Email Threading and Near-Duplicate Detection in E-Discovery

Email threading uses message relationships and metadata—such as reply references, conversation identifiers, and subject information—to group related messages into a thread. Reviewers can then understand context across replies instead of treating every message as isolated. A thread is not necessarily a perfect chronology: subject lines can repeat, headers may be missing, and messages can be forwarded or moved between systems. Near-duplicate detection groups documents that are highly similar but not identical. It can help reviewers compare drafts and identify small changes, but a similarity threshold is a technical setting, not a legal definition of duplicative or nonresponsive. A small change may contain the important admission, edit, recipient, or attachment. Exact deduplication is a separate task. EDRM’s Message Identification Hash specification supports cross-platform email duplicate identification by combining message identifiers with hash-based matching. Deduplication should preserve custodian and source information when those details matter to production or review. Before suppressing messages, determine whether the review method shows the most inclusive email and attachments, exposes unique content in replies, and preserves responsive information. Test with representative threads, missing or malformed headers, changed subjects, and near-duplicate drafts. Record the system, settings, criteria, and exceptions. Threading and similarity groups are aids to navigation and prioritization; they do not decide responsiveness, privilege, or production scope. Protocols and local rules can control the treatment of duplicates and email families.