There is incongruity between much of the research that is ostensibly directed at improving marine conservation outcomes, and the effort that is actually needed to generate better outcomes. I argue that this is partly due to inadequate critique of basic assumptions by marine researchers. One assumption that is frequently made (explicitly or implicitly) is that marine reserves need to contain representatives of all species if they are to be effective. This assumption is not supported by evidence: many (perhaps most) species of marine organisms do not need the protection that marine reserves confer, because they are not threatened by processes that can be effectively managed by marine reserves. Better conservation investments are likely to be made if researchers critically evaluate basic assumptions, and design their research to address real information needs. Improved recommendations are likely to be achieved by research directed at understanding which species need protection most, and on how marine protected areas (MPAs) can be designed and managed to conserve viable populations of those species.