This Article provides the first extended academic consideration of a new practice adopted by an increasing number of courts to resolve e-discovery disputes — the sampling of a small portion of the information sought in backup or other relatively inaccessible files. We provides a comprehensive overview and statistical analysis of contemporary sampling techniques, identifying issues where sampling practice is inconsistent or where additional guidance appears to be required. Our aim is to provide a coherent theoretical approach to the use of sampling, suggesting “best practices” for many unresolved issues, and locating sampling practice within broader contemporary debates about discovery.