A Dog and Its Tail: Don’t Let Version Uncertainty Cloud Linked Attachment Production

Key Takeaways

• Courts now require capability testing for linked attachment recovery
• Sedona 2025 commentary sets baseline for reasonable e‑discovery steps
• RG standard distinguishes preservation gap from context gap
• Versioning uncertainty should not justify non‑production of attachments
• Collect and search current versions; aim for reconstruction‑grade over time

Summary

Linked (cloud) attachments remain a critical e‑discovery blind spot, despite tools that can collect and search them. Recent developments—such as the Carvana court‑ordered capability test, the Sedona Conference’s 2025 commentary, and the proposed Reconstruction‑Grade eDiscovery (RG) standard—reinforce the duty to produce these files. However, the RG standard’s focus on "as‑sent" version resolution is being weaponized, allowing producers to claim versioning uncertainty and refuse collection altogether. The author argues that collecting the current version satisfies the floor of reasonable steps, while the industry works toward reconstruction‑grade fidelity.

Pulse Analysis

The rise of cloud‑based collaboration has outpaced traditional e‑discovery practices, leaving linked attachments—documents accessed via URLs—largely invisible to legacy collection tools. Courts are beginning to push back, as illustrated by the Carvana securities litigation where a judge demanded a bounded forensic test to prove what could be recovered. Simultaneously, the Sedona Conference’s 2025 commentary formalizes the expectation that parties take reasonable steps under Rule 37(e) to preserve and produce collaborative data, turning vague best‑practice claims into measurable obligations.

At the heart of the debate is versioning. The Reconstruction‑Grade eDiscovery (RG) standard proposes an "as‑sent" version requirement, arguing that the version a recipient saw when the link was clicked is the gold standard. In practice, this requirement can become a loophole: producers cite uncertainty over which version to deliver and opt to produce nothing. Empirical data on how often linked files change after transmission is scarce, but industry intuition suggests post‑send modifications affect fewer than twenty percent of cases, varying widely by organizational culture. Without solid metrics, courts risk basing decisions on speculation rather than evidence.

Practitioners should adopt a pragmatic two‑track approach. First, use existing platform tools—Microsoft Purview, Google Vault, or forensic collectors—to gather and search the current version of every linked attachment, documenting any gaps. Second, invest in capability testing and incremental improvements toward reconstruction‑grade fidelity, leveraging the RG framework’s transparency provisions. This balanced strategy satisfies the immediate legal floor while positioning firms for future compliance as tools mature and courts increasingly demand demonstrable collection capabilities.