Placebo response impacts the development of treatments for central nervous system disorders. Despite significant attention in scientific literature, the magnitude of placebo response continues to increase. The present work reviews empirical and theoretical literature on associations of trial design features, patient selection, and investigative site selection with signal detection and placebo response. Trials with greater percentages of patients randomized to placebo had smaller average placebo responses and larger average drug–placebo differences in comprehensive meta-analyses of major depressive disorder and schizophrenia. Excluding patients with large responses during double-blind placebo lead-ins reduced placebo response and increased drug–placebo differences, but that advantage was nearly offset by the decrease in sample size. Core factor subscales of the Hamilton Rating Scale for Depression (HAMD) yielded larger drug–placebo differences than the HAMD total score. Theoretical considerations suggest that the number of sites and number of patients per site can influence power, with the optimal combination of the two being influenced by the correlation structure and therefore being situation dependent. Similar considerations apply to the number of countries and number of sites per country. Sequential parallel comparison and randomized withdrawal designs hold promise for improved signal detection. Patient selection and assessment of changes in illness severity may benefit from remote and independent evaluation as well as patient self-reports. Given the complexities of signal detection and placebo response, no single strategy will solve all the problems and combinations of approaches may be most useful.