Performance budgets are a growth lever, not a dev preference

Speed work usually happens too late — after the hero video, the four font weights and the tag manager are already live, when every fix means taking something away from someone.

A performance budget flips the order. Agree the limits up front: a maximum payload, a target for largest contentful paint, and a cap on third-party scripts. Design and engineering then make trade-offs inside those limits.

Budgets also settle debates cheaply. A carousel isn't rejected on taste; it's rejected because it doesn't fit the budget, and the team chooses what to remove to make room.

Enforce it in CI. A check that fails the build when the bundle grows keeps the budget alive after launch, which is exactly when most sites quietly get slower.