Zoos exude a certain self-confidence regarding their roles as education providers. Indeed, the education outputs of zoos are, at face value, pretty impressive, with most investing in learning opportunities for leisure visitors, education groups and in some cases, as part of their in situ programs. However, these outputs are not necessarily reliable indicators of the educational achievements of zoos. Quantity does not necessarily equate to quality, just as outputs do not necessarily lead to outcomes. Zoo-accreditation organizations such as the AZA and EAZA offer us clear insight into the strategic vision underpinning the education goals for zoo visitors; a heightened appreciation of the value of biodiversity and a connectedness with the natural world. Unsurprisingly, most zoos have educational goals that ally neatly with the vision of their respective accreditation body. Consequently, we are left with fairly narrow, top-down educational goals. This does not necessarily sit well with what we know about the unpredictability of “free choice” learning in environments such as zoos and aquariums, or what is known about public science communication. Research that seeks to explore the impacts of zoo visits often focuses on evaluating performance based on educational goals and the findings are used as a means of providing evidence of institutional achievement. However, any visitor outcome that falls outside of this narrow range could well be missed by the research. In this article, we propose that research that takes unpredictable and unexpected outcomes into account is necessary and overdue.