Ask for too much so the real request feels like mercy.
Cialdini's 1975 study is chapter 2. Students asked to counsel juvenile delinquents two hours a week for two years refused essentially unanimously. Asked immediately afterwards to take a group of juveniles to the zoo for one afternoon, half agreed. Asked about the zoo trip cold, seventeen percent did.
Two mechanisms run together, and separating them matters because they can be defeated separately.
Reciprocal concession. The requester has visibly given something up, and movement in a negotiation obliges movement in return. This is the reciprocity of the previous law applied to positions rather than to objects.
Perceptual contrast. An afternoon is a real cost, but measured against two years it is nothing, and the target evaluates it against the rejected anchor rather than against their actual calendar.
The finding that makes this more than a trick is the follow-up. Subjects who agreed after the retreat were more likely to turn up, and more willing to volunteer again, than those who agreed to the direct request. The technique does not merely extract a yes; it produces a target who feels responsible for the outcome, because from their perspective they negotiated the requester down.
The operational requirements are two. The opening must be large enough to be refused but not so large as to appear unserious — an absurd anchor reads as bad faith and forfeits the relationship. And the retreat must appear to cost the requester something, which is why the second request is framed as a concession rather than as the plan.
The boundary is exactly there, at the point where the opening bid is not a position anyone holds. Union bargaining, litigation demands and salary negotiation all open at positions both sides know are theater, and this is understood by both and is not deception. Opening at a fabricated position with someone who does not know the convention — a consumer, a patient, a person negotiating alone against an institution that does this professionally — is exploiting an asymmetry of experience rather than making an offer.
And the technique destroys trust when detected, which is the practical constraint on its use. A target who realizes that the first request was never real will discount every subsequent one.
The counter is chapter 304's discipline: evaluate each request against what you actually want, not against the one before it.
Countered by Nothing You Say