Continuous measurement of collective atomic spin by an off-resonant optical probe can produce conditional spin-squeezing, the degree of which is limited by the signal to noise ratio of the measurement. Here, we relate the amount of achievable squeezing to characteristic experimental parameters including the number of atoms, the volume they occupy, the intensity and detuning of the optical probe field, and the measurement duration. Our approach combines techniques from filtering theory with an analysis of the atom-probe scattering physics based on a semiclassical treatment of the atoms. From this framework, we develop a calibration of conditional spin-squeezing and demonstrate the procedure for computing it using laboratory squeezing data.