Weed-suppressive bacteria (WSB), specifically the D7 and ACK55 strains of Pseudomonas fluorescens, were previously reported to selectively inhibit growth of invasive annual grasses (IAGs) that have caused severe ecosystem degradation across much of the western US. However, recent studies show highly mixed evidence for WSB effectiveness in field settings. We evaluated how the effectiveness of WSB, specifically selective inhibition of exotic annual but not native perennial grasses, was altered by (1) growth media (soil vs. agar culture), (2) pre-sterilization to remove competing microbes, and (3) WSB concentration, reduced by centrifuging with and without filtration. WSB had no effects in soil microcosms with or without pre-sterilization, but WSB did have effects in agar cultures. On agar, WSB effects were amplified by pre-sterilizing seeds prior to treatment with WSB inoculum – but effects were non-selective (i.e., native and exotic plants were equally inhibited) with or without pre-sterilization at high WSB concentration. The only selective effects of WSB were observed on agar at low WSB concentrations, made possible by centrifuging WSB cells out of the inoculum prior to treatment application (from ~10⁷ to ~10⁵ live cells/ml). Completely removing WSB cells from inoculum by filtration led to no WSB effects. While these results suggest the desired effects of WSB can be reproduced on agar with a supernatant containing the putative active agent and a low density of WSB cells, the effect is still not reproducible for plants in soil, even when competing microbes are removed. Thus, additional research is needed to determine if, when, where, and how WSB could be effective in field settings.