Advances in the availability and affordability of molecular and genomic data are transforming human health care. Surveillance aimed at supporting and improving food safety and animal health is likely to undergo a similar transformation. We propose a definition of ‘molecular surveillance’ in this context and argue that molecular data are an adjunct to rather than a substitute for sound epidemiological study and surveillance design. Specific considerations with regard to sample collection are raised, as is the importance of the relation between the molecular clock speed of genetic markers and the spatiotemporal scale of the surveillance activity, which can be control- or strategy-focused. Development of standards for study design and assessment of molecular surveillance system attributes is needed, together with development of an interdisciplinary skills base covering both molecular and epidemiological principles.