A Practical Discovery Refresher: Mastering the Tools, Rules, and Pitfalls of Federal Civil Litigation

In brief. Discovery is where most civil cases are actually won or lost—and where the most avoidable mistakes are made. This guide walks through every core discovery tool under the Federal Rules of Civil Procedure, the objections that surround them, the privilege practice that governs what stays hidden, the document-handling and e-discovery practices that keep a case out of trouble, and the full Rule 37 sanctions apparatus that punishes the lawyers who get it wrong. It is written to be useful to a judge, a seasoned litigator, and a newer attorney or paralegal alike, and reflects the rules as amended through December 1, 2025. It is not legal advice; rules and local practices vary, so always check the governing rules for your court.

Introduction: Why Discovery Mastery Matters

Here is a number that should concentrate the mind of every litigator: in the modern federal case, somewhere between fifty and ninety percent of the total legal spend is consumed not by trial—most cases never see one—but by discovery. The document review, the depositions, the fights over search terms and custodians and privilege logs—that is where the money goes and the calendar disappears. A trial is theater. Discovery is the engine room, and the engine room is where ships sink.

Discovery is at once the most powerful and the most misused phase of civil litigation: the process by which parties obtain the facts, documents, and testimony to build their own case and anticipate their opponent's. Done well, it narrows the issues and makes settlement meaningful; done poorly, it breeds delay, sanctions, and malpractice.

The Federal Rules of Civil Procedure—principally Rules 26 through 37—supply the governing framework, all under the overarching command of Rule 1 that the rules secure the "just, speedy, and inexpensive determination of every action and proceeding." Those rules have been amended repeatedly to respond to the explosion of electronically stored information, curb the cost and burden of discovery, and insist on proportionality between the scope of discovery and the genuine needs of the case. The most recent amendments, effective December 1, 2025, push privilege-log planning to the front of the case rather than leaving it for the end. Yet despite the clarity of the rules and the mountain of case law interpreting them, discovery errors remain among the most common and consequential mistakes practitioners make.

To keep the discussion concrete, this guide follows a single hypothetical dispute (used throughout for illustration only). Meridian Manufacturing, Inc., a mid-sized maker of industrial pumps, has sued Cardinal Components, LLC, a former supplier, for breach of a supply agreement. Meridian has also added a trade-secret claim against Dana Reyes, a senior design engineer who left Meridian for Cardinal and, Meridian suspects, took proprietary pump-housing designs along the way. As we move through each tool, we will watch how it plays out for Meridian, Cardinal, and Dana—because the same rule looks very different depending on whether you are sending the request or answering it.

This guide proceeds from public records requests—a form of pre-litigation and parallel discovery practitioners often overlook—through the formal architecture of the Federal Rules: initial disclosures, interrogatories, requests for production, requests for admission, depositions, and subpoenas. It then addresses the objections that surround those devices, the privilege doctrines and logs that govern what may be withheld, the mechanics of documenting an opponent's deficiencies, document review and redaction, electronic discovery, and—because every one of these obligations has teeth—the Rule 37 sanctions that enforce them. Throughout, the emphasis is practical: what to do, what not to do, and why.

Discovery is also where the relationship between opposing counsel is most severely tested. Courts have little tolerance for lawyers who cannot cooperate, who run to the courthouse over disputes a phone call would resolve, or who wield discovery as a weapon. The attorney who masters discovery pairs knowledge of the rules with professionalism and a genuine willingness to work things out. For the broader arc of a case into which discovery fits, see our comprehensive guide to federal civil litigation for small businesses, and for the threshold question of whether a matter is worth pursuing at all, our guide to evaluating and assessing a civil case.

Part One: Public Records Requests

Understanding What Public Records Are

Before formal litigation discovery begins—and often alongside it—practitioners frequently need information held by government agencies: police reports, corporate records, agency policies, regulatory filings, and inspection records. Accessing it is the domain of the public records request, governed federally by the Freedom of Information Act (FOIA), 5 U.S.C. § 552, and at the state level by each state's analogous open-records or "sunshine" law.

FOIA gives any person—not just a litigant—the right to request records held by federal executive-branch agencies. For litigators, public records are valuable and underused for two reasons. First, the documents sit with a government agency rather than the opposing party, so the opponent cannot obstruct, slow-walk, or "lose" them. Second, they often predate the litigation entirely and were created without any litigation motive, which makes them credible and hard to attack: a self-serving email written after a lawsuit looms gets discounted, but an inspection report filed two years before anyone thought about suing carries the quiet authority of the contemporaneous.

In our hypothetical, Meridian's counsel realizes that Cardinal's plant has been inspected by OSHA and is regulated by a state environmental agency. Inspection reports and regulatory correspondence might reveal when Cardinal retooled its line to produce the suspect pump housings—evidence bearing directly on when Cardinal began using Meridian's designs. None of that sits in Cardinal's litigation files; it sits in a government file room, beyond Cardinal's control—exactly the independent, hard-to-spin corroboration that can turn a "he said, she said" trade-secret claim into one anchored in dates no one can argue with.

The Four-Step Public Records Request Process

Step one: define what you need and who has it. Precision is everything. Federal agencies are sprawling bureaucracies, and a request aimed at the wrong sub-office, or written too vaguely, will be bounced or denied. If you want a particular kind of record—an inspection report, an enforcement action, an internal policy—learn which unit within the agency maintains it; that research pays for itself in faster, more accurate responses. And if several agencies might plausibly hold the records, send targeted requests to each, tailored to that agency's holdings.

Step two: search online before submitting. Before filing any formal request, search the agency's website and its FOIA "reading room." Agencies are required to post—and many proactively post—frequently requested documents, final opinions, policy statements, and staff manuals; if what you need is already public, a formal request just wastes everyone's time. Cast a wider net, too: journalists, researchers, and advocacy organizations routinely publish records obtained through earlier FOIA requests, so for records with public-interest value there is often a substantial body of already-disclosed material that no one has analyzed.

Step three: draft and submit the request. A good request describes the specific records sought, with dates, subject matter, and any other parameters that narrow the search. State your preferred delivery method—email is usually fastest—and give contact information so the agency can reach you with questions or a fee estimate. If you qualify for a fee waiver or reduction (for example, as news media or for an educational or public-interest purpose), request it expressly. Mark "FOIA Request" on both the envelope and the top of the letter, and use the agency's preferred submission method—many now use online portals—so your request is logged correctly. Free request generators maintained by FOIA-focused organizations can help identify the correct office and produce a compliant letter.

Step four: follow up persistently. Agencies must respond within twenty business days under 5 U.S.C. § 552(a)(6)(A)(i), but extensions are common and delay is the norm. If the deadline passes without a substantive response, follow up promptly and keep following up. Expedited processing is available in limited circumstances—an urgency to inform the public about alleged government activity, or a threat to someone's life or safety—but is not automatic and must be requested with supporting justification.

Be prepared for partial disclosure. FOIA contains nine exemptions, codified at 5 U.S.C. § 552(b), that let agencies withhold material—among them classified national-security information, statutorily confidential information, trade secrets and confidential commercial information, inter- and intra-agency deliberative communications, unwarranted invasions of personal privacy, and certain law-enforcement records. If you believe a withholding is improper, you may appeal administratively and, if that fails, seek judicial review in federal district court. Suppose the state environmental agency releases Cardinal's permit file but redacts a process diagram as a "trade secret." Meridian can challenge that designation, since the diagram may be exactly the evidence that shows Cardinal copied Meridian's housing geometry. A redaction is an invitation to argue, not a final word.

Part Two: Formal Discovery Under the Federal Rules

The Scope of Federal Discovery: Rule 26(b)(1)

Every discovery device operates within the boundaries set by Rule 26. Under Rule 26(b)(1), a party may obtain discovery of any nonprivileged matter that is relevant to any party's claim or defense and proportional to the needs of the case. Proportionality weighs six factors: the importance of the issues at stake, the amount in controversy, the parties' relative access to relevant information, the parties' resources, the importance of the discovery in resolving the issues, and whether the burden or expense outweighs its likely benefit.

The proportionality requirement, dramatically strengthened in the 2015 amendments, is one of the most important features of modern discovery practice. Those amendments did two things worth pausing on. First, they moved proportionality into the definition of the scope of discovery itself—so disproportionate material is now simply outside the boundary of what is discoverable, not merely something a court might limit on motion. Second, they deleted the old, endlessly recited phrase that discovery need only be "reasonably calculated to lead to the discovery of admissible evidence"—a formulation lawyers had pressed into service as a near-limitless mandate. It is gone, and a practitioner who still recites it in a 2026 objection is quoting a rule that no longer exists, as courts notice.

Proportionality obliges requesting parties, responding parties, and courts alike to calibrate discovery to the genuine needs of the case rather than deploy it as a scorched-earth tactic. The factors are nominally co-equal, but in practice courts often treat cost and burden as a threshold matter. In Nece v. Quicken Loans, Inc., the court found a request for some three million emails disproportionate at the class-certification stage where statutory damages were capped at $500 per violation. Conversely, courts allow broader discovery where important public issues are at stake—a principle rooted in the Advisory Committee's observation that some cases "may have importance far beyond the monetary amount involved," applied in Oxbow Carbon & Minerals LLC v. Union Pacific Railroad Co., 322 F.R.D. 1 (D.D.C. 2017).

Crucially, the burden of demonstrating disproportionality falls on the party resisting discovery, which is best positioned to know what compliance would cost. State Farm Mutual Automobile Insurance Co. v. Pointe Physical Therapy, LLC, 255 F. Supp. 3d 700, 705 (E.D. Mich. 2017). And a conclusory invocation fails: an objecting party "must submit affidavits or other evidence to substantiate its objections" rather than rest on adjectives. Cratty v. City of Wyandotte (E.D. Mich. 2017). Courts are notably unsympathetic to large organizations complaining of burden given their "sophisticated access to data," Labrier v. State Farm Fire & Casualty Co., 314 F.R.D. 637, 643 (W.D. Mo. 2016), and reject the excuse that a party's own disorganized recordkeeping should shrink its obligations—"[d]iscovery and its costs are neither a shield to ward off nor hammer to throttle the opposing party," Goes International, AB v. Dodur Ltd. (N.D. Cal. 2016).

For a defendant like Cardinal, proportionality is a shield against a sprawling document demand; for Meridian, a sword against a smaller opponent's refusal to search its own systems. The same standard cuts whichever way the facts point. We treat the mechanics of asserting and answering these objections in our companion guide to mastering document discovery: objecting and responding to requests for production.

The "Stop and Think" Rule: Rule 26(g)

Lurking behind every request, response, and objection is Rule 26(g), the discovery analog of Rule 11. It requires at least one attorney of record to sign every disclosure, request, response, and objection, certifying—after a reasonable inquiry—that the paper is consistent with the rules and warranted by existing law, not interposed for any improper purpose, and neither unreasonable nor unduly burdensome or expensive given the needs of the case. That last clause is, in effect, a proportionality certification baked into your signature. This "stop and think" rule has real teeth: where a violation occurs without substantial justification, sanctions on the signing attorney, the party, or both are mandatory. Courts have invoked it to punish boilerplate objections, Liguria Foods, Inc. v. Griffith Laboratories, Inc., 320 F.R.D. 168 (N.D. Iowa 2017); document dumps; abusive volumes of requests; and counsel who blindly accept a client's bare "we have no documents" without verifying the search, Brown v. Tellermate Holdings Ltd. (S.D. Ohio 2014). An attorney cannot outsource the integrity of discovery to the client and then disclaim responsibility. You sign; you own it.

Initial Disclosures: Rule 26(a)

Rule 26 also imposes mandatory initial disclosures—the one place in federal practice where a party must hand over information without anyone asking for it. Under Rule 26(a)(1), each party must disclose four things: each individual likely to have discoverable information the party may use to support its claims or defenses, with contact information; a copy or description by category and location of the documents, ESI, and tangible things it may use; a computation of each category of damages claimed, with supporting materials made available; and any insurance agreement that might satisfy or indemnify a judgment. These are generally due at or within fourteen days after the Rule 26(f) conference, made on then-available information—a party cannot duck the obligation by claiming its investigation is incomplete or that an opponent has not yet disclosed.

Two later tiers complete the structure. Rule 26(a)(2) governs expert disclosures: each testifying expert's identity and, for retained experts, a signed report with all opinions and their bases, the facts and data considered, supporting exhibits, qualifications and publications from the prior ten years, prior testimony from the past four years, and compensation. Rule 26(a)(3) governs pretrial disclosures—trial witnesses, deposition testimony, and exhibits—generally due at least thirty days before trial.

The disclosure regime matters far beyond its modest deadline, because Rule 37(c)(1) gives it self-executing teeth: a party that fails to disclose or supplement is automatically barred from using the withheld witness or information unless the failure was substantially justified or harmless. We unpack that preclusion sanction—one of the most lethal in discovery—in our guide to Rule 37(c)(1) preclusion of undisclosed evidence. The lesson is blunt: the witness you forgot to disclose and the damages theory you never computed may simply vanish from your case.

The 2025 Amendments: Privilege Logging Moves to the Front

That Rule 26(f) conference deserves special attention now, because the December 1, 2025 amendments to Rules 16 and 26 changed what must happen there. Rule 26(f)(3)(D) now requires the parties' discovery plan to state their views on the method and timing for complying with Rule 26(b)(5)(A)—which governs how privilege claims are asserted and described—and Rule 16(b)(3)(B)(iv) directs the court to address that timing and method in its scheduling order. In plain terms: privilege-log practice, long treated as an end-of-discovery chore, is now a front-of-case obligation.

The amendments also nudge parties toward two protective devices experienced practitioners already favor. The first is categorical logging—describing privileged material by defined categories rather than document-by-document. The second is a non-waiver order under Federal Rule of Evidence 502(d), under which production of privileged material does not waive the privilege in that or any other federal or state proceeding. A 502(d) order is the single strongest clawback protection available: unlike the safe harbor of FRE 502(b), it does not depend on proving the disclosure was inadvertent or that you took reasonable precautions, and it travels with the documents into every future case. The upshot for Meridian and Cardinal: at their very first conference, they should agree on how privilege will be logged and ask the court to enter a 502(d) order—saving both sides tens of thousands of dollars in later motion practice, and saving a privilege that a careless production would otherwise destroy.

A word on sequence: discovery generally cannot begin until after the Rule 26(f) conference, except that Rule 34 document requests may be delivered earlier (as early as twenty-one days after a defendant is served), though deemed served at the conference. After the conference, the methods may be used in any sequence.

Interrogatories: Rule 33

Interrogatories are written questions one party serves on another, answered in writing and under oath. They do the essential early work of establishing foundational facts about the parties, witnesses, claims, defenses, and damages—the names, dates, and document locations that map the terrain before the expensive depositions begin.

Under Rule 33(a)(1), a party may serve no more than twenty-five interrogatories—including all discrete subparts—on any other party without a stipulation or leave of court. That phrase is the battleground. The cap is routinely, and improperly, evaded through compound questions that pack several independent inquiries into one numbered item; courts count each discrete subpart, so an interrogatory crammed with unrelated sub-questions can be counted as several. The test is functional: subparts "logically or factually subsumed within and necessarily related to the primary question" count as one, while subparts introducing a genuinely new line of inquiry count separately. Keep each interrogatory tied to a single subject, and remember you need not fire all twenty-five at once—serve them in sequentially numbered sets (many districts require the numbering to run continuously across sets).

Responses are due within thirty days. Each interrogatory must be answered separately and fully, and the answers must be verified—signed under oath, under penalty of perjury, by the party, not merely by counsel (Rule 33(b)(3), (b)(5)). This is not a formality. Interrogatory answers are admissible and can be used against the responding party at trial; they are, in a real sense, sworn testimony delivered in writing with time to think. Answers should be accurate, carefully drafted, and no broader than the question requires: answer the question asked, completely, without volunteering more.

A party may object instead of answering, but the objection must be specific and stated with particularity; if only part of an interrogatory is objectionable, the party must answer the rest. Often the better course is to both object and answer—preserving the objection while responding to the extent the question is proper—rather than lodge a bare objection that simply invites a motion to compel. And the duty does not end with the first response: under Rule 26(e), a party who learns that an answer was incomplete or incorrect in a material respect must supplement in a timely way, on pain of having the evidence excluded at trial under Rule 37(c)(1).

Two special features of Rule 33 deserve mention. First, contention interrogatories—which ask a party to state the facts or legal theories supporting a specific contention in its pleadings—are expressly permitted by Rule 33(a)(2): an interrogatory is not objectionable merely because it asks for an opinion or contention relating to fact or the application of law to fact. Capacchione v. Charlotte-Mecklenburg Schools, 182 F.R.D. 486 (W.D.N.C. 1998). A court may defer such interrogatories until later in discovery, but deferral is a question of timing, not a license to refuse.

Second, Rule 33(d) offers an escape hatch when the answer can be derived from business records and the burden is substantially the same for either party: the responding party may produce the records and let the requesting party do the arithmetic—a genuine convenience, but one that cannot be abused to bury a simple answer in a haystack of files.

When drafting, keep each question focused on a single fact, and define key terms so the responding party cannot adopt a strategically narrow reading. Meridian might ask Cardinal to "identify each person who accessed, copied, or modified any file containing Meridian's pump-housing designs, stating for each the date and means of access"—precise, single-purpose, and hard to dodge. Contrast a sloppy version—"describe everything Cardinal knows about the designs and how it got them and who was involved"—which is compound, vague, and an open invitation to evasion. For a deeper treatment, see our guide to mastering interrogatories: objecting and responding in civil litigation.

Requests for Production of Documents: Rule 34

Requests for production—RFPs, or document requests—are among the most consequential tools in civil litigation, because documents do not forget, do not get nervous, and do not change their story on cross. They require the responding party to produce documents, ESI, and tangible things within its possession, custody, or control that fall within the request's scope, and may also demand entry onto property for inspection (Rule 34(a)(2)). "Control" is a term of art and a frequent fight: a party controls a document not only when it physically holds it but when it has the legal right to obtain it on demand—from a corporate affiliate, an agent, a custodian. That a sister company or vendor physically holds the file does not necessarily put it beyond a party's control.

Responses are due within thirty days (Rule 34(b)(2)(A)). The responding party may produce everything responsive, withhold under stated objections, or state that responsive documents do not exist or are not within its control. Three procedural requirements sharpened by the 2015 amendments now structure the response and largely killed the old game of objecting in the dark:

First, objections must be stated with specificity (Rule 34(b)(2)(B)); reflexive boilerplate without explanation of how the labels apply to the particular request is improper. Second, and most consequentially, Rule 34(b)(2)(C) now requires the responding party to state whether any responsive materials are being withheld on the basis of an objection. An opponent can no longer object and quietly hold everything back—the rule forces the producing party to show its cards, so the requesting party knows whether an objection hides a single trivial document or a smoking gun. Third, if production is to occur, the response must state a reasonable time for it; rolling productions are permissible if specified.

When documents are produced, Rule 34(b)(2)(E) requires that they be produced either as kept in the ordinary course of business or organized and labeled to correspond to the request's categories. A producing party may neither dump an undifferentiated mass of files to bury the key document, nor cherry-pick a favorable subset while silently withholding the rest. Anything withheld for privilege or work product must appear on a privilege log described with enough particularity to let the requesting party assess the claim (Rule 26(b)(5)(A)).

The recurring mistakes cut across practice areas: answering partially without flagging the incompleteness; objecting without citing supporting authority; relying on the client to gather documents without independent attorney review; over-producing rather than reviewing and redacting; producing a disorganized jumble; failing to identify and collect ESI; and preservation errors that lose material before it can be produced. When Cardinal responds to Meridian's request for "all communications referring to the pump-housing designs," it cannot simply hand over a folder of emails and leave Dana's text messages, Slack channels, and personal-cloud files unaddressed—if those sources hold responsive material, Cardinal must search them, produce what is responsive, and—under Rule 34(b)(2)(C)—account for anything it withholds.

Requests for Admission: Rule 36

Requests for admission ask the opposing party to admit or deny specific factual statements. They are unlike every other discovery device: not designed to gather information but to eliminate it as a subject of dispute. Every fact admitted is a fact no one has to prove at trial, and a well-built set can shrink a sprawling case to its genuine points of contention—and authenticate a stack of documents along the way. They are powerful and chronically underused.

Under Rule 36(a)(3), a matter is deemed admitted if the responding party fails to serve a written answer or objection within thirty days—a self-executing default that can be catastrophic if a deadline slips. The requesting party need not even move; the admission happens automatically. Lauter v. Rosenblatt (C.D. Cal. 2019). And under Rule 36(b) a matter admitted is conclusively established unless the court permits withdrawal or amendment—not mere evidence the jury weighs but a fact removed from the case entirely. (Withdrawal is available when it would promote the merits and the requesting party would not be prejudiced, but it is discretionary and never to be relied upon.) An admission binds only the party who made it, not co-parties. Becerra v. Asher, 921 F. Supp. 1538 (S.D. Tex. 1996).

Each request should be a single, simple, unambiguous factual statement; Rule 36(a)(2) requires a separate request for each matter, and compound or argumentative requests that smuggle in legal conclusions invite evasion. A useful discipline, especially for authentication, is to keep separate the request to admit a document's authenticity, the request to admit a signature, and the request to admit the truth of a fact stated in the document—three distinct requests, not one. Doe v. Mercy Health Corp. (E.D. Pa. 1993). The art is to draft statements the opponent cannot honestly deny—Meridian might ask Cardinal to "admit that Dana Reyes signed the Employee Confidentiality Agreement attached as Exhibit A on or about June 3, 2023."

When responding, a party must admit, deny, or object—and may plead lack of sufficient information only after a reasonable inquiry, and only when its information is genuinely insufficient (Rule 36(a)(4)). Lack of knowledge is not a dodge for a party that simply prefers not to look. And there is a sting in the tail: under Rule 37(c)(2), if a party denies a request and the requesting party later proves it true, the court must order the denying party to pay the reasonable expenses, including attorney's fees, of making that proof, absent a narrow exception. So if Cardinal flatly denies that Dana signed the agreement, and Meridian proves the signature at trial, Cardinal can be made to pay for the proof. We develop the offensive and defensive strategy of admissions in our guide to strategic responses to requests for admission.

Depositions: Rule 30

Depositions—oral examinations under oath, taken outside court and recorded by a court reporter and, increasingly, by video—are the most versatile and powerful discovery tool. They let counsel explore a witness's knowledge in real time, lock in testimony for trial, gauge credibility and demeanor, and surface information no document captures. They are also the most expensive phase of discovery, and where professional conduct is most visibly tested.

Under Rule 30(b)(1), a party must give reasonable written notice to every other party, stating the deponent's name and the time and place. By default a party may take only one deposition of any given deponent and must seek leave for a second (Rule 30(a)(2)(A)(ii)), and each is limited to one day of seven hours unless the parties stipulate or the court orders otherwise (Rule 30(d)(1)). A practical subtlety: the seven-hour clock runs only against time on the record—breaks and lunch do not count—so defending counsel should track the elapsed on-record time.

If the deponent is a party, notice alone compels attendance; if a non-party, a subpoena under Rule 45 is required, and a subpoena duces tecum if you want the witness to bring documents. For a party deponent, coupling a Rule 34 request with the notice is the more reliable practice.

Corporate depositions under Rule 30(b)(6)

When the witness is an organization rather than a person, Rule 30(b)(6) governs, and it is one of the most strategically important devices in the rules. The noticing party names the entity and "describe[s] with reasonable particularity the matters for examination"; the organization must then designate people to testify about information known or reasonably available to the organization. The designee testifies not to personal knowledge but to the institution's, and the organization must prepare the witness—reviewing documents and consulting employees—to speak to the noticed topics. A poorly prepared designee who repeatedly answers "I don't know" can expose the organization to a motion to compel a better witness and to sanctions. A December 1, 2020 amendment requires the parties to confer in good faith about the matters for examination. For Cardinal, a 30(b)(6) notice on a topic like "Cardinal's document-retention practices for engineering files" forces the company itself, not just a convenient individual, to commit to an answer.

Conduct, objections, and instructions not to answer

Under Rule 30(c)(2), objections must be stated "concisely in a nonargumentative and nonsuggestive manner," and the examination proceeds—the witness still answers, subject to the objection. The reason is structural: most deposition objections (to the form of a question—leading, compound, vague, assumes facts not in evidence, asked and answered) are waived if not raised on the spot, because they can be cured by rephrasing; objections to relevance, hearsay, or prejudice are generally preserved for trial. So the only objections defending counsel should make are the form objections that must be made to be preserved, stated in two or three words.

What defending counsel may not do is coach. The speaking objection—the argumentative objection that telegraphs the desired answer ("objection, vague—but if you remember a specific date, you can say so")—is prohibited and sanctionable under Rule 30(d)(2). Specht v. Google, Inc., 268 F.R.D. 596 (N.D. Ill. 2010). Even the reflexive "if you know" has been condemned as coaching. The genuine tool for protecting a witness is not the speaking objection but thorough preparation.

Counsel may instruct a witness not to answer in only three situations under Rule 30(c)(2): to preserve a privilege, to enforce a court-ordered limitation, or to present a motion under Rule 30(d)(3) to terminate a deposition conducted in bad faith or to harass. A relevance objection is emphatically not a basis to instruct a witness to stay silent. Brincko v. Rio Properties, Inc., 278 F.R.D. 576 (D. Nev. 2011) ("there are very few circumstances in which an instruction not to answer a deposition question is appropriate"). And if counsel invokes the bad-faith ground, the rule requires action: counsel must state the basis on the record and promptly move to terminate or limit—sitting on the objection without moving is itself sanctionable. Redwood v. Dobson, 476 F.3d 462 (7th Cir. 2007).

Review and signature: the errata sheet

If the witness or a party requested review before the deposition concluded, Rule 30(e) gives the deponent thirty days—running from the court reporter's notification that the transcript is available, not from physical receipt—to submit changes "in form or substance" on a statement of changes, the errata sheet, with a reason for each. EBC, Inc. v. Clark Building Systems, Inc., 618 F.3d 253 (3d Cir. 2010). Courts split on how far a witness may go: some permit any change, even one that contradicts the original testimony (which remains in the record for impeachment), Podell v. Citicorp Diners Club, Inc., 112 F.3d 98 (2d Cir. 1997); others hold a deposition "is not a take home examination" and limit the errata to corrective rather than contradictory changes, Garcia v. Pueblo Country Club, 299 F.3d 1233 (10th Cir. 2002). Know your circuit before drafting an errata sheet that tries to walk back a damaging admission.

When Meridian deposes Dana, its lawyer will walk her methodically through which design files she accessed in her final weeks, when, and from what device, locking in those answers under oath; Dana's counsel may lodge a crisp form objection but cannot instruct her to stay silent simply because the answers hurt. The craft of taking and defending depositions is explored in our companion pieces on the art and science of depositions in federal civil litigation and the art of defending depositions in federal court.

Part Three: Discovery Objections—Proper Use and Abuse

The Problem of Boilerplate Objections

The most pervasive and damaging practice in civil discovery is the reflexive boilerplate objection—stock phrases copied wholesale into responses without any analysis of whether they apply. General objections, the laundry list lodged collectively against every request, are universally disfavored and often treated as legally ineffective—a Rule 26(g) certification boilerplate cannot honestly carry.

Few opinions say this more bluntly than Judge Andrew Peck's in Fischer v. Forrest, No. 14-cv-1304 (S.D.N.Y. Feb. 28, 2017), which warned that any lawyer still using the old "reasonably calculated" language and unexplained general objections "has not read the 2015 amendments," and struck a set of stock objections wholesale. The point was echoed in Liguria Foods, Inc. v. Griffith Laboratories, Inc., 320 F.R.D. 168 (N.D. Iowa 2017), where Judge Mark Bennett issued an order to show cause against both sides' counsel—experienced lawyers at fine firms—for trading boilerplate objections that had become so habitual they no longer noticed. The objections most often abused—vague, ambiguous, argumentative, overbroad, unduly burdensome—are each legitimate when they genuinely describe a defect in a specific request, and meaningless when fired reflexively. Repeated boilerplate also costs counsel credibility: a judge who has watched you cry "unduly burdensome" against an obviously reasonable request will discount the cry when the burden is real.

Calls for a Legal Conclusion

The "calls for a legal conclusion" objection is similarly overused. As noted, Rule 33(a)(2) provides that an interrogatory is not objectionable merely because it asks for an opinion or contention relating to fact or the application of law to fact. Contention interrogatories are entirely proper; a court may defer such an interrogatory to a later stage, but that is timing, not a basis to refuse the question.

Cost and Burden Objections

Objections based on cost or burden are valid when genuinely applicable, but they must be substantiated with specifics—the volume of documents, the cost of collection and review, the architecture of the systems—and the objecting party bears that burden with affidavits or other evidence. A bare "this is burdensome" is no better than any other boilerplate. Where the burden is real, the better practice is to propose alternatives rather than refuse: offer inspection at a storage site, propose cost-sharing, or suggest the requesting party bear the cost of especially voluminous records. If Meridian demands every email Cardinal sent over six years, Cardinal can legitimately object—but it should pair the objection with a concrete counterproposal, such as agreed search terms applied to a defined custodian list over the relevant eighteen-month period, supported by an affidavit on what the unrestricted search would cost.

Other Objections: What Works and What Doesn't

Several other objections recur, each with a narrow zone of legitimacy and a wide zone of abuse. The Fifth Amendment privilege against self-incrimination is valid where an answer would expose a party to criminal liability, but must be asserted question by question (and in civil cases, unlike criminal ones, a court may draw an adverse inference from the invocation). Objections that a request is harassing, oppressive, or embarrassing are valid when a request is truly designed to harass, but courts overrule them where the request serves a legitimate purpose. The objection that information is "equally available" holds only when genuinely true, and is frequently asserted in bad faith by the party better positioned to find the material. Finally, many districts impose specific objection requirements by local rule, and failing to comply can waive objections or draw sanctions. The boilerplate objection cut-and-pasted from a brief in another district is not just lazy; it may be affirmatively wrong.

Part Four: Privilege, Work Product, and the Privilege Log

Discovery's reach stops at the privileged. Two doctrines do most of the work: the attorney-client privilege and the work-product protection. They are the most legitimate grounds for withholding material—and, handled carelessly, among the most easily forfeited; a single botched production can hand an adversary the very advice you gave your client.

The Attorney-Client Privilege

Under federal common law (governing federal-question cases, while state privilege law governs diversity cases under FRE 501), the attorney-client privilege protects communications that are (1) between a client and counsel, or their agents whose participation does not destroy confidentiality, (2) made and intended to be confidential, and (3) made to obtain or provide legal advice. Because the privilege withholds relevant evidence, courts construe it narrowly, applying it only where it serves its purpose of encouraging candid client communication. Fisher v. United States, 425 U.S. 391, 403 (1976).

The privilege is assessed document-by-document, and it cannot be conjured by ritual. You do not create privilege by stamping "privileged and confidential" across a business memo, by copying a lawyer on an email that seeks no legal advice, or by routing an ordinary commercial document through outside counsel. Neuder v. Battelle Pacific Northwest National Laboratory, 194 F.R.D. 289 (D.D.C. 2000). The communication must actually be for legal advice. This is why mixed business-and-legal communications—the daily bread of in-house counsel—are so heavily litigated; the privilege attaches only where the primary purpose is legal.

For corporate clients, the governing question is whose communications with counsel are privileged. The Supreme Court answered it in Upjohn Co. v. United States, 449 U.S. 383 (1981), rejecting the narrow "control group" test (which limited the privilege to senior management) and holding that the privilege protects communications between counsel and employees throughout the hierarchy when made at the direction of corporate superiors to secure legal advice on matters within the employees' duties. Upjohn is why a properly structured internal investigation—the kind Meridian's general counsel likely launched the moment Dana was suspected—can shield interviews of rank-and-file employees. But the protection is not automatic: the investigation must be conducted for the purpose of legal advice, with appropriate Upjohn warnings, and documented as such. One framed as routine business fact-finding may yield no privilege at all, and Cardinal will surely argue Meridian's "investigation" was ordinary management activity in lawyer's clothing. The cleaner the legal framing at the outset, the stronger the privilege later.

The Work-Product Doctrine

The work-product protection traces to Hickman v. Taylor, 329 U.S. 495 (1947), and is now codified in Rule 26(b)(3). It shields documents and tangible things prepared in anticipation of litigation or for trial by or for a party or its representative, in two grades. Ordinary (or fact) work product—witness interviews, factual summaries—is discoverable only on a showing of substantial need and that the party cannot obtain the substantial equivalent without undue hardship; the party seeking it bears that burden. Opinion work product—the attorney's mental impressions, conclusions, and legal theories—receives near-absolute protection almost never overcome. That distinction is why a litigator can sometimes pry loose an investigator's factual notes but essentially never the lawyer's strategic analysis.

Waiver: How Privilege Is Lost

Privilege is fragile and not self-executing—it must be affirmatively asserted or it is forfeited. The most common waiver traps are voluntary disclosure to a third party outside the privilege, the "at issue" waiver that arises when a party puts its counsel's advice in play (for instance, an advice-of-counsel defense), and inadvertent production. On inadvertent production, FRE 502(b) provides a safe harbor: disclosure does not waive the privilege if it was inadvertent, the holder took reasonable steps to prevent it, and promptly took reasonable steps to rectify it—including by invoking the clawback procedure of Rule 26(b)(5)(B), which obligates a party notified of an inadvertent production to promptly return, sequester, or destroy the material. FRE 502(a) limits subject-matter waiver—the doctrine that disclosing one privileged communication can waive privilege over all communications on the same subject—to intentional disclosures where fairness requires the disclosed and undisclosed material be considered together.

The strongest protection of all, as noted earlier, is a court order under FRE 502(d), which renders production non-waiving in that and every other proceeding, without any showing of inadvertence. In a case like Meridian's, where hundreds of thousands of documents change hands and a few privileged ones inevitably slip through, a 502(d) order is the difference between a recoverable mistake and a catastrophe—precisely why the 2025 amendments push parties to address it at the outset.

The Privilege Log

When a party withholds material on privilege or work-product grounds, Rule 26(b)(5)(A) requires it to expressly claim the protection and describe the withheld documents "in a manner that, without revealing information itself privileged or protected, will enable other parties to assess the claim"—in practice, a privilege log. A bare objection that a request "calls for privileged information" is not enough. Burlington Northern & Santa Fe Railway Co. v. United States District Court, 408 F.3d 1142 (9th Cir. 2005).

A traditional document-by-document log identifies, for each withheld item, the document type, date, author, all recipients (by name and role—not "et al."), a subject-matter description sufficient to assess the claim, and the specific privilege asserted. The description is where logs live or die. Merely parroting the legal standard ("document reflecting legal advice") is inadequate and risks waiver, Johnson v. Ford Motor Co., 309 F.R.D. 226 (S.D. W. Va. 2015); but a description so detailed it discloses the actual advice can itself waive the privilege, Veolia Water Solutions v. Siemens Industry, Inc., 63 F. Supp. 3d 558 (E.D.N.C. 2014). The art is to say enough and no more—"Memorandum from in-house counsel to CEO analyzing the company's legal obligations under new FDA regulations" hits the mark.

The risks are concrete and severe. An inadequate log can waive the privilege it was meant to protect, In re Grand Jury Subpoena, 274 F.3d 563 (1st Cir. 2001), and so can an untimely one, though most courts apply the multi-factor Burlington Northern balancing test—weighing the adequacy of the information, the timeliness, the volume, and the difficulty of the task—rather than a per se rule. And in a trap that catches the unwary, a privilege log is itself a discovery response that must be signed by an attorney under Rule 26(g); serving an unsigned log has been held to waive the very protections it asserts, Rhodes v. Ingram (E.D.N.C. 2015). Logs auto-generated by a review platform, without human-supplied descriptions, are routinely held insufficient.

Given the volume of modern ESI, categorical logging—grouping withheld documents into defined categories rather than listing each one—has become increasingly accepted, and is exactly the arrangement the 2025 amendments invite parties to negotiate at the Rule 26(f) conference. Many districts now permit it where a document-by-document log would be unduly burdensome and the category descriptions still allow the opponent to assess the claims, though some jurisdictions (the Seventh Circuit's district courts among them) traditionally insist on individualized logs. For Meridian and Cardinal, agreeing up front to log "all post-suit communications between Cardinal and its litigation counsel" as a single category can spare both sides a fortune in fees.

Part Five: Subpoenas and Third-Party Discovery

Rule 45 governs subpoenas used to compel testimony and document production from non-parties—and in many cases the decisive evidence lives not with the opponent but with a third party who has no stake in the fight. A properly served subpoena commands a non-party to appear for deposition, produce documents, or permit inspection; failure to comply without adequate excuse exposes the recipient to contempt under Rule 45(g). But the rule also carries unusual protections for non-parties, who did not choose to be drawn into someone else's lawsuit.

Geographic limits. A subpoena may command attendance only within 100 miles of where the person resides, is employed, or regularly transacts business in person, measured as a straight line rather than by driving route (Rule 45(c); Universitas Education, LLC v. Nova Group, Inc. (S.D.N.Y. 2013)). The same limit governs document production.

Objecting to a document subpoena. A recipient of a document subpoena may serve written objections before the earlier of fourteen days after service or the compliance date (Rule 45(d)(2)(B)). Timely objections suspend the obligation to produce until the issuing party obtains a compelling order. But this route works only for document subpoenas: a recipient resisting a testimonial subpoena cannot merely object—it must move to quash or modify under Rule 45(d)(3) or seek a protective order under Rule 26(c). Failing to object on time risks waiving all objections.

Moving to quash. Under Rule 45(d)(3)(A), the court must quash or modify a subpoena that fails to allow reasonable time to comply, requires travel beyond the geographic limits, requires disclosure of privileged matter, or subjects a person to undue burden; under Rule 45(d)(3)(B), it may quash one requiring disclosure of trade secrets or an unretained expert's opinion. The recipient bears the burden, ATS Products, Inc. v. Champion Fiberglass, Inc., 309 F.R.D. 527 (N.D. Cal. 2015), and the motion is filed in the court for the place of compliance.

The duty to protect non-parties—and cost-shifting. Rule 45(d)(1) affirmatively requires the issuing party to "take reasonable steps to avoid imposing undue burden or expense" on a non-party, and authorizes sanctions—including lost earnings and attorney's fees—for violations. One important consequence is cost-shifting: several circuits hold that when compliance imposes significant expense, Rule 45(d)(2)(B)(ii) makes shifting that cost to the requesting party mandatory. Legal Voice v. Stormans Inc., 738 F.3d 1178 (9th Cir. 2013) (treating roughly $20,000 as significant); Leonard v. Martin, 38 F.4th 481 (5th Cir. 2022). Courts typically shift a fraction—half or more—rather than parsing each line item, and a burden is often deemed undue where the same documents could be obtained from a party. Haworth, Inc. v. Herman Miller, Inc., 998 F.2d 975 (Fed. Cir. 1993).

The sensible sequence for responding to a subpoena: identify every potential objection to its scope, manner of compliance, or burden; raise them through the right vehicle (written objections for documents, motion to quash, or protective order); and comply with whatever is not subject to a valid objection—while continuing to preserve everything responsive until the matter is resolved. Litigators should also think about subpoenas proactively, mapping early which non-party witnesses and custodians hold relevant information.

Third-party discovery is often where the decisive evidence lives. In our hypothetical, the most damaging material may not be in Cardinal's files at all but in Dana's personal cloud-storage account, on a former contractor's laptop, or in a recruiter's emails. Meridian can subpoena the cloud provider for access logs and stored files—subject to the provider's objections and to the Stored Communications Act, which restricts what a provider may disclose about the contents of a user's account (often forcing the requesting party to obtain the material from the account holder directly). Because so much modern evidence lives in personal devices, cloud accounts, and messaging apps, the preservation and collection of that material raises issues we take up in trade secrets in the age of remote work and cloud computing. And when a non-party witness or defendant sits abroad, service and compulsion become far more complex—a problem explored in our guide to serving legal documents to foreign defendants in cross-border IP litigation.

Part Six: Documenting Discovery Deficiencies and the Motion to Compel

When an opponent's responses fall short—incomplete interrogatory answers, deficient productions, improper objections, missing disclosures—the right first move is not a reflexive motion to compel. It is to document the deficiencies thoroughly and then attempt informal resolution. Rule 37(a)(1) requires it: a motion to compel must include a certification that the movant conferred or attempted in good faith to confer before filing. Most districts reinforce this with local rules mandating a meet-and-confer, and many judges require a pre-motion conference. Skip it and your motion gets bounced on procedure before the court ever reaches the merits.

The workhorse here is the discovery deficiency memo: an internal document, often prepared by a paralegal or associate and reviewed by supervising counsel, that walks through each of the opponent's responses and identifies, with specificity, what was requested, what was produced or answered, and what remains outstanding. It ensures nothing is overlooked, supplies the factual basis for a meet-and-confer letter, and—if informal efforts fail—becomes the backbone of the motion.

A concrete illustration shows the value. Suppose Meridian's Request No. 14 sought "all documents reflecting Cardinal's income, expenses, assets, and liabilities for fiscal years 2023 and 2024," and Cardinal produced only summary income statements while omitting tax returns, balance sheets, and profit-and-loss detail. The memo records the request number, the full scope sought, what Cardinal produced, and a precise statement of what is missing—and, replicated across every deficient response, produces a clean record that converts directly into a meet-and-confer letter demanding supplementation by a stated deadline, and, if that fails, into the motion itself.

The fee consequences are worth understanding before you file, because Rule 37(a)(5) cuts both ways and is mandatory in direction. If the motion is granted (or the discovery produced after it is filed), the court must order the losing party, its attorney, or both to pay the movant's reasonable expenses including attorney's fees—unless the movant filed without first conferring in good faith, the opposing position was substantially justified, or other circumstances make an award unjust. But if the motion is denied, the rule flips: the court must order the movant to pay the opponent's expenses, subject to the same exceptions. A poorly considered motion to compel is therefore not just a waste of effort; it is an affirmative fee risk. The deficiency memo protects counsel here too: it documents the good-faith effort and shows the motion is targeted rather than reflexive.

Part Seven: Document Review, Redaction, and Production Oversight

The Two-Step Redaction Process

When a production contains privileged, confidential, or otherwise protected information, proper redaction is a two-step process. First, the information must be accurately identified throughout the production set—careful review of every page, not a spot check. Second, the redactions must be applied so recipients genuinely cannot recover the hidden content by visual means or digital manipulation.

That second point is where modern productions go wrong. A black box drawn over text in a PDF, without removing the underlying text, can be defeated by anyone who copies the "redacted" passage or inspects the file's data—a failure that has embarrassed sophisticated parties in high-profile filings, exposing everything from settlement figures to the names of confidential witnesses. Courts have sanctioned parties whose ineffective redactions exposed privileged communications. Proper redaction requires tools that permanently remove the underlying data, not merely obscure it. If Meridian produces Dana's personnel file with her Social Security number "redacted" by a box that copy-paste reveals, it has not protected that information—it has disclosed it while creating a false sense of compliance.

The Recurring Production Mistakes

A consistent set of production errors recurs. Ineffective redaction is the most technically dangerous, because it disguises disclosure as compliance. Extra or orphaned files—stray attachments, embedded objects, system files—slip into bulk-collected ESI and, if not culled, can disclose sensitive material. Producing too much or too little reflects a failure to review: over-production exposes privileged documents and can support a waiver argument, while under-production without proper objection is sanctionable. Failure to communicate compounds all of these; if a production will be late or documents are withheld, say so, because most contested motions could have been avoided by a timely call. Disorganized production violates Rule 34's organization requirement. Missing certifications cause trial problems, because third-party records usually must be certified to be admissible; request certified copies at the outset. And procrastination is the most preventable error of all: a production of thousands of pages, ESI in multiple formats, privilege review, and redaction cannot be assembled the night before it is due.

A note on authentication, because it bridges discovery and trial. Documents gathered in discovery still must be admissible, and electronic evidence in particular—website captures, social-media posts, screenshots—carries its own authentication requirements. If Meridian intends to use a screenshot of Cardinal's product page as evidence of the copied housing, it should preserve and authenticate that capture properly rather than assume a printout will be admitted—a problem we address in capturing the web: authenticating website screenshots as evidence in federal court.

Part Eight: Electronic Discovery

The EDRM Framework

Electronic discovery—the identification, preservation, collection, review, and production of electronically stored information—is now the dominant form of document discovery. Virtually every modern case involves significant volumes of ESI: emails, databases, text messages, chat platforms, social-media content, ephemeral-messaging apps, and cloud-stored material. Mishandling ESI is among the most fertile grounds for sanctions. The Electronic Discovery Reference Model (EDRM) supplies the most widely accepted map, describing a workflow that runs from information governance through identification, preservation, collection, processing, review, analysis, production, and presentation. The stages overlap, but understanding each helps a practitioner spot where a case is at risk.

Information governance is the starting point: the policies by which an organization creates, stores, retains, and deletes its electronic information. How a company manages its data before any dispute determines what is available when one arises, and clients with mature governance respond faster and more cheaply. Encouraging sound governance before litigation is among the most valuable preventive counseling a lawyer can offer—a theme that connects to building disciplined internal programs, as in building a trade secret protection program from scratch.

Identification determines what ESI is potentially relevant and where it lives—reviewing the facts, interviewing the people who created or held relevant material, and mapping the data environment. For Cardinal, that means recognizing early that relevant data may sit not only in its email system but in Dana's company laptop, her personal phone, a shared engineering drive, and a chat tool the team used.

Preservation, the Legal Hold, and Zubulake

Preservation prevents the destruction or alteration of relevant ESI—the conduct the law calls spoliation. Its primary mechanism is the legal hold: a formal instruction from the legal team to the relevant custodians to suspend ordinary deletion practices for the categories of data the hold covers.

The threshold question is when the duty attaches, and the answer is earlier than many clients assume: it arises when litigation is reasonably anticipated, which can precede the filing of any complaint. Silvestri v. General Motors Corp., 271 F.3d 583 (4th Cir. 2001). Recognized triggers include deciding to sue, receiving a demand letter, the filing of an incident report, a government investigation, and internal awareness of a likely-litigated complaint—and a plaintiff who has decided to sue is usually held to have anticipated litigation before the defendant did. Apple Inc. v. Samsung Electronics Co., 888 F. Supp. 2d 976 (N.D. Cal. 2012).

The duty is exacting, and it runs first to counsel. The seminal Zubulake decisions established that counsel must not merely issue a hold but actively oversee it—communicating with key players, identifying sources of relevant data, suspending auto-deletion, and confirming that preservation is actually happening. Zubulake v. UBS Warburg LLC, 220 F.R.D. 212 (S.D.N.Y. 2003), and 229 F.R.D. 422 (S.D.N.Y. 2004). A defensible hold program assembles a team (business units, IT, records, counsel), identifies custodians and date ranges, maps data sources, suspends routine destruction and auto-deletion, issues a written hold from a figure of authority, collects acknowledgments, sends periodic reminders, and documents every step—reaching not only email but text and instant messages, voicemail, drives and external media, collaboration tools, ephemeral-messaging apps, and cloud storage. (On Pension Committee of University of Montreal Pension Plan v. Banc of America Securities, LLC, 685 F. Supp. 2d 456 (S.D.N.Y. 2010): Judge Scheindlin treated the failure to issue a written hold as gross negligence per se, but the Second Circuit rejected that bright-line rule in Chin v. Port Authority, 685 F.3d 135 (2d Cir. 2012)—so cite Pension Committee for its hold best practices, not a per se rule.)

The common preservation failures are predictable: a hold too narrow to capture all relevant custodians; reliance on employees to self-collect without verification; failure to suspend auto-deletion after the duty attaches; and neglect of mobile devices, personal email, chat apps, and cloud storage. Cardinal's gravest risk is mundane: if its email system automatically deletes messages after ninety days and no one suspends that rule when Meridian's claim becomes foreseeable, the very communications showing when and how Dana's design files arrived could vanish—and that loss, as we will see, can shift the case decisively in Meridian's favor.

Collection, Processing, Review, and TAR

Collection gathers and centralizes the preserved ESI. Methods vary, but one rule is constant: collection must not alter the metadata—creation and modification dates, authorship, file path—because that metadata may itself be relevant or may be needed to authenticate a document later. Processing prepares the data for review, using specialized software to extract files from archives, deduplicate, convert formats, and filter out system files, often applying search terms to cull the volume. Review—the most time-consuming and expensive stage by far—is the human, and increasingly machine-assisted, evaluation of the processed material for relevance, privilege, and confidentiality.

Technology-assisted review (TAR), also called predictive coding, uses machine learning trained on attorney coding decisions to prioritize and classify large document sets, and has become an accepted method for reducing the cost of large-scale review. The landmark judicial approval came from Judge Peck in Da Silva Moore v. Publicis Groupe SA, 287 F.R.D. 182 (S.D.N.Y. 2012), the first federal opinion to bless TAR. A party facing a large review—Meridian, perhaps, confronting hundreds of thousands of Cardinal emails—should ask counsel whether TAR can bring the cost within reason without sacrificing thoroughness, and expect to negotiate validation and transparency protocols with the other side.

Production Format, Analysis, and Presentation

Analysis evaluates the reviewed material for content and context—patterns, themes, custodians, key exchanges—and runs throughout discovery. Production delivers the reviewed material in the agreed format—no triviality, and a frequent, avoidable source of disputes, because neither Rule 34 nor Rule 45 dictates one. Under Rule 34(b)(2)(E), if the request does not specify a form, the producing party must produce ESI as it is ordinarily maintained or in a reasonably usable form, and need not produce it in more than one form. The practical choices fall along a spectrum:

• Native production preserves the original format and full metadata fidelity and is cheapest for the producing party, but native files cannot be Bates-stamped on their face or redacted (they must be converted to images first).
• Image (TIFF or PDF) with load files and extracted text—the litigation-platform standard—allows Bates numbering, confidentiality stamping, and redaction, and gives the requesting party full searchable text and metadata in a review platform, but it is most expensive for the producing party and renders spreadsheets awkwardly.

The right answer depends on the document types and the needs of the case, which is why the form of production, the metadata fields, and the Bates conventions should be settled in a written ESI protocol at the Rule 26(f) conference. Producing ESI in a format the other side cannot open is a common, easily avoided, self-inflicted error. Presentation, finally, is the use of the electronic evidence at hearings, depositions, and trial—where everything collected, preserved, reviewed, and produced becomes part of the evidentiary record.

ESI Management: Practical Considerations

A few practical points. First, designate a point person responsible for ESI on a given matter; a single person deeply familiar with the client's data environment will outperform a responsibility spread thinly. Second, when engaging e-discovery vendors, mind confidentiality: information shared with a vendor retained by counsel is protected only if the engagement is structured that way. Third, know your court's local rules and standing orders, many of which address ESI preservation, production format, inadvertent disclosures, and e-discovery disputes. And finally, given the December 2025 amendments, raise privilege logging and a Rule 502(d) non-waiver order at the very first conference—before a single document changes hands.

Part Nine: When It Goes Wrong—The Rule 37 Sanctions Machine

Every obligation described in this guide has a remedy behind it, and the remedy is Rule 37. Understanding the sanctions architecture is how a litigator calibrates risk, both in pressing an opponent who has misbehaved and in protecting a client from its own missteps. The architecture has distinct compartments.

Rule 37(a)—the motion to compel and fee-shifting. We covered the mechanics above. The key point is that the fee-shifting of Rule 37(a)(5) is mandatory in direction, subject to the "substantially justified" and "unjust" safety valves, and runs against the loser—movant or respondent.

Rule 37(b)—violation of a discovery order. The most fearsome compartment, reserved for a party that disobeys a court order compelling discovery. Beyond the mandatory fee award, the court may impose the escalating menu of Rule 37(b)(2)(A): deeming facts established, prohibiting the party from supporting or opposing claims or introducing evidence, striking pleadings, staying the proceedings, dismissing the action, rendering a default judgment, or treating the disobedience as contempt. Courts impose these case-ending sanctions on parties who flout orders. Red Wolf Energy Trading, LLC v. Bia Capital Management, LLC (D. Mass. 2022) (default judgment for repeated violations).

Rule 37(c)—failure to disclose or admit. Rule 37(c)(1) is the self-executing exclusion sanction for failures to disclose or supplement under Rule 26(a) and (e): the withheld witness or information cannot be used unless the failure was substantially justified or harmless, and the non-disclosing party bears that burden. Because exclusion can be effectively case-dispositive, this is among the most lethal provisions in the rules, treated at length in our guide to Rule 37(c)(1) preclusion of undisclosed evidence. Rule 37(c)(2), as noted, requires a party that unjustifiably denied a Rule 36 request later proven true to pay the cost of proof.

Rule 37(d)—total failure to respond. When a party simply fails to appear for its own deposition, serve interrogatory answers, or respond to a Rule 34 request at all, Rule 37(d) authorizes immediate sanctions—without a prior motion to compel. A complete no-show is treated more harshly than a deficient response, which routes back through Rule 37(a). Rule 37(f) reaches a party that fails in bad faith to participate in framing the Rule 26(f) discovery plan.

Rule 37(e)—the spoliation of ESI. The compartment everyone fears, and the one most litigated. Rule 37(e), rewritten in 2015, governs the consequences when ESI that should have been preserved is lost because a party failed to take reasonable steps to preserve it and it cannot be restored or replaced. Only when those predicates are met does the court reach remedies—which divide into two sharply different tracks:

• Under Rule 37(e)(1), if the loss prejudiced another party, the court may order measures "no greater than necessary to cure the prejudice"—excluding evidence, letting the parties argue the loss to the jury, shifting fees. What it may not do under (e)(1) is impose the severe sanctions.
• Under Rule 37(e)(2), only if the court finds the party "acted with the intent to deprive another party of the information's use" may it impose the heavy artillery: a presumption that the lost information was unfavorable, an adverse-inference instruction, dismissal, or default judgment. And here is the doctrinal heart of modern spoliation law: negligence—even gross negligence—is not enough. The 2015 amendment deliberately abrogated the older line (in the Second Circuit, Residential Funding Corp. v. DeGeorge Financial Corp., 306 F.3d 99 (2d Cir. 2002)) that allowed adverse inferences for mere negligence. For ESI, you now need intent.

That negligence/intent line determines whether the most powerful sanctions are even on the table, and courts infer intent from circumstantial conduct: selective preservation (backing up some data while letting the damaging data go), Ronnie Van Zant, Inc. v. Pyle, 270 F. Supp. 3d 656 (S.D.N.Y. 2017); a multi-year failure to implement any hold combined with continuous deletion, CrossFit, Inc. v. National Strength & Conditioning Ass'n (S.D. Cal. 2019); and supervisors instructing employees to delete, wiping software, or deleting laptop data after a preservation demand, Roadrunner Transportation Services, Inc. v. Tarwater, 642 F. App'x 759 (9th Cir. 2016) (default judgment affirmed). One unsettled wrinkle: Rule 37(e) is silent on the standard of proof, and courts split between preponderance, Ramirez v. T&H Lemont, Inc., 845 F.3d 772 (7th Cir. 2016), and clear-and-convincing evidence for the severest sanctions—so check your circuit and, where unsettled, brief to the higher standard. We dig deeper in our overview of sanctions for ESI spoliation under FRCP 37(e).

Beyond Rule 37. Rule 26(g) sanctions reach discovery requests, responses, and objections directly; Rule 30(d)(2) sanctions deposition misconduct; and the court's inherent authority provides a backstop for bad-faith conduct the rules do not reach, Chambers v. NASCO, Inc., 501 U.S. 32 (1991)—though inherent-authority fee awards must satisfy but-for causation, Goodyear Tire & Rubber Co. v. Haeger, 581 U.S. 101 (2017), and a leading post-Goodyear decision affirmed substantial compensatory sanctions calibrated to the costs caused by discovery misconduct, Klipsch Group, Inc. v. ePRO E-Commerce Ltd., 880 F.3d 620 (2d Cir. 2018). The message runs through this entire guide: discovery obligations are not aspirational. They are enforced, sometimes ruinously, against the parties and lawyers who ignore them.

For our hypothetical, the architecture is not abstract. If Cardinal's ninety-day auto-delete swallowed the emails showing when Dana's design files arrived, the case may turn on which track of Rule 37(e) applies. If Cardinal was merely careless—nobody thought to suspend the purge—Meridian gets (e)(1) relief. But if Meridian can show someone at Cardinal understood litigation was coming and let the purge run anyway, or ran wiping software the weekend after the demand letter, that is intent, and the door to an adverse-inference instruction or even default swings open. The difference between those outcomes is, quite literally, the case.

A Note on Alternatives: Discovery Outside Federal Court

The federal discovery machine described here is not the only model, and part of advising a client well is knowing when to avoid it. Arbitration typically offers far more limited discovery—often a modest document exchange and a few depositions—which can be a feature rather than a bug for a party that wants to control cost and avoid full-blown e-discovery. A client weighing a dispute-resolution clause, or deciding whether to litigate or arbitrate, should understand the choice carries profound discovery consequences. We compare these tradeoffs in our guides to alternative dispute resolution, the intricacies of AAA commercial arbitration, and arbitration, mediation, and choosing a dispute-resolution forum. For Meridian the calculus is fixed—its supply agreement sends the contract claim to court—but a different client, drafting a different agreement, might keep a future fight out of the discovery process altogether.

Frequently Asked Questions

How many interrogatories, document requests, and requests for admission can I serve? Rule 33 caps interrogatories at twenty-five per party, including discrete subparts, absent a stipulation or court order. Rules 34 and 36 impose no federal numerical limit—but many districts cap them by local rule, and a court may limit any of them as cumulative or disproportionate under Rule 26(b)(1)–(2). Always check the local rules and scheduling order before assuming there is no ceiling.

When does my duty to preserve evidence begin? When litigation is reasonably anticipated—frequently before a complaint is filed. Deciding to sue, receiving a demand letter, an internal report of a serious incident, or a government inquiry can each trigger the duty. The safest practice is to issue a written litigation hold the moment a dispute looks likely, suspend any automatic deletion, and document the steps you took. Waiting for service of a complaint is often too late.

What is the difference between Rule 37(e)(1) and 37(e)(2) sanctions for lost ESI? Everything, if you lost the data. Under (e)(1), a court that finds prejudice may impose curative measures "no greater than necessary"—but not the severe ones. Under (e)(2), a court may impose the severe sanctions (adverse-inference instruction, dismissal, default) only on a finding that the party acted with intent to deprive the other side of the information. Negligence, even gross negligence, will not get an opponent to (e)(2). The whole fight in most ESI-spoliation disputes is whether the conduct crosses from carelessness into intent.

Can I refuse to answer a deposition question I think is irrelevant? Generally no. Under Rule 30(c)(2) the examination proceeds even over objection, and a relevance objection is not a basis to instruct a witness not to answer. Counsel may instruct silence only to preserve a privilege, to enforce a court-ordered limitation, or to move to terminate a deposition conducted in bad faith. Improper instructions and coaching "speaking objections" are sanctionable under Rule 30(d)(2).

What happens if I miss the deadline to respond to requests for admission? The matters are automatically deemed admitted under Rule 36(a)(3)—no motion required—and, under Rule 36(b), conclusively established unless the court permits withdrawal. A blown RFA deadline can decide a case. Calendar these with extra care, and if a deadline has slipped, move promptly under Rule 36(b) to withdraw before your opponent relies on them.

Do I have to log every privileged document individually? Rule 26(b)(5)(A) requires a description sufficient to let the other side assess each claim, which traditionally meant a document-by-document log. But categorical logging is increasingly accepted where an individualized log would be unduly burdensome, and the December 2025 amendments push parties to agree on the method and timing of logging at the outset. Negotiate categorical logging and a log deadline at your Rule 26(f) conference; do not leave it to the end.

Is a black box over text in a PDF a safe redaction? No—and this mistake has burned major institutions. A visual box that does not remove the underlying text can be defeated by copying the passage or inspecting the file. Use redaction tools that permanently delete the underlying data, and verify before producing. An ineffective redaction is worse than none, because it discloses while creating a false belief the information is protected.

What is a Rule 30(b)(6) deposition, and why is it dangerous? It is a deposition of an organization, which must designate a witness to testify about its knowledge—not personal knowledge—on the noticed topics, prepared by reviewing documents and consulting employees. It is dangerous because the testimony binds the entity, and an unprepared designee who repeatedly says "I don't know" can expose the organization to a motion to compel and sanctions. A 2020 amendment requires the parties to confer in good faith about the topics in advance.

Conclusion: Building a Discovery Practice That Works

Discovery is not the glamorous part of practicing law. It is detailed, time-consuming, frequently contentious, and often thankless. But it is the phase where cases are won and lost, where evidence is gathered or lost forever, where the bulk of every client's legal bill is spent, and where the reputations of counsel are quietly made. The fundamentals covered here are the core of federal discovery practice, and none is especially mysterious. What separates the effective practitioner from the ineffective one is rarely arcane knowledge; it is attention to detail, discipline about deadlines, the willingness to do the hard work of preparation, and the judgment to tell a fight worth having from one better resolved by cooperation.

A closing thought, then, and the single most useful habit a litigator can build: when in doubt about a discovery issue, communicate. Pick up the phone. Send a thoughtful email. Most of the disputes that consume court time and client money could have been resolved by a timely, professional conversation—and Rule 37(a)(1) makes that expectation a precondition to relief. The lawyers who earn reputations for cooperative, good-faith practice find those reputations pay dividends—in judicial goodwill, in the quality of their work, and in the results they achieve. Master the engine room, and the ship sails.

Related Articles

• A comprehensive guide to federal civil litigation for small businesses
• Evaluating and assessing a civil case
• Mastering interrogatories: objecting and responding in civil litigation
• Mastering document discovery: objecting and responding to requests for production
• Strategic responses to requests for admission
• The art and science of depositions in federal civil litigation
• The art of defending depositions in federal court
• Rule 37(c)(1) preclusion of undisclosed evidence
• Trade secrets in the age of remote work and cloud computing
• Building a trade secret protection program from scratch
• Capturing the web: authenticating website screenshots as evidence in federal court
• Navigating the paper trail: a guide to federal civil litigation filings
• Arbitration: a comprehensive guide to alternative dispute resolution
• The intricacies of AAA commercial arbitration

This article is for general informational and educational purposes only and does not constitute legal advice. The Meridian–Cardinal scenario is entirely hypothetical. Discovery practice varies by jurisdiction, court, and case type, and the rules change—most recently with the privilege-logging amendments effective December 1, 2025. Consult the applicable Federal Rules of Civil Procedure, the local rules and standing orders of your court, and current case law for the requirements governing your specific matter.

Selected Authorities

Statutes and rules: Freedom of Information Act, 5 U.S.C. § 552 (including the nine exemptions at § 552(b) and the twenty-business-day response requirement at § 552(a)(6)(A)(i)); Fed. R. Civ. P. 1, 16(b)(3)(B)(iv), 26(a), 26(b)(1), 26(b)(3), 26(b)(5)(A), 26(e), 26(f)(3)(D), 26(g), 30, 30(b)(6), 30(c)(2), 30(d), 30(e), 33, 34, 36, 37(a), 37(b), 37(c), 37(d), 37(e), 45; Fed. R. Evid. 501, 502(a)–(d); the December 1, 2025 amendments to Rules 16 and 26 governing the timing and method of privilege claims.

Cases: Upjohn Co. v. United States, 449 U.S. 383 (1981) (corporate attorney-client privilege); Hickman v. Taylor, 329 U.S. 495 (1947) (work product); Zubulake v. UBS Warburg LLC, 220 F.R.D. 212 and 229 F.R.D. 422 (S.D.N.Y. 2003, 2004) (preservation and counsel's oversight); Pension Comm. of Univ. of Montreal Pension Plan v. Banc of Am. Sec., LLC, 685 F. Supp. 2d 456 (S.D.N.Y. 2010), abrogated in part by Chin v. Port Auth., 685 F.3d 135 (2d Cir. 2012); Silvestri v. Gen. Motors Corp., 271 F.3d 583 (4th Cir. 2001) (preservation trigger); Da Silva Moore v. Publicis Groupe SA, 287 F.R.D. 182 (S.D.N.Y. 2012) (technology-assisted review); Fischer v. Forrest (S.D.N.Y. 2017) and Liguria Foods, Inc. v. Griffith Labs., Inc., 320 F.R.D. 168 (N.D. Iowa 2017) (boilerplate objections); Burlington N. & Santa Fe Ry. Co. v. U.S. Dist. Court, 408 F.3d 1142 (9th Cir. 2005) (privilege-log waiver); Ramirez v. T&H Lemont, Inc., 845 F.3d 772 (7th Cir. 2016) (spoliation standard of proof); Ronnie Van Zant, Inc. v. Pyle, 270 F. Supp. 3d 656 (S.D.N.Y. 2017), Roadrunner Transp. Servs., Inc. v. Tarwater, 642 F. App'x 759 (9th Cir. 2016), and Klipsch Grp., Inc. v. ePRO E-Commerce Ltd., 880 F.3d 620 (2d Cir. 2018) (intent and ESI sanctions); Legal Voice v. Stormans Inc., 738 F.3d 1178 (9th Cir. 2013) (cost-shifting for non-party subpoenas); Specht v. Google, Inc., 268 F.R.D. 596 (N.D. Ill. 2010) and Brincko v. Rio Props., Inc., 278 F.R.D. 576 (D. Nev. 2011) (deposition conduct).

Secondary and reference sources: The Electronic Discovery Reference Model (EDRM) workflow; Advisory Committee Notes to the 2015, 2020, and 2025 amendments to the Federal Rules of Civil Procedure; Thomson Reuters Practical Law practice notes on discovery scope, proportionality, the written discovery devices, depositions, privilege and privilege logs, e-discovery, third-party subpoenas, and discovery sanctions.