Nearly 50 years after Hardin's 'tragedy of the commons' we have not yet found predictive tools to guide us towards sustainable management of common-pool resources (CPR). We often have a good understanding of the qualitative relationships between the principal actors in socioecological systems (SESs), but classical quantitative approaches require a tremendous amount of data to understand the drivers of SESs sustainability. Here we show that qualitative modelling approaches can provide important governance insights for SESs that are understood but not quantified. We used Loop Analysis to test the outcomes of different management regimes on a simple nature-based tourism SES described by economic, social and environmental variables. We tested the sustainability of different management scenarios on this system and we identified the necessary conditions to achieve it. We found that management regimes where property rights and responsibilities are shared between different stakeholders are more likely to be successful. However, the system is generally highly unstable and it is important to tune each strategy to each particular situation. The conditions for sustainability found across the different systems tested were: a low reinvestment rate of tourist revenues into new infrastructures and a low growth rate of the environment. Management strategies based on maximum sustainable yield, which keep the environment far from its carrying capacity, have less chance to be sustainable. Qualitative models of SESs are powerful diagnostic tools; they can help identifying variables that play an important role in determining socioecological sustainability in data-poor circumstances and guide the design of efficient data collection programmes. Our results highlight the importance of careful planning when designing management strategies for nature-based tourism. The application of one-size-fits-all solutions to every situation is likely to lead to the failure of the commons; however tourism-based SESs can be sustainable if management strategies are tuned to each particular case.