Every hook, loop, setup and question is a promise, and a piece is largely judged on its ratio of promises made to promises kept. This is why a modest piece that pays everything can outperform an ambitious one that pays most things: the unpaid remainder is not neutral, it registers as a small betrayal.
Timing has a shape. A payoff too soon after its setup reads as obvious; too long after and the audience has stopped holding the setup, so the return costs them the work of remembering. The interval that works is usually longer than instinct suggests and shorter than a first draft delivers.
The strongest payoffs recontextualise rather than repeat. The element returns with a different meaning, so the audience gets the recognition and a second thing on top of it.
What is a payoff?
The moment an earlier promise — a setup, a loop, a hook, a question — is kept. Audiences judge a piece largely on the ratio of promises made to promises kept.
Why do endings feel unsatisfying even when nothing is wrong?
Usually unpaid setups. Elements the piece emphasised and never returned to do not register as neutral; they accumulate as a sense that it did not add up.