Photometric redshifts, i.e. redshifts derived by comparing an observed spectral-energy distribution (SED) to a range of empirical or theoretical SED templates, are commonly used in studies of the high-redshift Universe. Often, the next step is to use these redshifts as fixed input parameters for SED fitting to derive physical properties for each galaxy. However, this two-step approach ignores degeneracies between redshift and, e.g., stellar mass. Here I present first results using an improved approach that integrates both methods. I find that mass determinations are, on average, three times more uncertain than they seem from the common two-step approach. If not accounted for, these underestimated uncertainties can impact our ability of making meaningful comparisons between observations and simulations of galaxy evolution.