Decisions.

The framework

1. 01

   Write your number before you hear theirs.

   The only reliable defence against an anchor is a figure you generated before the other one existed. Not a range you sketch once the meeting invitation lands, not a position you refine after the proposal arrives — an independent estimate, written down, with the reasoning beside it, while you are still ignorant of what anyone else will say. Do it roughly if you must; a crude number built from your own evidence is a better origin than a precise one built from theirs. And write it somewhere you cannot quietly revise it later, because you will want to, and you will call it updating.

2. 02

   When their number lands first, rebuild — never adjust.

   The instinct is to counter, and countering keeps the argument on their number. Splitting the difference, shaving twenty per cent, arguing the figure down: every one of these accepts their number as the origin and haggles over the distance, which is precisely the mechanism the effect runs on. So refuse the axis rather than the amount. Say plainly that the number is not a basis you can work from, then rebuild the figure in front of them from your own inputs — what the work costs, what comparable deals cleared, what you would have proposed had nobody spoken. That is not obstruction. It is restoring your own origin.

3. 03

   Move first when you know the range; move second when you don't.

   Adam Galinsky and Thomas Mussweiler showed in 2001 that the party making the first offer usually finishes better off, because that offer defines the field everyone then works inside. The advice this yields is conditional, not universal: open first when you have a defensible read on the range, and let them open when you genuinely do not, because in ignorance your own first number is as likely to anchor you low as high. When you do open, pitch at the ambitious end of what you can justify and attach the reasoning in the same breath — a figure with a rationale is far harder to dismiss as posturing.

4. 04

   Audit your incumbent numbers — the anchors you set for yourself.

   The costliest anchors are rarely set by opponents. Last year's budget decides this year's, the estimate an engineer gave in March survives four changes of scope, the price you once paid quietly determines what you now think the thing is worth. Call these your incumbent numbers, and treat their survival as suspicious rather than reassuring. Once a quarter, take one recurring figure and rebuild it from zero, as though you had arrived this morning knowing the objective and nothing about the number. If the rebuild lands near the incumbent, you have lost an hour. If it doesn't, you have just priced the anchor.