The rules and cases behind eDiscovery decisions

Most eDiscovery disputes turn on a small group of federal rules and an even smaller group of decisions. The useful professional skill is knowing not only what each authority says, but also what it leaves unresolved.

These are United States federal authorities. State courts and other jurisdictions have their own rules, which this reference does not cover.

Rule 26(f) and Rule 16(b): the conference and the order#

Rule 26(f) requires the parties to confer before the scheduling conference, and Rule 16(b) governs the scheduling order. That timing makes the conference a practical point of control over eDiscovery. Issues that remain unsettled at that stage are more likely to return as disputes over responsiveness, burden, production, or privilege.

The conference is where eDiscovery cost is effectively decided. Agreement on scope, sources, formats, search methods, and privilege handling is less expensive than litigation over the same subjects. A resulting agreement also gives the court a concrete benchmark when later evaluating whether the parties conducted discovery as promised.

Custodian lists are a frequent point of review. An early agreement can identify the relevant custodians and explain how sources are being attributed to them, avoiding a later dispute about whether the population selected for processing was adequate. The agreement should also address chat and collaboration sources rather than leaving those systems outside the defined scope. A search plan built around email and document repositories may not answer a dispute involving messages preserved in another system.

Production format deserves the same attention. Native files, TIFF, and PDF—with or without load files—serve different purposes, and the choice affects how a production is handled, reviewed, and used at trial. Early agreement can also identify whether a production will include family relationships, extracted text, or other metadata needed for later use. Left unaddressed, the format is settled later by whoever produces first, and reformatting a production already delivered is more expensive than specifying it once.

Reprocessing is another subject best addressed before data movement begins. The conference can clarify the circumstances that would justify additional processing, who bears its cost, and how a changed collection or agreed production format affects the existing set. Without that clarity, a request for reprocessing can become a second round of negotiation after the costs have already been incurred.

Privilege handling should also be addressed while the discovery process is being designed. The parties can consider how documents will be identified for privilege review, how the privilege log will be produced, and how a clawback process will operate. These arrangements do not displace the court’s authority, but they can reduce avoidable disputes over whether the production and review process followed the agreed method.

Rule 26(f) does not make every issue a conference issue or eliminate later enforcement. Rule 16(b) likewise does not make the scheduling order itself a substitute for the governing discovery rules. The conference and order establish a framework, but they do not decide disputes the parties have not actually resolved or terms the order does not contain.

Rule 26(b)(1): scope and proportionality#

The 2015 amendment moved the proportionality factors into the definition of the scope of discovery itself and added “the parties’ relative access to relevant information” as a factor. That structure matters. Proportionality is not an objection bolted onto discovery after the requesting party has defined the demand; it is part of what “discoverable” means under the rule.

Proportionality therefore cuts in both directions. A responding party may invoke it to resist discovery that is disproportionate to the needs of the case. A requesting party may also invoke it against an overbroad objection when the requested material is proportionate and the objection does not justify withholding it.

The relative-access factor is most useful in asymmetric matters, particularly when one party is better positioned to identify, collect, or produce relevant information than the other. It connects proportionality to the practical distribution of information rather than treating every request as though the parties had equal access and equal resources.

Rule 26(b)(1) supplies a framework for analyzing a request or objection; it does not make proportionality a self-executing answer to every dispute. A court still must determine how the rule applies to the matter, its sources, and the information sought. The relative-access factor is informative rather than conclusive, and it cannot substitute for a concrete examination of the request.

Rule 34(b)(2)(B) and (C): objections and withholding#

Rule 34(b)(2)(B) requires a responding party to state the grounds for objecting to a request with specificity. Rule 34(b)(2)(C) requires the objection to state whether responsive materials are being withheld on the basis of that objection.

Together, these provisions make the basis and consequence of an objection visible. The first requires the responding party to explain the ground rather than merely announce resistance. The second requires the party to identify whether the objection is being used to withhold responsive materials.

That changes the practical posture of boilerplate objections. An objection that reserves everything and discloses nothing no longer complies with the specificity requirement. Similarly, an objection that does not say whether responsive material is being withheld leaves the requesting party without a reliable account of the production’s completeness.

The withholding statement also surfaces disputes early. A party can see that materials are being withheld and examine the stated ground before the end of production, when the objection is easier to correct. Without that information, the significance of the objection may remain hidden until the requesting party compares the requested population with what was produced.

The requirements do not require a responding party to produce privileged material or material outside the permissible scope of discovery. They also do not turn every objection into a merits ruling. They require specificity and an accurate disclosure of the withholding consequence; the validity of the ground remains a separate question.

Rule 37(e): lost ESI and the two tiers#

Rule 37(e) was rewritten in 2015 into a two-tier framework addressing electronically stored information that should have been preserved and was lost. The threshold question is therefore logically prior: the rule reaches information that a preservation obligation should have protected. Rule 37(e) does not create the preservation duty in the first place.

Under subsection (e)(1), a court may order measures no greater than necessary to cure the prejudice if it finds prejudice. The provision therefore links the remedy to the demonstrated harm. A cure must be no greater than necessary, which requires attention to the nature of the lost information and the reason a proposed measure is needed.

Subsection (e)(2) is materially more severe. Only upon a finding that a party acted with the intent to deprive another party of the information may the court presume that the information was unfavorable, instruct the jury that it may so presume, or dismiss or enter default judgment. Intent is the operative threshold for those measures.

Negligence alone does not satisfy the intent requirement, and the volume of lost information does not independently trigger the severe measures under subsection (e)(2). That distinction makes contemporaneous records concerning what was preserved, and why, especially important. Records of preservation decisions and the circumstances surrounding them bear directly on the question whether conduct amounted to negligence, intent, or something else.

The first tier does not ignore the seriousness of loss. It permits a finding of prejudice and a proportionate curative measure. The second tier does not apply merely because a large amount of information disappeared. It requires the specific finding of intent to deprive another party of the information.

Rule 37(e) consequently separates the question of preservation from the question of sanction. It does not make the amount of data lost a proxy for bad faith, and it does not excuse a party from responding when the required intent cannot be established. It also does not displace the need to establish the preservation duty and the loss before selecting the applicable tier.

Federal Rule of Evidence 502(d): the clawback order#

Federal Rule of Evidence 502(d) permits a federal court to order that privilege or work-product protection is not waived by disclosure in the proceeding before it. The order provides a mechanism for handling privileged or protected material that is produced subject to the court’s direction.

Practitioners treat a 502(d) order as cheap insurance, because it can stop a disclosure in the proceeding from operating as an unintended waiver. The order supplies a judicial basis for producing material while preserving protection under the terms the court adopts. That can make it easier to exchange sensitive documents without resolving every privilege dispute first.

A 502(d) order is not a private agreement between the parties. The court enters the order, and the parties operate within its terms. The order’s scope, procedures, and treatment of disclosed material are defined by the order itself, not by an informal understanding that never receives judicial treatment.

The rule also does not decide the underlying merits of a privilege or work-product claim. It addresses the effect of disclosure in the proceeding before the court. Nor does a general practice of using clawbacks substitute for analysis of the order’s terms and the treatment of material within it.

Da Silva Moore v. Publicis Groupe (S.D.N.Y. 2012)#

Magistrate Judge Andrew Peck held that computer-assisted review is an acceptable way to search for relevant ESI in appropriate cases. The parties’ agreement to use that method mattered to the decision. District Judge Andrew Carter affirmed in April 2012, and the decision is widely described as the first judicial opinion approving predictive coding.

The authority is significant because it treated technology-assisted review as a permissible discovery method rather than an automatic substitute for conventional searching. It recognized that the suitability of computer-assisted review depends on the circumstances, including the parties’ agreement to employ it.

What the decision does not say is equally important. It did not make computer-assisted review mandatory, establish that the method suited every matter, or remove the need to examine the search process. It approved a method in a case where the parties had agreed to use it; it did not announce a universal protocol. Cited accurately, it supports considering an agreed technology-assisted process in an appropriate case — not the propositions that agreement is irrelevant, that validation and documentation are unnecessary, or that every producing party must work the same way.

Rio Tinto PLC v. Vale S.A. (S.D.N.Y. 2015)#

Judge Peck explained that where the producing party wants to use technology-assisted review, courts will permit it, describing that proposition as black-letter law. Maura Grossman was later appointed special master in the case to resolve disputes concerning TAR.

The decision separates permission from compulsion. A producing party that wants to use TAR does not have to win an argument about whether the method is allowed at all. Treating it as generally permissible moves the argument to the only question left worth having: whether this use of it, in this matter, is defensible in how it was designed, documented, and administered.

The special-master appointment is a useful signal about the remaining dispute. General judicial acceptance of TAR did not resolve disagreements over how the producing party was conducting the review. The parties still needed a mechanism for resolving implementation disputes, which is why method documentation, validation, and agreed procedures remain important in a technology-assisted workflow.

What the decision does not say is that permission resolves every dispute or that one method is always superior. Nor does it establish that a producing party must use TAR. Its treatment of judicial permission should not be converted into a broader claim about method selection or implementation.

Hyles v. City of New York (S.D.N.Y., 1 August 2016)#

Judge Peck held that a court will not force a producing party to use TAR when it prefers keyword searching. That conclusion aligns with Sedona Principle 6, which places the responding party in the best position to decide how to search for and produce its own ESI.

Read together, the decision and the Sedona principle locate method choice with the responding party, subject to the result being defensible. The responding party may select TAR or Boolean/keyword search, but method preference does not end the inquiry. The resulting discovery process must withstand scrutiny and meet the governing requirements.

The decision does not establish that keyword search is always adequate or that TAR is always unnecessary. It does not insulate a search method from review. Its holding is narrower: a court will not compel TAR merely because the producing party prefers keyword searching.

The authority also does not settle every related disagreement. Questions about custodian selection, search terms, validation, or the treatment of a particular source may still require agreement or adjudication. Method selection is a starting point for resolving a discovery dispute, not a substitute for examining how the selected method produces the result.

Reading the three cases honestly#

Some of the language in Rio Tinto and Hyles is dicta rather than holding. That distinction belongs in the analysis of each decision, not in a footnote. In those cases, the dicta signals judicial attitude; it does not itself establish a binding rule.

All three decisions are from the Southern District of New York. A district-court decision binds no one outside its own case, and its persuasive influence depends on the force of its reasoning rather than a nationwide precedential rule. The influence of these three decisions comes from being early, reasoned, and widely cited.

None of the cases, by themselves, establishes a complete discovery framework for every matter. Da Silva Moore addresses computer-assisted review in appropriate circumstances and highlights the importance of agreement. Rio Tinto treats permission to use TAR more generally while showing that implementation can still generate disputes. Hyles locates method choice with the responding party, subject to a defensible result.

Citing them accurately means identifying the proposition for which each decision is relevant and avoiding the common mistake of treating broad language about TAR as a holding on every question involving search, review, or production.

EDRM: a map, not a source of obligation#

EDRM is the Electronic Discovery Reference Model. Its stages are information governance, identification, preservation, collection, processing, review, analysis, production, and presentation. It provides a common vocabulary for describing eDiscovery work across the matter lifecycle.

EDRM is a reference model, not a mandatory sequence. Real matters frequently loop between stages, particularly processing and review. A collection may require additional processing, and review findings may send data back for further collection, analysis, or revised production.

The model is a framework for discussing work, not a source of legal obligation. It has no legal force of its own. Confusing its descriptive stages with binding duties can create the opposite of its intended utility: an EDRM label may suggest a mandatory sequence when the actual dispute turns on a rule, an order, an agreement, or the facts of the matter.

EDRM also does not determine whether a particular search method is appropriate, whether an objection is sufficiently specific, or what remedy applies when information is lost. It organizes the workflow; the governing authority supplies the legal standard.

Which authority governs which argument#

Matched to the argument being made:

• Scope and proportionality — Rule 26(b)(1).
• Timing, and what had to be settled when — Rule 26(f) and Rule 16(b).
• Objections, and whether anything is being withheld — Rule 34(b)(2)(B) and (C).
• Choice of search method — Da Silva Moore, Rio Tinto, Hyles and Sedona Principle 6.
• ESI that should have been preserved and was lost — Rule 37(e).
• Protection for privileged material produced anyway — Federal Rule of Evidence 502(d).

Where an argument spans two of those, the rule governs and the cases inform. A district-court decision on method does not displace a rule on scope.

Common questions#

What does Rule 26(f) require?

Rule 26(f) requires the parties to confer before the scheduling conference, and Rule 16(b) governs the scheduling order.

What does Rule 26(b)(1) say about proportionality?

Proportionality is part of the definition of the scope of discovery itself. It cuts in both directions, allowing resistance to disproportionate requests and opposition to overbroad objections.

What must Rule 34(b)(2)(B) and (C) require?

Rule 34(b)(2)(B) requires a responding party to state the grounds for objecting with specificity. Rule 34(b)(2)(C) requires the objection to state whether responsive materials are being withheld.

What are the two tiers under Rule 37(e)?

Subsection (e)(1) permits measures no greater than necessary to cure demonstrated prejudice. Subsection (e)(2) permits severe measures only upon a finding that a party acted with intent to deprive another party of the information.