Deletion Alone Is Not Sanctionable — Intent to Deprive Is

Deleting data is not itself sanctionable under Rule 37(e). The severe measures — an adverse-inference instruction, dismissal, default — require a finding that the party acted with intent to deprive another party of the information's use in the litigation. Everything else is capped at measures no greater than necessary to cure prejudice. Forensics rarely proves intent directly; it documents the pattern from which a court infers it.

This article is educational. It is not legal advice, and it is not a substitute for reading Rule 37(e) and the opinions themselves.

What has to be true before Rule 37(e) applies at all?

The rule is gated, and the gates are frequently skipped in briefs that jump straight to sanctions. Four predicates:

1. The information was electronically stored information. Rule 37(e) does not govern the loss of paper, physical objects, or a witness's memory.
2. It should have been preserved in the anticipation or conduct of litigation. That imports the common-law trigger rather than creating a rule-based one.
3. It was lost because the party failed to take reasonable steps to preserve it. A loss that reasonable steps would not have prevented is outside the rule.
4. It cannot be restored or replaced through additional discovery.

Fail any one and the analysis ends. The fourth gate does more work than practitioners expect, because messages deleted from a phone frequently survive on a counterparty's device, in a platform's server-side store, or in a cloud backup. A motion that has not tested those sources is vulnerable on its own terms. The reference entry for Rule 37(e) sets out the provision and its structure.

The trigger question is the one most often litigated on a thin record. Cache La Poudre Feeds, LLC v. Land O'Lakes, Inc. held that the duty requires more than a mere possibility of litigation: an equivocal expression of discontent that produced no lawsuit for nearly two years did not trigger it, and on those facts the duty attached when the complaint was filed. The same opinion states the scope of the duty once triggered — it reaches the employees likely to have relevant information, the key players.

What is the difference between Rule 37(e)(1) and 37(e)(2) in practice?

Everything, because they run on different findings and produce different remedies.

	Rule 37(e)(1)	Rule 37(e)(2)
Required finding	Prejudice from the loss	Intent to deprive
Remedy ceiling	Measures no greater than necessary to cure the prejudice	Presumption of unfavourability, adverse-inference instruction, dismissal, default
Separate prejudice finding needed?	Yes — prejudice is the trigger	No
Typical proof	Comparing what was produced against what should exist	Circumstantial pattern evidence

Two decisions fix the contours. Hoffer v. Tellone held that Rule 37(e)(2) sanctions require a finding, by a preponderance of the evidence, that a party acted with intent to deprive, and that the older and looser "culpable state of mind" standard — which encompassed negligence — does not apply to subdivision (e)(2). The Ninth Circuit in Jones v. Riot Hospitality Group LLC held that intent under the rule means wilful destruction for the purpose of avoiding discovery, that it may be proved by circumstantial evidence, and — importantly for motion drafting — that Rule 37(e)(2) requires no separate finding of prejudice.

On the (e)(1) side, Barbera v. Pearson Education, Inc. is the template for a cure that is genuinely proportionate: where an employer lost a key email exchange, the magistrate judge barred the employer from disputing the plaintiff's description of the messages and declined further sanctions. The Seventh Circuit affirmed, and then affirmed summary judgment anyway, because the claim failed on the merits even accepting the plaintiff's version of the missing emails entirely. That sequence is a useful warning to movants: winning the spoliation motion and winning the case are not the same project.

What can forensics actually show about state of mind?

An examiner cannot image intent. What an examiner can do is document a set of facts whose combination is difficult to explain innocently, and courts have been explicit that this is how the finding gets made.

The clearest inventory comes from Jones v. Riot Hospitality, where the court pointed to selective deletion alongside preserved contemporaneous messages, the acquisition of new phones just after the imaging order, and a surviving screenshot of a deleted message bearing directly on the case. Each fact alone is ambiguous. Together they are not.

Categories that carry weight, and the artefact classes behind them:

• Selectivity. Messages in a thread are gone while adjacent messages on the same day, in the same app, survive. The supporting analysis is comparative message-volume work between device pairs over time, plus surviving thread fragments. That is close to the method the Ninth Circuit accepted in Jones.
• Timing against a known legal event. Deletion or device replacement immediately after a hold notice, a preservation letter, a complaint, or an imaging order. The evidence is filesystem and application timestamps, app install and uninstall records, device-provisioning and backup-restore records.
• Tool use. Installation or execution of wiping or privacy software, or clearing of unallocated space. In Victor Stanley, Inc. v. Creative Pipe, Inc., a sustained pattern of deletion, use of wiping software, and failure to preserve devices led to a recommended default judgment on the copyright count, a contempt finding, and a fee award. In Genger v. TR Investors, LLC, deletion of files and wiping of unallocated free space on a work computer supported contempt of a status quo order and a multi-million-dollar fee and expert-cost award — though the Delaware Supreme Court cautioned that its affirmance should not be read as extending beyond the narrow facts.
• Coordination with others. Jones involved the plaintiff coordinating with witnesses to delete theirs, and the court upheld orders directing the plaintiff and non-party witnesses to surrender their phones to a jointly retained forensic specialist.
• Alteration rather than deletion. Cat3, LLC v. Black Lineage, Inc. treated the manipulation of documents so that the authentic version was unavailable as a loss of information within Rule 37(e), and relied on a forensic examination showing the altered addresses could not have arisen from any automatic process. The examination's finding — that the state of the data resulted from intentional human action rather than an automatic or inadvertent computer process — is the exact shape of a useful opinion on this issue. The court imposed preclusion plus fees, and refused dismissal and an adverse inference as more than necessary.
• Internal impossibility. In Rossbach v. Montefiore Medical Center, the plaintiff offered a photograph of text messages that forensic analysis showed could not have been produced on the device she said had received them. Dismissal was affirmed; only the portion sanctioning her counsel was vacated, because the district court had applied the wrong legal standard to the attorney.

The methodological caveat matters here, and Jones states it. The court found the examiner's methodology sound notwithstanding the examiner's acknowledgement that no industry standard for that analysis exists. That is a narrow endorsement rather than a general licence. NIST SP 800-101 Rev. 1 on mobile device forensics and the SWGDE mobile device collection best practices anchor the acquisition half of the work, where standards do exist; the interpretive half is where a Rule 702 fight is most likely.

Does failing to issue a litigation hold decide the case?

No, and the framing has shifted. Under the pre-2015 federal framework, Zubulake v. UBS Warburg LLC set out a three-element test for an adverse inference — an obligation to preserve at the time of destruction, a culpable state of mind, and relevance of the lost evidence — and imposed an instruction plus costs. Pension Committee of the University of Montreal Pension Plan v. Banc of America Securities, LLC found that several plaintiffs failed to issue written litigation holds or to preserve and collect relevant records, and imposed graduated sanctions calibrated to each party's conduct, including a rebuttable-presumption instruction for the more culpable. The Fifth Circuit's more demanding approach appears in Rimkus Consulting Group, Inc. v. Cammarata, which permitted an adverse inference only if the jury found the deletions were intended to prevent the evidence's use in litigation, and denied the request to strike the pleadings and enter default.

That patchwork is what the 2015 amendment replaced in federal court, and the negligence-based route to an adverse inference did not survive it — that is the express holding of Hoffer. Read those earlier cases for their factual anatomy of preservation failure, not for their sanctions standard.

State practice diverges, and this is a genuine split rather than a lag. New York's First Department in VOOM HD Holdings LLC v. EchoStar Satellite L.L.C. adopted the Zubulake rule for New York, holding a defendant grossly negligent where it issued no hold until after suit and the hold it then issued did not switch off automatic email deletion — with relevance presumed and an adverse-inference sanction affirmed. The New York Court of Appeals then held in Pegasus Aviation I, Inc. v. Varig Logistica S.A. that failure to institute a litigation hold is not gross negligence per se, but one factor among many in assessing culpable state of mind. Texas runs a different framework again: Brookshire Brothers, Ltd. v. Aldridge makes spoliation a two-step judicial process — the court decides as a question of law whether spoliation occurred, then assesses a proportionate remedy — with those findings made outside the jury's presence and spoliation evidence kept from the jury except as it relates to the substance of the suit. If you practise in state court, the federal intent standard is not necessarily your standard.

When does a preservation failure produce something other than an inference?

Often, and this is the under-used part of the toolkit.

• Further discovery as the remedy. Treppel v. Biovail Corp. found an adverse inference unjustified on the record but ordered additional discovery, including a forensic search of the opposing party's computer systems. The preservation failure supplied the justification for an intrusion that would not otherwise have been ordered.
• Compensatory monetary sanctions. Klipsch Group, Inc. v. ePRO E-Commerce Ltd. affirmed sanctions equal to the costs the opposing party reasonably incurred in remedying non-compliance, plus jury instructions and a $2.3 million bond, and rejected the argument that a compensatory sanction becomes impermissibly punitive because it dwarfs the likely value of the case.
• Permissive rather than mandatory inference. Goodman v. Praxair Services, Inc. imposed an instruction allowing but not requiring the jury to draw an inference, and denied the other relief requested.
• Terminating sanctions where court orders were violated too. Calsep A/S v. Dabral affirmed a default judgment against a trade-secret defendant who destroyed electronic evidence, including deleting material from a source-code control system, in violation of several court orders — and held the district court was not required to analyse lesser sanctions in any greater detail than it did, given its earlier leniency. Order violations and spoliation compound.
• Nothing. Reinsdorf v. Skechers U.S.A., Inc. declined to reopen discovery or impose sanctions where the evidence of misconduct was insufficient, and Integrated Direct Marketing, LLC v. May denied an adverse inference while granting sanctions on different grounds — failure to produce documents, submission of a false affidavit, and refusal to answer deposition questions.

One structural limit is worth knowing before you ask for it. In Clientron Corp. v. Devon IT, Inc., the Third Circuit held the district court committed legal error in piercing a corporate veil as a discovery sanction so as to reach one shareholder but not the other, and vacated the sanctions order for a new sanction to be imposed — the misconduct findings stood; the chosen remedy did not.

What does this mean for how the examination is scoped?

If the legal question is intent, the examination should be scoped to the intent indicators rather than to content recovery. Content recovery answers "what did we lose", which matters to prejudice under (e)(1). Intent lives in the metadata of the loss: when, in what order, by what mechanism, against what backdrop of legal events, and with what asymmetry between what went and what stayed.

Two practical consequences. First, the examiner needs the legal chronology — hold date, complaint date, order dates — before the analysis rather than after, because the whole inference depends on alignment with it. Second, the report has to be written to survive Rule 702, and that is not automatic: Jones upheld admission of the spoliation expert's report without a Daubert hearing, but did so while noting the absence of an industry standard for the analysis. Expect the opposing motion. The Digital Forensics for Lawyers track covers the scoping conversation between counsel and examiner that most of these motions turn on, and the Sedona Conference Commentary on Discovery of Mobile Device Data is the best guide to the preservation decisions that avoid the motion entirely.