The bone marrow is the principal site of blood and immune cell production throughout the adult lifespan. Changes within bone marrow niches which regulate haematopoiesis directly contribute to age-related immune decline, an increased risk of cancer and cardiovascular disease, and the onset and progression of blood cancers. Targeting bone marrow dysfunction remains a major translational challenge, where species differences limit the predictive power of animal models and current in vitro systems fail to capture the cellular and architectural complexity that enables lifelong haematopoiesis. The pressing need in the field is for human alternatives. These should accelerate basic discovery (where direct experimentation on living marrow is neither ethical nor practical), personalised medicine (where efficacy and toxicity could be tested in patient relevant models to guide treatment decisions), and preclinical drug development. Advances in stem cell engineering, biomaterials, and micro-physiological systems have brought the idea of building a human bone-marrow niche ex vivo closer to experimental reality. In this review, we examine recent advances in bone marrow organoid development through the lens of how these systems can meet translational needs. We propose that to coordinate progress across the field, there is a need to develop consensus benchmarks to define an ideal bone marrow organoid that can overcome translational barriers. We suggest that a bone marrow organoid should (1) reproduce the diverse stromal and haematopoietic lineages of the bone marrow, (2) self-organize into spatial architectures that recapitulate human haematopoietic niches, and (3) sustain lympho-myeloid output from progenitors over time.