We develop a fitness functional for freezing molecular energy flow that relies only on experimental observables. The functional allows us to implement a modular control algorithm where simulation data and experimental data can be used interchangeably. This interchangeability could be useful as a spectroscopic tool and for reactive control because the controllability of the experimental system and its model can be compared directly. The fitness functional performs as well as functionals based on complete knowledge of the wave function. We compare our simulation results with an analytical theory of control, and find good agreement between the simulated and predicted times over which the system can be controlled.