The vast majority (> 99.99%) of spiders are solitary, but in some species, individuals form stable groups or, alternatively, looser aggregations that fluctuate over space and time. While previous work has documented trade-offs associated with being part of intraspecific groups, rather than being solitary, we know less about why grouping behavior would be facultative. Residing in a group can reduce predation risk, but also lead to competition for food, and we hypothesized that variation in these trade-offs could provide insights into the optional nature of clustering. We used a Costa Rican population of golden orb-web spiders, Nephila clavipes, to ask two questions. First, we tested whether a spider’s size influenced the trade-offs of clustering, and second, whether inter-annual variation in predator and prey abundance could help explain clustering patterns. We also examined the process of cluster formation and dissolution in relation to spider size. Data collected on over 1500 spiders over 4 years indicated that in each year, small spiders experienced the greatest reduction in predation if clustered and also were most likely to initiate aggregations. Overall rates of predation varied significantly among years, and small spiders were most likely to cluster in the year with highest predation. We also detected a cost of clustering in the form of reduced prey capture, but only in years with relatively high overall prey acquisition. Together, these findings suggest that trade-offs that vary individually and over time could influence a spider’s decision to cluster and, hence, explain the facultative nature of aggregating in this population. Animals within a population can show variability in a range of behaviors, such as foraging and predator avoidance, and at times, the ideal way of behaving depends on individual or environmental conditions. In this study, we asked why some female golden orb-web spiders (Nephila clavipes) form groups, while in the same population, others are solitary. This is an interesting question given how rare aggregations are among spiders. In our study, small spiders benefited most from clustering in terms of higher survival when aggregated compared to when solitary. At the same time, spiders of all sizes captured less prey when aggregated, although this cost was only detected in years with relatively abundant prey. Together, these findings suggest that trade-offs that vary individually and over time could influence a spider’s decision to cluster and, hence, explain the optional nature of aggregating in this population.