With the continuing threat to aviation security from homemade explosive devices, the restrictions on taking a volume of liquid greater than 100 ml onto an aircraft remain in place. From January 2014, these restrictions will gradually be reduced via a phased implementation of technological screening of Liquids, Aerosols and Gels (LAGs). Raman spectroscopy offers a highly sensitive, and specific, technique for the detection and identification of chemicals. Spatially Offset Raman Spectroscopy (SORS), in particular, offers significant advantages over conventional Raman spectroscopy for detecting and recognizing contents within optically challenging (Raman active) containers. Containers vary enormously in their composition; glass type, plastic type, thickness, reflectance, and pigmentation are all variable and cause an infinite range of absorbances, fluorescence backgrounds, Rayleigh backscattered laser light, and container Raman bands. In this paper we show that the data processing chain for Cobalt Light Systems’ INSIGHT100 bottlescanner is robust to such variability. We discuss issues of model selection for the detection stage and demonstrate an overall detection rate across a wide range of threats and containers of 97% with an associated false alarm rate of 0.1% or lower.