Reaching developmental milestones younger and at larger sizes is commonly claimed to reflect increased fitness. However, the amount of fitness gained from being larger and younger at a milestone may vary with several attributes, particularly evolutionary history, life history, and environmental characteristics. We used a meta-analysis to investigate whether these attributes affected the utility of developmental milestones to be used as predictors of future fitness. We chose amphibian size at and time to metamorphosis (SAM and TTM, respectively) as model developmental milestones, because studies have examined SAM and TTM's efficacy for fitness prediction (via post-metamorphic fitness proxies), and they are commonly used in a variety of studies testing ecological and evolutionary theory and more applied research on the effects of anthropogenic stressors. We found variation in the predictive power of SAM and TTM for post-metamorphic performance. SAM was a more consistent predictor of post-metamorphic performance than TTM, but also had a higher sample size. Life history and study design (i.e., laboratory vs. field studies), but not evolutionary history, were important for explaining variation in predictive power for post-metamorphic performance. The correlation between SAM and performance increased with the proportion of time to maturity reached at metamorphosis, suggesting that species can compensate for initial fitness reductions through ontogeny. Because numerous researchers use size and age at developmental milestones to indicate fitness, we urge caution in interpreting their results due to the species- and system-specific nature of fitness surrogates.