Acute respiratory distress syndrome (ARDS) is a heterogeneous syndrome, presenting challenges to both frontline clinicians and researchers testing novel therapies. Identifying more homogeneous subgroups within the broader ARDS population could allow for more efficient testing of interventions in targeted cohorts of patients, as well as better-tailored therapy at the bedside. Over the past decade, various physiologic, clinical, and biologic characteristics have been used to identify ARDS patients at highest risk for poor clinical outcomes. Using those markers to select patients for enrollment in clinical trials is called prognostic enrichment. Similarly, some studies have shown that different subgroups of ARDS patients respond differently to a specific therapy. Selecting patients for trials based on a higher likelihood of responding to a therapy is called predictive enrichment. For example, a mix of clinical and biologic markers has been used to identify two distinct subphenotypes of ARDS—termed hypoinflammatory and hyperinflammatory—that not only are present in multiple different clinical trial populations, but have divergent clinical outcomes and may respond differently to several experimental therapies. The future of the field will involve developing more pragmatic methods to stratify ARDS patients, harnessing new ‘omics technology to improve subphenotyping, and using subphenotypes prospectively to guide enrollment in clinical trials of focused therapies for ARDS. Advances in our ability to discern subphenotypes within the general ARDS population could help us move from a one-size-fits-all approach to ARDS management to therapy tailored to the unique clinical and biologic profile of each ARDS patient.