Flywheels take time to get up to speed.

Time to bake; time for the acorn to grow.

If the output metric is evaluated too early or with too high a target then there are no moves to start the flywheel (or plant the acorn) that hit the target in time.

In that case the only move for the owner of the metric is linear strategies.

The solution is to either delay the timeline of evaluation or set the goals more modestly at the beginning to allow a flywheel-shaped strategy to be viable.