked you. It is not always malice — often somebody repeated a number they heard in a corridor and did not know it was a forecast.

Do not spend the meeting on whether they should have. It is the satisfying conversation and it changes nothing, and it makes the next person less likely to tell you early. Instead, in this order: establish what is now true externally, restate the forecast without adjusting it to be polite, state the gap in plain units, and put all of it in writing the same day — which is the no-surprises discipline applied to a surprise somebody else created.

The written record matters more here than anywhere else in the course. This is the exact situation that later becomes a disagreement about who said what, and the person without a timestamp loses it.

The most expensive sentence in planning

It is “yeah, we can probably do that”, said in a room, to end a moment of discomfort that would have passed in fifteen seconds.

It costs so much because of what it buys: nothing. The discomfort was going to end anyway. What you have purchased is a commitment nobody planned, that you now cannot revise without it looking like a slip rather than the forecast it always was. Silence is available and costs nothing. “Let me check and come back to you this afternoon” is available and costs nothing.

Now apply it: rehearse the room

What this does not cover

You committed, honestly, with the trade made explicit. The plan is now real and other people are relying on it — which is where keeping them informed as it moves comes in, and eventually what to do when the decision underneath it turns out to be the thing that was wrong.