Datasets represent an instance of Marketing Cloud Personalization for use in personalization. User profiles, campaign configurations, and other features of Personalization are all scoped to a dataset. This scoping strategy enables you to use different datasets on different sites, geographies, or deployment environments. For example, while splitting datasets between your QA and production environments is common, it is also helpful to maintain separate datasets for your French and German sites.

Datasets are placed within Identity Groups to coordinate user identities between environments, ensuring that you're not emailing the same user from the French and German environments while still letting each dataset coordinate that user's personalization.