Decoupling Goals from Rewards

Definition

Setting a "stretch goal" — the best possible outcome if everything goes right — purely to stimulate ambitious thinking, while explicitly and permanently disconnecting that goal from how performance will later be evaluated and rewarded. The point of the target is to change what people imagine is possible, not to define what they must deliver.

In the Book

Chapter 4 opens its first principle with Jean-Marie Descarpentries's practice at Carnaud Metal Box in the 1980s and Groupe Bull in the mid-1990s: each business unit proposed its own "best possible outcome" stretch target, and then — the crux of the method — Descarpentries "would promptly forget about the target." He evaluated and rewarded managers instead on relative indicators (performance versus last year, versus the competition), never against the stretch number itself. His stated purpose was "to get managers to dream the impossible dream" without triggering the sandbagging that occurs the moment a number becomes a promise someone will be judged against. The book contrasts this with ordinary target-setting, where, as Jack Welch is quoted saying, "making a budget is an exercise in minimalization... everyone is negotiating to get the lowest number," because any target that will be scored creates an incentive to lowball it.

Why It Matters

Most organizations conflate two jobs a target can do — inspiring bigger thinking, and defining a measurable bar for reward — and bundling them into one number destroys the first job to protect the second: nobody proposes their true ceiling if it becomes the floor they'll be punished for missing. Separating "what would be great" from "what you'll be scored against" is a reusable move anywhere aspiration and accountability collide, from stretch goals in OKRs to sprint estimates a team fears will be treated as a commitment.