Two of the most fundamental questions regarding the broad line region (BLR) are "what is its structure?" and "how is it moving?” Baldwin et al. (1995) showed that by summing over an ensemble of clouds at differing densities and distances from the ionising source we can easily and naturally produce a spectrum similar to what is observed for AGN. This approach is called the `locally optimally emitting clouds' (LOC) model. This approach can also explain the well-observed stratification of emission lines in the BLR (e.g. Clavel et al. 1991, Peterson et al. 1991, Kollatschny et al. 2001) and `breathing’ of BLR with changes in the continuum luminosity (Netzer & Mor 1990, Peterson et al. 2014) and is therefore a generally accepted model of the BLR. However, LOC predictions require some assumptions to be made about the distribution of the clouds within the BLR. By comparing photoionization predictions, for a distribution of cloud properties, with observed spectra we can infer something about the structure of the BLR and distribution of clouds. I use existing reverberation mapping data to constrain the structure of the BLR by observing how individual line strengths and ratios of different lines change in high and low luminosity states. I will present my initial constraints and discuss the challenges associated with the method.