Identifying Trade Secrets in Litigation: Reasonable Particularity, Sequencing Discovery, and Protecting the Secret While You Sue

Summary. Trade secret litigation has a discovery problem no other case has: the plaintiff must describe the thing it is suing over, in writing, to the very competitor it accuses of stealing it, before it knows exactly what was taken. This article explains why courts require identification of the asserted trade secrets with reasonable particularity before discovery begins, tracing the rule from California Code of Civil Procedure section 2019.210 and the four policy reasons in DeRubeis v. Witten Technologies through the federal appellate decisions in InteliClear, Mallet, Oakwood Laboratories, TLS Management, IDX Systems, and Imax. It explains what an identification must contain, how compilation and combination secrets and negative know-how are described, when the identification can be amended, and what happens at summary judgment and trial when a plaintiff identifies too much or too little. It then covers the sequencing fight under Rule 26(b)(1), the mechanics of two-tier protective orders, source code review rooms, prosecution bars, sealing under 18 U.S.C. section 1835, Rule 502(d) clawback, and in camera review. The final sections address forensic discovery, including neutral examiners, imaging protocols, privacy objections, the DTSA ex parte seizure remedy, and Rule 37(e) spoliation in departing-employee cases, and close with an extended worked example, strategy for each side, a cost discussion, common mistakes, and a FAQ.

Three weeks after a senior firmware engineer leaves Juniper Labs for a competitor, the IT department finds that a 64 GB USB drive was plugged into her workstation on her last Friday and that 2,900 files left with it. Juniper's general counsel wants to sue by Monday. Its outside counsel asks a question that sounds simple and is not: "Which trade secrets, exactly, did she take?" The honest answer is "we think the ones in those 2,900 files, but we have not read them all, and some of them are public." That answer will not survive the defendant's first motion.

Every other kind of intellectual property case begins with a public document that defines the right: a patent with claims, a copyright registration with a deposit, a trademark registration with a drawing. A trade secret case begins with the opposite. The right exists only because the information is not written down anywhere the public can see, and the plaintiff must now write it down for a court, an adversary, and an adversary's lawyers. How much it must say, when it must say it, who gets to read it, and what the other side must give up in return are the questions that decide most trade secret cases long before anyone talks about damages.

The short answer

Most courts, state and federal, require the plaintiff in a trade secret case to identify each asserted trade secret with reasonable particularity (enough detail to separate the claimed secret from general knowledge in the field and to let the defendant and the court know what is actually in dispute) before the plaintiff takes discovery into the defendant's own technology. California made this a statutory precondition to discovery in Cal. Civ. Proc. Code § 2019.210, and Massachusetts later adopted a similar rule. Many federal courts impose a similar sequence as a matter of case management under Rule 16 and Rule 26, although the Ninth Circuit held in 2025 that the federal Defend Trade Secrets Act itself imposes no up-front identification requirement and that a court may not use an identification order to strike secrets before the merits are tested. The Sedona Conference and the Federal Judicial Center have both published guidance endorsing early, written, numbered identifications. The identification is not a one-time pleading; it is a living document that can be amended for good cause as discovery reveals what the defendant actually took, but a plaintiff who identifies its secrets vaguely, or changes them late, loses at summary judgment or has the late theory excluded at trial.

The second half of the problem is protecting the secret while litigating over it. The tools are a two-tier protective order under Rule 26(c)(1)(G) (a "confidential" tier that the parties may see and an "attorneys' eyes only" tier that only outside counsel and approved experts may see), source code review protocols, prosecution and competitive-decision-making bars, sealing orders supported by 18 U.S.C. § 1835, clawback orders under Fed. R. Evid. 502(d), and in camera review. The third half (trade secret cases have three halves) is forensic discovery of the defendant's devices and accounts, usually through a neutral examiner working under a court-approved protocol, backed by the preservation duty and the sanctions power in Rule 37(e).

This article is the analytical companion to Managing Discovery in a Trade Secret Case: A Practical Guide, which walks through the same material chronologically, the Trade Secret Identification and Discovery Checklist, and the Trade Secret Litigation Toolkit, which contains model identification statements, a two-tier protective order, and a forensic protocol.

Why discovery in a trade secret case is unlike any other

Three structural features drive everything that follows.

The claim defines itself during the case. A patent plaintiff cannot change its claims to match what the defendant turns out to be doing; the Patent Office fixed them years earlier. A trade secret plaintiff can, in principle, read the defendant's documents and then announce that whatever it found there was its secret. Courts call this "molding" the claim to the discovery, and preventing it is the single most-cited reason for requiring identification first. See the discussion of DeRubeis v. Witten Technologies, Inc., 244 F.R.D. 676 (N.D. Ga. 2007), below.

Discovery is itself a form of the harm. In an ordinary commercial case, producing documents to the other side is a nuisance. In a trade secret case, the information the defendant must produce (its own formulas, its own code, its own customer terms) is often its most valuable property, and the party receiving it is a direct competitor who has just accused it of theft. The plaintiff's production carries the same risk in reverse: the identification statement is a written description of the secret, handed to the competitor. Congress recognized this when it directed courts in 18 U.S.C. § 1835(a) to "enter such orders and take such other action as may be necessary and appropriate to preserve the confidentiality of trade secrets" in any prosecution or civil proceeding under the chapter.

The best evidence is on the defendant's devices, and the defendant has every incentive to lose it. A departing employee who copied files to a USB drive and then, after receiving a demand letter, ran a disk-wiping utility has converted a defensible case into an indefensible one. The forensic record (file system metadata, USB connection logs, cloud sync histories, deleted-file artifacts) is both the plaintiff's proof of misappropriation and the thing most at risk in the first thirty days. See Litigation Holds, Spoliation, and Rule 37(e).

Put the three together and the shape of the litigation becomes clear. The plaintiff wants the defendant's devices imaged yesterday and wants to see the defendant's documents before committing to what its secrets are. The defendant wants the plaintiff to commit first, in detail, and wants no one touching its systems. The court wants a sequence that lets the case proceed without converting discovery into a second misappropriation.

The legal framework

The statutes that define the thing being identified

The federal Defend Trade Secrets Act (DTSA) defines a trade secret in 18 U.S.C. § 1839(3) as "all forms and types of financial, business, scientific, technical, economic, or engineering information," including "patterns, plans, compilations, program devices, formulas, designs, prototypes, methods, techniques, processes, procedures, programs, or codes," provided the owner has taken reasonable measures to keep it secret and it derives independent economic value from not being generally known or readily ascertainable through proper means. Section 1 of the Uniform Trade Secrets Act (UTSA), adopted in nearly every state (New York is the conspicuous holdout, and North Carolina has its own similar statute), is materially the same. Misappropriation is defined in § 1839(5), the civil cause of action is in 18 U.S.C. § 1836(b), and 18 U.S.C. § 1838 preserves state law, which is why most complaints plead the DTSA and the state statute side by side.

Nothing in either statute says when or how the plaintiff must identify what it is suing over. The DTSA contains no identification requirement, and the federal courts that impose one draw it from the Federal Rules of Civil Procedure and their case-management power. The explicit statutory identification rules are state rules.

California Code of Civil Procedure § 2019.210 and Massachusetts

Section 2019.210 provides that in any action alleging misappropriation under the California UTSA, "before commencing discovery relating to the trade secret, the party alleging the misappropriation shall identify the trade secret with reasonable particularity subject to any orders that may be appropriate under Section 3426.5 of the Civil Code" (the provision directing courts to preserve secrecy through protective orders, sealing, and in camera hearings). The Legislature enacted the rule in 1984, when it adopted the UTSA, as subdivision (d) of former section 2019; it was renumbered without substantive change effective July 1, 2005. Massachusetts, which adopted the UTSA in 2018, added a parallel provision, Mass. Gen. Laws ch. 93, § 42D(b), requiring identification "with sufficient particularity under the circumstances of the case" before discovery relating to the secret begins.

The California courts describe the statute as serving four interrelated goals, borrowed from Computer Economics, Inc. v. Gartner Group, Inc., 50 F. Supp. 2d 980 (S.D. Cal. 1999): it promotes well-investigated claims and dissuades meritless ones, keeps plaintiffs from using discovery to obtain the defendant's secrets, helps the court frame the scope of discovery, and lets defendants build their defenses without waiting until the eve of trial to learn what they are accused of taking.

Does California's rule apply in federal court, and what changed in 2025

For years the federal district courts in California split over whether § 2019.210 governs California UTSA claims litigated in federal court. Computer Economics applied its predecessor under Erie; others declined to treat it as binding but imposed the same sequence under Rule 16 and Rule 26. In practice, a plaintiff pleading a California UTSA claim in the Northern or Central District of California should expect to identify its secrets before it takes discovery about them; in Loop AI Labs Inc. v. Gatti, 195 F. Supp. 3d 1107 (N.D. Cal. 2016), the court recounted its earlier order applying § 2019.210 to the state claim and staying discovery on that claim until a compliant disclosure was served.

DTSA-only cases are now different in the Ninth Circuit. In Quintara Biosciences, Inc. v. Ruifeng Biztech, Inc., No. 23-16093 (9th Cir. Aug. 12, 2025), the district court borrowed § 2019.210 for a case that pleaded only a DTSA claim, ordered the plaintiff to list each secret "as claims would appear at the end of a patent," and then, on a motion under Rule 12(f), struck nine of the plaintiff's eleven asserted secrets. The Ninth Circuit reversed. It held that, unlike the California statute, the DTSA does not require a plaintiff to identify its secrets with particularity at the start; that whether a secret has been identified with "sufficient particularity" is a question of fact usually resolved on summary judgment or at trial; that Rule 12(f) is not a vehicle for striking trade secrets; and that functionally dismissing claims as a discovery sanction was an abuse of discretion absent extreme circumstances. The court was equally clear that district judges keep their Rule 16 management powers and their authority under Rule 26(c)(1)(G) to order that trade secrets "be revealed only in a specified way." The practical lesson: federal courts may still sequence discovery around an identification, but they may not turn the identification order into an early merits ruling.

The federal rules that govern the fight

Rule 26(b)(1) limits discovery to matter "relevant to any party's claim or defense and proportional to the needs of the case." In a trade secret case, the claim is not fully defined until the secrets are identified, so relevance itself is indeterminate until the identification exists. That is the doctrinal hook most federal courts use.

Rule 26(c)(1)(G) authorizes a protective order "requiring that a trade secret or other confidential research, development, or commercial information not be revealed or be revealed only in a specified way." It is the source of every two-tier protective order, source code protocol, and prosecution bar.

Rule 34 governs document requests and, in subsection (a)(1)(A), requests to "inspect, copy, test, or sample" electronically stored information, which is the textual basis for forensic imaging of an opponent's devices. Rule 45 does similar work for non-parties, including the new employer that has not been sued. Rule 33 is the usual vehicle for the identification itself when no standalone disclosure is ordered, and Rule 30(b)(6) governs the corporate deposition in which each side tests the other's account of its secrets and its systems. Rule 37(e) supplies the sanctions regime for lost electronically stored information. Fed. R. Evid. 502(d) lets the court order that privileged material produced in the case is not waived, which matters enormously when a forensic image of a laptop is handed to an examiner.

Finally, 18 U.S.C. § 1836(b)(2) authorizes an ex parte civil seizure "in extraordinary circumstances," and 18 U.S.C. § 1835 both directs courts to preserve confidentiality and, in subsection (b), bars a court from authorizing disclosure of information the owner asserts is a trade secret without first giving the owner a chance to explain, under seal, its interest in keeping it confidential. We discuss below why the seizure remedy is sought often and granted rarely.

Why courts make the plaintiff go first: DeRubeis, the commentaries, and the California cases

The DeRubeis framework

The most quoted explanation of the identification requirement comes from DeRubeis v. Witten Technologies, Inc., 244 F.R.D. 676 (N.D. Ga. 2007). Two former employees sued their old employer, Witten, over their employment agreements; Witten counterclaimed that they were using its underground-mapping technology in their competing business and demanded discovery into how they performed their services. Asked to identify its secrets, Witten served a list that, the court observed, did not specify any trade secret at all but described "the end results of, or the functions performed by," the claimed secrets. That pattern (describing what the technology does rather than what the secret is) remains the most common identification failure.

The court catalogued three policies favoring the trade secret claimant's right to early discovery: the broad scope of discovery under the federal rules, the fact that a company with many secrets may not know which were taken until it sees how the defendant operates, and the "Catch-22" that a list drafted blind will be either too general to survive or too specific to capture what the defendant is actually doing. It then identified four policies favoring identification first:

1. Discovery should not become a vehicle for acquiring the defendant's secrets. If the plaintiff may demand the defendant's formulas and code before saying what it claims, a lawsuit becomes a fishing expedition, and a competitor who sues can learn in discovery what it could not learn in the market.
2. Relevance cannot be measured without a target. Quoting Xerox Corp. v. International Business Machines Corp., 64 F.R.D. 367 (S.D.N.Y. 1974), the court explained that the burden is on the plaintiff to specify its charges, "not upon the defendant to guess at what they are," and that until it does, neither the court nor the parties can know whether discovery is relevant.
3. The defendant cannot defend what it cannot see. Every trade secret defense (non-use, general knowledge, independent development, reverse engineering) depends on knowing what is claimed.
4. The plaintiff should not be allowed to mold its claim to the discovery. If the plaintiff is not forced to define its secrets before it gets into the defendant's files, it may simply claim whatever it finds there.

The court rejected any "talismanic procedure" and noted that courts had developed at least nine different approaches, from staged discovery to simultaneous exchanges. On the facts before it, it required Witten to identify its secrets with reasonable particularity, which it defined in two parts that most federal courts still use: a description sufficient (1) to put the other side on notice of the nature of the claims and (2) to let it discern the relevancy of any requested discovery. A protective order with an attorneys' eyes only tier was already in place, which the court treated as answering the claimant's fear that identifying its secrets would expose them.

The Sedona Conference commentary and the FJC judicial guide

Two secondary sources now shape how judges handle identification. The Sedona Conference, a think tank whose working groups write influential commentaries on discovery, published its Commentary on the Proper Identification of Asserted Trade Secrets in Misappropriation Cases, 22 Sedona Conf. J. 223 (2021). Its principles reduce to four propositions: identification is a procedural step, not an adjudication that the information is a trade secret; the claimant should identify its asserted secrets in writing early in the case; the particularity required is what is reasonable given the nature of the information and the stage of the case; and the identification may be amended for good cause.

The Federal Judicial Center's Trade Secret Case Management Judicial Guide (2023), written by Peter Menell and a group of experienced practitioners and academics for federal judges, goes further and recommends specific practices: identification after the pleadings but before or early in discovery; a written, numbered list in which each number is a separate secret; documents used as support rather than as a substitute for a narrative description; no general stay of discovery while the parties argue about sufficiency; and a good-cause standard for amendments. It also warns judges not to import patent-claim drafting requirements into trade secret identification. Parties negotiating a case-management order should assume the judge has read both documents, and should cite them.

California's statutory rule: Advanced Modular Sputtering, Brescia, and Perlan

Because § 2019.210 has been litigated longer than any federal equivalent, the California courts of appeal have produced the most developed body of law on what "reasonable particularity" means. Three decisions do most of the work, and they are cited in federal courts nationwide.

Advanced Modular Sputtering, Inc. v. Superior Court, 132 Cal. App. 4th 826 (2005), is the cautionary tale about overenforcement. A maker of sputtering equipment (machines that deposit thin films on silicon wafers) sued two laid-off employees whose new company was soon selling competing machines to its customers. The plaintiff served three successive designations of eight combination secrets; each time, the defendants submitted expert declarations calling them vague and generally known, and each time a discovery referee and the trial court agreed. The parties spent "thousands of dollars on attorney fees and expert witnesses" without ever beginning discovery. The court of appeal granted writ relief. It held that reasonable particularity "does not mean that the party alleging misappropriation has to define every minute detail of its claimed trade secret at the outset of the litigation," nor does it require "a miniature trial on the merits" before discovery. More exacting detail may be needed where the secrets are incremental advances in a specialized technical field, but where credible experts declare that they can understand the designation and distinguish it from what is already known, the designation should generally be considered adequate. The court also held that because every one of the plaintiff's ten causes of action depended on the misappropriation allegation, discovery on all of them waited for the identification.

Brescia v. Angelin, 172 Cal. App. 4th 133 (2009), concerned a formula and manufacturing process for a high-protein, low-carbohydrate pudding with an extended shelf life (the defendants on appeal included the actor Sylvester Stallone, chairman of a company accused of making a competing product). The trial court demanded that the plaintiff explain how his secrets differed from the general knowledge of food scientists. The court of appeal reversed. It held that § 2019.210 does not require that explanation in every case; it is required only when the details themselves are inadequate, given the nature of the secret and the field, to let the defendant learn the boundaries of the secret and investigate defenses or to let the court fashion discovery. The designation is to be liberally construed, and reasonable doubts about its adequacy are resolved in favor of letting discovery proceed. Notably, the court agreed that the plaintiff's earlier attempts had been properly rejected; what finally worked was a succinct, specific statement of the formula and process.

Perlan Therapeutics, Inc. v. Superior Court, 178 Cal. App. 4th 1333 (2009), is the counterweight. A biotechnology company accused its founder and his wife of taking protein-based antiviral technology to a new company. Its amended statement opened with a page of boilerplate objections, repeated the public complaint, described its technology in general scientific terms, and referred to hundreds of pages of attached documents. The trial court found it inadequate, and the court of appeal denied writ relief, emphasizing that appellate courts review these rulings for abuse of discretion and that trial judges are best placed to spot gamesmanship. Its key sentences are quoted in identification disputes everywhere: a plaintiff "is not entitled to include broad, 'catch-all' language as a tactic to preserve an unrestricted, unilateral right to subsequently amend its trade secret statement," and "[i]f Perlan does not know what its own trade secrets are, it has no basis for suggesting defendants misappropriated them." Nor may a plaintiff "hide its trade secrets in 'plain sight' by including surplusage and voluminous attachments." The court added that if discovery revealed that additional secrets were taken, the plaintiff might show good cause to amend.

Read together, the three cases set the California standard: enough detail to draw a boundary, not proof of secrecy; no prior-art essay unless the description is otherwise unintelligible; and no hiding behind volume, boilerplate, or generality.

The federal appellate decisions, explained

Federal courts required identification at summary judgment and trial for decades before they began imposing it as a discovery sequence. The two lines have now merged, and the appellate cases below are cited in both contexts. They also reveal real circuit differences in tone, if not in doctrine.

Imax and IDX: the summary judgment cases

Imax Corp. v. Cinema Technologies, Inc., 152 F.3d 1161 (9th Cir. 1998), remains the standard citation for under-identification. Imax claimed the "dimensions and tolerances" of its large-format "rolling loop" projector, "including every dimension and tolerance that defines or reflects that design," but through four rounds of supplemental interrogatory answers never identified the precise numerical values it claimed. The Ninth Circuit affirmed summary judgment. The catchall phrase did not "clearly refer to tangible trade secret material"; the defendant "could not be expected to prepare its rebuttal" without a concrete identification of exactly which dimensions and tolerances were at issue; and neither a judge nor a jury had the expertise to sort through a complex projector to find them. On those facts, reasonable specificity "could only be achieved by identifying the precise numerical dimensions and tolerances."

IDX Systems Corp. v. Epic Systems Corp., 285 F.3d 581 (7th Cir. 2002), is the companion case for over-identification. IDX, a medical practice software vendor, tendered a 43-page description of its software and claimed it as a trade secret. Judge Easterbrook's opinion held the description "both too vague and too inclusive," because it effectively asserted that all information in or about the software was secret, and it did not separate the secrets from the rest of what goes into any software package. The opinion's most quoted line is the one every plaintiff should tape above the desk: "a plaintiff must do more than just identify a kind of technology and then invite the court to hunt through the details in search of items meeting the statutory definition." The same opinion also acknowledged the plaintiff's dilemma with unusual candor: "Reluctance to be specific is understandable; the more precise the claim, the more a party does to tip off a business rival to where the real secrets lie."

InteliClear and Quintara: the Ninth Circuit's modern synthesis

InteliClear, LLC v. ETC Global Holdings, Inc., 978 F.3d 653 (9th Cir. 2020), is the case federal practitioners now cite first. InteliClear licensed its securities clearance and settlement system to ETC under a confidentiality agreement. While the license was still running, ETC began building its own system, and InteliClear's architect later noticed a table in ETC's system with the same unusual column names as his. A forensic expert compared the two systems with ETC's consent and reported "abundant evidence" of identical elements. ETC moved for summary judgment the day after discovery opened, arguing that InteliClear had not identified its secrets with sufficient particularity, and the district court granted the motion before any discovery had occurred.

The Ninth Circuit reversed on two grounds. First, on substance, InteliClear's sealed declaration specified the tables, columns, account identifiers, codes, and methods of populating data that it claimed, which was enough to create a triable issue. The court restated the limits (plaintiffs "may not simply rely upon 'catchall' phrases or identify categories of trade secrets," and it is inadequate to "cite and incorporate by reference hundreds of documents"), but held that where no discovery had occurred it was "not fatal" that InteliClear's hedging language left open the possibility of expanding its identification later; its burden at that stage was "only to identify at least one trade secret with sufficient particularity to create a triable issue." Second, on sequence, the district court abused its discretion under Rule 56(d) by ruling before any discovery. Identification disputes, the court observed, "typically arise in the battleground of discovery," which provides "an iterative process where requests between parties lead to a refined and sufficiently particularized trade secret identification." Refining identifications through discovery "makes good sense."

Quintara, discussed above, applied InteliClear to the procedural question it left open. Read together, the two cases mean that in the Ninth Circuit an identification order is a case-management tool: a court may require one, may enforce it with ordinary discovery remedies, and may resolve particularity on a full record at summary judgment, but it should not use the order to end claims at the threshold.

Mallet and Oakwood: the Third Circuit in the same year

In Mallet & Co. v. Lacayo, 16 F.4th 364 (3d Cir. 2021), a maker of baking release agents (the lubricants that let bread and cakes slide out of commercial pans) obtained a preliminary injunction against two former employees, their new employer Synova, and Synova's parent company, Bundy. The injunction barred the employees from working for Bundy or Synova and restrained the companies from competing. The Third Circuit vacated it. The district court had never identified which information it was protecting; its order described "general categories of business and technical information, a list that could be used to describe documents found in any number of corporations." Mallet itself had described its secret as "the overall body of knowledge" connecting the development, production, and application of its products. The court held that "a district court's injunction order must first adequately identify the information to which it accords trade secret status"; otherwise the order lacks a foundation for the likelihood-of-success finding and fails Rule 65(d), which requires an injunction to state its terms specifically and describe the restrained acts in reasonable detail. The court also sent the bond amount back for a case-specific analysis. Mallet is the case to cite when a plaintiff seeks emergency relief on a thin identification. For injunction strategy generally, see the Trade Secret Remedies Toolkit.

Oakwood Laboratories LLC v. Thanoo, 999 F.3d 892 (3d Cir. 2021), addressed the pleading stage and came out the other way. Oakwood alleged that its former vice president of product development took its microsphere drug-delivery know-how to a competitor that had also received a confidential memo during earlier business discussions. Over two years the district court dismissed four successive complaints. The Third Circuit vacated the dismissal, holding that the final complaint "easily" met the pleading standard. A plaintiff need not "spell out the details of the trade secret" in a public complaint; it must describe the subject matter with enough particularity to separate it from general knowledge and to let the defendant ascertain the boundaries within which the secret lies. The court also held that the district court erred by demanding heightened specificity about which secrets were used and how, and it adopted a broad reading of "use," which includes relying on a trade secret "to assist or accelerate research or development." The two decisions together describe a sliding scale: notice at the pleading stage, reasonable particularity before discovery into the defendant's technology, and specificity sufficient for an injunction or a verdict at the remedies stage.

TLS Management: losing at trial for failure to identify

TLS Management & Marketing Services, LLC v. Rodriguez-Toledo, 966 F.3d 46 (1st Cir. 2020), shows that the identification problem does not disappear once a plaintiff survives summary judgment. TLS, a Puerto Rico tax-planning firm, won a bench trial against its former managing director on claims that he took its "Capital Preservation Reports" and its "U.S. Possession Strategy." The First Circuit reversed and directed judgment for the defendants. Writing for the court, Judge Dyk observed that trade secrets, unlike patents, copyrights, and registered marks, have no boundaries defined before litigation, which creates the risk "that the trade secret owner will tailor the scope of the trade secret in litigation to conform to the litigation strategy"; in this case, the alleged secrets "were not identified by TLS until the eve of trial," even though Puerto Rico's statute requires a description "as specifically as possible" before discovery. The reports mixed public tax law with client-specific information, and at oral argument TLS could not say which parts were secret; its principal testified to "approximately 53 different methods or techniques" without describing them. The strategy consisted largely of well-known steps. A plaintiff that cannot articulate the secret loses even after winning the facts.

The Seventh Circuit's demand for "concrete secrets"

The Seventh Circuit has long been the least forgiving forum on identification, and its most recent decision confirms it. In NEXT Payment Solutions, Inc. v. CLEAResult Consulting, Inc., No. 24-1377 (7th Cir. Jan. 14, 2026), a software developer claimed that a customer studied its appointment-scheduling tool to improve a replacement system. After years of litigation and an order to explain which features were secret and why, the developer produced a spreadsheet of thirty-four software "modules" and five "combinations of modules and features." The district court granted summary judgment, and the Seventh Circuit affirmed, repeating that its case law "requires a high level of specificity" and that a plaintiff must show "concrete secrets" rather than "broad areas of technology." The district court had found that the developer offered "broad descriptions and jargon-laden terminology" describing what its software did, not how it did it. When the developer later tried to recast its remaining unjust enrichment claim around "other proprietary information" it had never identified, the court excluded that theory in limine, and the Seventh Circuit upheld that ruling too. Compare Life Spine, Inc. v. Aegis Spine, Inc., 8 F.4th 531 (7th Cir. 2021), where the same court affirmed a preliminary injunction for a spinal-implant maker that claimed "the precise dimensions and measurements" of its device's components and their interconnectivity, information unavailable from its patent and marketing materials and shared with its distributor only under a confidentiality agreement. Life Spine succeeded where Imax failed because it claimed the numbers, not the category.

The district court decisions that fill in the practice

Three district court decisions are cited constantly in case-management disputes. Vesta Corp. v. Amdocs Management Ltd., 147 F. Supp. 3d 1147 (D. Or. 2015), recognized a "growing consensus" requiring identification before the claimant may compel discovery of its adversary's secrets, applied that approach outside California to a claim about payment-solutions methods, and excused the defendants from responding to discovery on that topic until the plaintiff amended its interrogatory answers to identify its secrets with reasonable particularity. The court explained that a plaintiff claiming a method must identify its steps and explain what makes the combination unique, not merely label it a "method." The same court had earlier gone the other way in Nike, Inc. v. Enter Play Sports, Inc., 305 F.R.D. 642 (D. Or. 2015), declining to require more detail before discovery, which shows how fact-dependent the decision is.

Loop AI Labs, a later order in the case discussed above, shows what not to do. Ordered to serve a thorough and complete disclosure, the plaintiff filed one on the public docket, which the court observed "belies the proposition that it contains information specific enough to be considered 'confidential' trade secrets." It opened with catchall paragraphs incorporating everything identified in its pleadings and a range of Bates numbers, and much of the rest "reads like an inventory of categories" of scientific and business information. The court held the disclosure insufficient and in violation of its order, and referred the question of sanctions to the trial judge.

Sit-Up Ltd. v. IAC/InterActiveCorp, No. 05 Civ. 9292, 2008 WL 463884 (S.D.N.Y. Feb. 20, 2008), is the due-diligence case. A British television retailer that ran falling-price auctions let a would-be acquirer study its business, the deal died, and the acquirer launched its own channel. Asked by interrogatory to identify its secrets, the plaintiff answered that they were "strewn across at least 13,000 pages" produced in discovery. The court warned that the plaintiff would have to open to a jury, sum up, and resist summary judgment, and that it "can't proceed on that theory" if it could not identify its secrets now; it then gave the plaintiff a final opportunity and made that submission binding. The plaintiff listed more than one hundred secrets plus a "secret sauce" combining them. The court granted summary judgment on the combination claim because the plaintiff never showed how the components fit together to form a unique whole.

What "reasonable particularity" looks like on the page

The cases converge on a handful of drafting rules. An identification should be a written, numbered list in which each number is one secret. Each entry should say what the secret is (the formula, the parameter values, the sequence of steps, the selection criteria, the code module and what it does differently), not what it achieves. Documents may support an entry, but they cannot replace it: if a document is cited, the entry should say which part of the document embodies the secret, because a reference to an entire file invites the IDX objection that the court is being asked to hunt. Catchall phrases ("including but not limited to," "and all related know-how") add nothing and, after Perlan, may hurt. And the identification should be filed under seal or served under the protective order, never publicly docketed.

The level of detail depends on the kind of information. A useful way to think about it is to ask what a competent engineer, chemist, or salesperson in the defendant's position would need in order to search its own files and say "we use that" or "we do not."

Compilations and combination secrets

A trade secret can exist in a combination of elements each of which is public. The Second Circuit put it classically in Integrated Cash Management Services, Inc. v. Digital Transactions, Inc., 920 F.2d 171 (2d Cir. 1990): a "unified process, design and operation" of public components "in unique combination" can be protectable. The Ninth Circuit applied the same idea to compiled public data in Experian Information Solutions, Inc. v. Nationwide Marketing Services Inc., 893 F.3d 1176 (9th Cir. 2018), holding that a compilation of names and addresses could qualify even though its component parts were available elsewhere.

The identification burden is heavier for combinations, because the public pieces tell the defendant nothing. Since Struthers Scientific & International Corp. v. General Foods Corp., 51 F.R.D. 149 (D. Del. 1970), courts have required the claimant to describe "what particular combination of components it has in mind, how those components are combined, and how they operate in a unique combination." Sit-Up lost its "secret sauce" claim for failing to do exactly that, and Vesta required the plaintiff to explain what made its method a unique combination rather than a label. If the claim is that one subset of a combination is the secret, say so; a defendant who shows that the full combination is not used will otherwise win on non-use.

Negative know-how

Knowledge of what does not work can be valuable, because it lets a competitor skip the dead ends. In Metallurgical Industries Inc. v. Fourtek, Inc., 790 F.2d 1195 (5th Cir. 1986), the Fifth Circuit rejected a district court's view that "negative know-how" is categorically unprotectable. Negative know-how is also the easiest kind of secret to describe vaguely and the hardest to prove was used, because a defendant that avoided a dead end leaves no trace in its product. The FJC guide suggests identifying it either as the specific fact that particular approaches failed or by reference to identified research records. Pair it with evidence that the defendant's development timeline was shorter than independent work would allow.

Do not draft it like a patent claim

Patent claims serve a different function: they define a public, exclusive right and must satisfy 35 U.S.C. § 112. The FJC guide warns courts not to import those requirements into trade secret identification, and Quintara shows why: the district court there ordered the plaintiff to list its secrets "as claims would appear at the end of a patent" and then struck most of them. A good identification borrows the patent claim's discipline (numbered, element-by-element, no catchalls) without its legal consequences. It need not prove novelty, and in California it need not explain the prior art unless the description is otherwise unintelligible.

Amending the identification and carrying it through trial

Amending without moving the target

Identifications change. Discovery reveals that the defendant took a module nobody suspected, or that an asserted secret was published in a 2014 conference paper. The question is not whether amendment is allowed but on what terms. California and federal courts generally require good cause. Perlan said that a plaintiff who learns through discovery that additional secrets were misappropriated "may have good cause to amend," and Neothermia Corp. v. Rubicor Medical, Inc., 345 F. Supp. 2d 1042 (N.D. Cal. 2004), is regularly cited for requiring that showing; in Loop AI Labs the court warned at the outset that future amendments would be permitted only for good cause.

The FJC guide lists the factors courts actually weigh: whether the plaintiff was diligent, whether the defendant will be prejudiced, whether the amendment rests on newly learned facts, and how late in the case it comes. Amendments that add secrets after the close of fact discovery also run into Rule 16(b)(4)'s good-cause standard for modifying a scheduling order and into Rule 37(c)(1), which excludes information a party failed to disclose or supplement as Rule 26(e) requires, unless the failure was substantially justified or harmless. Dropping secrets, by contrast, is almost always allowed and usually wise.

Two housekeeping rules prevent chaos. Never renumber: if Trade Secret No. 6 is withdrawn, the amended list should say "No. 6: no longer asserted," because deposition transcripts, expert reports, and orders will have used the original numbers. And add new secrets at the end of the list, so that No. 14 means the same thing in every document in the case.

What happens at summary judgment and trial

The identification follows the case to judgment. At summary judgment, a defendant can win on the ground that the plaintiff has not identified a protectable secret with enough specificity for a fact finder to assess it, as in Imax, IDX, NEXT Payment Solutions, and Givaudan Fragrances Corp. v. Krivda, 639 F. App'x 840 (3d Cir. 2016), where the Third Circuit held that summary judgment was proper as to fragrance formulas about which the plaintiff "failed to provide enough specific information." Courts in other circuits say the same: "To survive summary judgment, [plaintiff] must identify specific documents or information that constitute trade secrets, not simply list categories or general topics of information." Kuryakyn Holdings, LLC v. Ciro, LLC, 242 F. Supp. 3d 789 (W.D. Wis. 2017). The plaintiff's response is usually a declaration from a knowledgeable witness or expert walking through each numbered secret, as in InteliClear, plus, where discovery has been cut short, a Rule 56(d) declaration explaining what discovery would show.

At trial, the numbered identification becomes the architecture of the case. Jury instructions and verdict forms should ask, secret by secret, whether each one exists and whether it was misappropriated; that structure makes the verdict reviewable and makes any damages award traceable to specific secrets, which matters for the apportionment fights discussed in Trade Secret Damages and Remedies. Courts routinely require plaintiffs to narrow a long list before trial, and a plaintiff that tries to present a secret, or a theory, that was never in the identification invites the exclusion the Seventh Circuit upheld in NEXT Payment Solutions. TLS Management shows the final risk: a plaintiff that wins the facts but cannot articulate the secret on appeal loses the judgment.

Sequencing discovery under Rule 26(b)(1)

Most sequencing fights are really proportionality fights. Rule 26(b)(1) asks whether discovery is proportional "considering the importance of the issues at stake in the action, 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 of the proposed discovery outweighs its likely benefit." In a trade secret case, the "parties' relative access" factor cuts both ways: the plaintiff has better access to its own secrets (which is why it should go first in describing them), and the defendant has better access to how it built its product (which is why the plaintiff needs discovery to refine its claim).

Courts have resolved the tension in roughly four ways, and a well-prepared party proposes one rather than arguing in the abstract:

1. Identification first, then technical discovery. The plaintiff serves its identification under the protective order; discovery into the defendant's technology opens only afterward. This is the California default and the approach in DeRubeis and Vesta.
2. Parallel tracks. Non-technical discovery (communications, departure facts, forensic evidence of copying) proceeds immediately, while discovery into the defendant's confidential technology waits for the identification. The FJC guide favors this approach and recommends against a general stay of all discovery.
3. Staged mutual disclosure. The plaintiff identifies, the defendant produces the specific technology that corresponds to the identified categories, and the plaintiff then serves a refined identification within a fixed period. This works well in software cases.
4. No sequencing. Some courts, as in Nike, decline to require more before discovery, usually where the plaintiff's description is already reasonably concrete or the defendant appears to be using the identification demand to delay.

Discovery runs in both directions, and the defendant's own secrets are not immune. The Supreme Court observed in Federal Open Market Committee v. Merrill, 443 U.S. 340 (1979), that there is no absolute privilege for trade secrets and similar confidential information in civil discovery; courts instead weigh the claim to confidentiality against the need for disclosure and typically protect the information through limits on who may see it. The courts of appeals use a burden-shifting test: the party resisting discovery must first show that the information is a trade secret and that disclosure might be harmful, and the requesting party must then show that it is relevant and necessary. Centurion Industries, Inc. v. Warren Steurer & Associates, 665 F.2d 323 (10th Cir. 1981); In re Remington Arms Co., 952 F.2d 1029 (8th Cir. 1991); see also American Standard Inc. v. Pfizer Inc., 828 F.2d 734 (Fed. Cir. 1987). The most famous application is Coca-Cola Bottling Co. of Shreveport, Inc. v. Coca-Cola Co., 107 F.R.D. 288 (D. Del. 1985), in which the court ordered production of the secret formulas under a protective order because they were relevant and necessary to the bottlers' contract claims. The identification is what makes the "relevant and necessary" showing possible: a plaintiff that has said exactly what it claims can explain exactly which of the defendant's files it needs.

For the wider set of discovery tools, see A Practical Discovery Refresher and, for the chronological version of this section, Managing Discovery in a Trade Secret Case: A Practical Guide.

Protecting the secret while you sue

The protective order

Protective orders in discovery rest on Rule 26(c)'s "good cause" standard and do not offend the First Amendment, because restraints on information obtained only through discovery are not restrictions on a traditionally public source of information. Seattle Times Co. v. Rhinehart, 467 U.S. 20 (1984). Good cause requires a particularized showing of the harm disclosure would cause, document by document if challenged, rather than broad allegations. Phillips ex rel. Estates of Byrd v. General Motors Corp., 307 F.3d 1206 (9th Cir. 2002); Pansy v. Borough of Stroudsburg, 23 F.3d 772 (3d Cir. 1994). In practice, trade secret cases proceed under stipulated umbrella orders, often based on a district's model order (the Northern District of California's model order for cases involving patents, highly sensitive information, and trade secrets is the most widely borrowed), with the good-cause showing deferred until someone challenges a designation.

The defining feature is the second tier. "Confidential" material may be seen by the parties' employees who need it; "Highly Confidential, Attorneys' Eyes Only" material is limited to outside counsel, approved experts, the court, and vendors. Courts approve two-tier orders on a showing that disclosure of especially sensitive information to a competitor would cause business harm. Layne Christensen Co. v. Purolite Co., 271 F.R.D. 240 (D. Kan. 2010); MGP Ingredients, Inc. v. Mars, Inc., 245 F.R.D. 497 (D. Kan. 2007). Both sides need the tier, because the defendant's technology is as sensitive as the plaintiff's.

Who sees the secrets: in-house counsel, the accused employee, and experts

Access disputes turn on whether the person is involved in "competitive decisionmaking," which the Federal Circuit defined in U.S. Steel Corp. v. United States, 730 F.2d 1465 (Fed. Cir. 1984), as advising on decisions such as pricing and product design made in light of similar information about a competitor. The question is the person's actual role, not the title. In Brown Bag Software v. Symantec Corp., 960 F.2d 1465 (9th Cir. 1992), the Ninth Circuit upheld an order denying a small company's in-house counsel access to the opponent's trade secrets because his work advising the company on contracts, marketing, and employment would put him in the "untenable position" of having to refuse his employer legal advice or risk revealing what he had learned; the order gave the company access through an independent consultant instead.

The accused employee is the hardest case. Denying the individual defendant access to the very information he is accused of taking can cripple his defense; granting it re-exposes the plaintiff. Courts typically place the burden on the plaintiff to show good cause for excluding him and then tailor access: supervised review at counsel's office, no copies, no screenshots. Independent experts are usually approved after advance disclosure of their identity, CV, and recent engagements, with a short window to object; disputes arise when the expert consults for competitors, and courts differ on whether that alone justifies exclusion. Compare Layne Christensen with Streck, Inc. v. Research & Diagnostic Systems, Inc., 250 F.R.D. 426 (D. Neb. 2008).

Prosecution bars, source code rooms, and clawbacks

When litigation counsel also prosecute patents, a protective order limited to "litigation counsel" offers little protection, because knowledge of the opponent's unpublished technology could shape claim drafting. In re Deutsche Bank Trust Co. Americas, 605 F.3d 1373 (Fed. Cir. 2010), holds that the party seeking a prosecution bar must show that the information designated to trigger the bar, the activities prohibited, the bar's duration, and the subject matter it covers "reasonably reflect the risk presented by the disclosure of proprietary competitive information," and that a lawyer seeking an exemption must show, counsel by counsel, that his or her patent office work does not implicate competitive decisionmaking related to the subject matter of the litigation. Trade secret cases with patent overlap should negotiate the bar at the outset.

Source code is usually produced only for inspection on a standalone, non-networked computer in a secure room, with sign-in logs, a ban on recording devices, limits on printed pages, and tight rules for using printouts in depositions and filings. Clawback protection belongs in every order. A Fed. R. Evid. 502(d) order provides that production in the case does not waive privilege in this or any other federal or state proceeding, which matters when forensic images of laptops sweep in communications with counsel. For drafting the full order, see Negotiating a Protective Order and Filing Under Seal: A Practical Guide and the Confidentiality Toolkit.

Sealing, public access, and in camera review

Protective orders govern what parties exchange; sealing governs what the court files. The standards differ sharply. There is a common-law right of access to judicial records, but the Supreme Court recognized in Nixon v. Warner Communications, Inc., 435 U.S. 589 (1978), that courts may refuse to let their files serve as "sources of business information that might harm a litigant's competitive standing." In the Ninth Circuit, records attached to dispositive motions may be sealed only for "compelling reasons," while a lesser good-cause showing suffices for non-dispositive motions. Kamakana v. City & County of Honolulu, 447 F.3d 1172 (9th Cir. 2006). The test turns on whether the motion is "more than tangentially related to the merits," not its label. Center for Auto Safety v. Chrysler Group, LLC, 809 F.3d 1092 (9th Cir. 2016). The Second Circuit treats summary judgment papers as judicial documents carrying a strong presumption of access. Lugosch v. Pyramid Co. of Onondaga, 435 F.3d 110 (2d Cir. 2006). And the Seventh Circuit has repeatedly refused to let parties seal by agreement, insisting on a document-by-document justification. Baxter International, Inc. v. Abbott Laboratories, 297 F.3d 544 (7th Cir. 2002); Citizens First National Bank of Princeton v. Cincinnati Insurance Co., 178 F.3d 943 (7th Cir. 1999). The parties' agreement never controls. Procter & Gamble Co. v. Bankers Trust Co., 78 F.3d 219 (6th Cir. 1996).

Genuine trade secrets nearly always clear the bar: the Ninth Circuit has said that the publication of materials that could result in infringement of trade secrets has long been considered a factor that overcomes the presumption of access. Apple Inc. v. Psystar Corp., 658 F.3d 1150 (9th Cir. 2011). What fails is overreach: sealing whole briefs, sealing information that is merely embarrassing, or sealing material already public. The working rules are to redact narrowly (the formula, not the paragraph about it), to file public redacted versions promptly, to use code names ("Trade Secret No. 4") in public filings and at trial, and to use the § 1835(b) procedure, a sealed submission explaining the owner's interest, before any order authorizes disclosure. When the court itself needs to see the secret, in camera review with a sealed transcript is the norm. A sealing ruling is not a merits ruling; courts generally keep contested material under seal until they decide whether it is a secret, because unsealing first would destroy the very thing the case is about, a point the Supreme Court made long ago about the nature of trade secret property. Ruckelshaus v. Monsanto Co., 467 U.S. 986 (1984). For the broader doctrine, see Protective Orders, Confidentiality Designations, and Sealing.

Forensic discovery of devices and accounts

Preservation comes first

The duty to preserve arises when litigation is reasonably anticipated, not when the complaint is served. Once it attaches, a party must suspend routine deletion and "must not destroy unique, relevant evidence that might be useful to an adversary." Zubulake v. UBS Warburg LLC, 220 F.R.D. 212 (S.D.N.Y. 2003). The Federal Circuit applies a "reasonably foreseeable" standard. Micron Technology, Inc. v. Rambus Inc., 645 F.3d 1311 (Fed. Cir. 2011). For a departing employee, the duty attaches at the latest when the former employer asserts its claim, see Cohn v. Guaranteed Rate, Inc., 318 F.R.D. 350 (N.D. Ill. 2016), and often earlier, because an employee who copies files on the way out knows exactly why someone might come looking.

Both sides have preservation problems. The plaintiff must stop the IT department from reimaging the departed employee's laptop (the single most common and most damaging self-inflicted wound in these cases), preserve server and cloud logs before rolling retention deletes them, and document chain of custody. The defendant, and the new employer, must preserve the employee's personal devices and accounts, which the employee may regard as nobody's business. Courts can and do enter preservation orders at the outset; in Henry Schein, Inc. v. Cook, 191 F. Supp. 3d 1072 (N.D. Cal. 2016), the court entered a temporary restraining order requiring a departing sales representative to preserve all data, including cloud storage, while refusing the employer's ex parte request to take "mirror images" of her personal devices, because that intrusion was not justified without giving her a chance to respond. For the mechanics, see Litigation Holds, Spoliation, and Rule 37(e) and the Litigation Hold and Evidence Preservation Checklist.

Neutral examiners and imaging protocols

A forensic image is a bit-for-bit copy of a storage device, including unallocated space where deleted files linger, verified by hash values so that the copy can be proven identical to the original. See A.M. Castle & Co. v. Byrne, 123 F. Supp. 3d 895 (S.D. Tex. 2015). Rule 34(a)(1)(A) allows a party to "inspect, copy, test, or sample" electronically stored information, but the 2006 advisory committee note cautions that the rule is not meant to create a routine right of direct access to an opponent's systems and that courts should guard against undue intrusiveness. Appellate courts have enforced that caution through mandamus. In In re Ford Motor Co., 345 F.3d 1315 (11th Cir. 2003), the Eleventh Circuit vacated an order giving a plaintiff unlimited direct access to Ford's databases, and in John B. v. Goetz, 531 F.3d 448 (6th Cir. 2008), the Sixth Circuit granted mandamus against an order requiring forensic imaging of state and personal computers where less intrusive means were available.

What justifies imaging is "something more" than suspicion: evidence that the device was used to take or transfer the information, or discrepancies and gaps in the other side's production. The leading model is the three-step protocol from Ameriwood Industries, Inc. v. Liberman, 2006 WL 3825291 (E.D. Mo. Dec. 27, 2006): a neutral expert images the devices and reports what was imaged; the expert recovers documents and provides them to the producing party's counsel; and that counsel reviews for privilege and responsiveness before producing to the requesting party, with a log of anything withheld. Genworth Financial Wealth Management, Inc. v. McMullan, 267 F.R.D. 443 (D. Conn. 2010), adopted that approach where former employees had taken copies of a client database and had not diligently preserved their devices, and ordered them to bear 80 percent of the neutral's cost because their conduct made the expert necessary; see also Playboy Enterprises, Inc. v. Welles, 60 F. Supp. 2d 1050 (S.D. Cal. 1999); Simon Property Group L.P. v. mySimon, Inc., 194 F.R.D. 639 (S.D. Ind. 2000); Covad Communications Co. v. Revonet, Inc., 258 F.R.D. 5 (D.D.C. 2009).

Modern protocols add three refinements. Hash searching lets the examiner look for exact copies of the plaintiff's files (identified by their cryptographic fingerprints) without anyone reading the employee's personal material. Artifact reports (USB connection logs, link files, cloud sync records, and deletion and wiping traces) often prove more than the documents themselves. And privacy screens (date limits, search-term limits, a privilege review by the producing side first, and a 502(d) order) answer the defendant's best objection. A good protocol also says who pays, who owns the images, and when they are destroyed. Parties should negotiate it alongside their ESI protocol; see Negotiating an ESI Protocol and Managing E-Discovery Costs.

Personal email and cloud accounts

The departing employee's personal Gmail or Dropbox account is often where the files went, and the plaintiff's instinct is to subpoena the provider. That usually fails. The Stored Communications Act, 18 U.S.C. § 2702, generally bars providers from disclosing the contents of communications in response to a civil subpoena, and courts quash such subpoenas. Crispin v. Christian Audigier, Inc., 717 F. Supp. 2d 965 (C.D. Cal. 2010). An overbroad subpoena can even expose the issuing party to liability. Theofel v. Farey-Jones, 359 F.3d 1066 (9th Cir. 2004). The workable route is a Rule 34 request to the party, who controls the account and can be required to produce from it or to consent to the provider's disclosure. Flagg v. City of Detroit, 252 F.R.D. 346 (E.D. Mich. 2008). Non-content records (login histories, IP addresses) are treated differently, and a subpoena for those may succeed. For the remote-work dimension of all this, see Trade Secrets in the Age of Remote Work and Cloud Computing.

The DTSA ex parte seizure remedy

Section 1836(b)(2) lets a court order, without notice, the seizure of property "necessary to prevent the propagation or dissemination of the trade secret," but only "in extraordinary circumstances" and only on detailed findings, including that a Rule 65 order would be inadequate because the defendant would evade it, that the applicant describes the matter to be seized with reasonable particularity, and that the defendant would destroy or hide the material if given notice. The order must provide for the narrowest seizure necessary, a hearing within seven days, security, and protection against publicity, and a person harmed by a wrongful or excessive seizure may sue for damages not limited by the bond.

Reported grants are few and share a pattern: prior lies, demonstrated technical sophistication, or efforts to hide data. Denials are more common, usually because a noticed TRO with a preservation order and a device turnover would do the job. Courts frequently order alternatives, such as delivering devices to the court or a neutral for safekeeping. The FJC published best practices for seizures in 2017, and its 2023 guide observes that very few seizure orders have been reported. In the ordinary departing-employee case, a noticed TRO with a forensic protocol accomplishes the same thing with far less risk. For the emergency-relief playbook, see Preliminary Injunctions and Temporary Restraining Orders.

Spoliation under Rule 37(e) in departing-employee cases

Rule 37(e) applies when electronically stored information that should have been preserved is lost because a party failed to take reasonable steps, and it cannot be restored or replaced. If the loss prejudices the other side, the court may order measures "no greater than necessary to cure the prejudice." Only on a finding of "intent to deprive" may it presume the information was unfavorable, give an adverse-inference instruction, or enter dismissal or default.

Trade secret cases produce the most dramatic examples. In Organik Kimya, San. ve Tic. A.S. v. International Trade Commission, 848 F.3d 994 (Fed. Cir. 2017), days before a court-ordered forensic inspection of a former Dow scientist's laptop, the respondent copied a program folder onto the drive at least 108 times to overwrite free space, backdated the computer's clock, ran a deletion utility repeatedly, and used a recovery tool to confirm that the deleted data could not be found. The Federal Circuit affirmed a default judgment and a 25-year exclusion order. Courts also impose large compensatory sanctions: the Second Circuit affirmed more than $2.5 million in sanctions for deletion and wiping in a counterfeiting case with roughly $20,000 at stake, holding that sanctions are measured by the cost of the misconduct, not the size of the claim. Klipsch Group, Inc. v. ePRO E-Commerce Ltd., 880 F.3d 620 (2d Cir. 2018). Fee sanctions under the court's inherent power must be compensatory and causally tied to the misconduct. Goodyear Tire & Rubber Co. v. Haeger, 581 U.S. 101 (2017).

Plaintiffs should not overplay spoliation. In Epic Systems Corp. v. Tata Consultancy Services Ltd., Nos. 19-1528 & 19-1613 (7th Cir. Aug. 20, 2020, amended Nov. 19, 2020), the district court gave an adverse-inference instruction because the defendant failed to preserve evidence, but the Seventh Circuit held that the inference, standing alone, could not support damages for use of "other information" beyond what the evidence showed: Epic sought "to use the adverse inference not just to plug evidentiary holes but to hold all the water." Spoliation levels the field; it does not replace proof.

Depositions, Rule 30(b)(6) topics, and experts

Depositions in trade secret cases test the identification from both directions. The defendant's Rule 30(b)(6) notice to the plaintiff will cover the development of each numbered secret, who knew it, what measures protected it, prior disclosures (patents, papers, trade shows, customer presentations), and the basis for believing it was taken. The plaintiff's notice to the defendant will cover the development timeline of the accused product, sources consulted, clean-room procedures, onboarding of the departed employee, and searches performed for the plaintiff's material. The FJC guide recommends handling contention questions (why the plaintiff contends a numbered item is secret, or how the defendant contends it developed the accused feature independently) through interrogatories under Rule 33 rather than through a corporate witness, and courts often limit or reshape 30(b)(6) topics accordingly. For the mechanics, see Preparing and Defending a Rule 30(b)(6) Deposition.

Experts do three jobs. A technical expert explains why each numbered item is or is not generally known and whether the defendant's product reflects it; in California, competing expert declarations on the adequacy of an identification are common, and Advanced Modular Sputtering tells courts not to resolve a credible disagreement at that stage. A forensic expert reconstructs what was copied, when, and where it went, often summarizing large volumes of log data under Fed. R. Evid. 1006. A damages expert ties harm to specific secrets. All three must satisfy Fed. R. Evid. 702 and the gatekeeping framework of Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993), and a technical expert who opines on "the trade secrets" without tying each opinion to a numbered secret invites exclusion. For deposing and defending these witnesses, see Expert Depositions: Taking Apart the Opinion, Protecting Your Own Expert, and Building the Daubert Record.

One more trap: forensic investigations commissioned before litigation may not stay privileged. In the Waymo litigation, Uber and the startup it acquired had jointly retained a forensic firm to investigate the departing engineer before the deal closed; the district court ordered the report produced, and the Federal Circuit dismissed the engineer's appeal and denied his petitions for mandamus. Waymo LLC v. Uber Technologies, Inc., 870 F.3d 1350 (Fed. Cir. 2017). Structure investigations through counsel, for the purpose of legal advice, and assume the facts they uncover will eventually be discoverable.

Strategy for each side

If you are the trade secret owner

• Investigate before you identify. Have your own forensic examiner reconstruct what left (USB logs, cloud syncs, email forwards) and build a hash list of the files that went out the door. Then triage: in most departing-employee cases a large share of the copied files are public, personal, or stale, and saying so up front buys credibility.
• Draft the identification before you file. Writing it forces you to decide what you actually own. If you cannot describe a secret in a paragraph that an engineer at the defendant could search against, you are not ready to sue on it.
• Separate the emergency ask from the identification. A temporary restraining order limited to preservation, return of identified files, and non-use of what was taken can rest on the files themselves; the FJC guide treats preservation and return orders as the narrow exception to full identification. An injunction against using "trade secrets" cannot, after Mallet.
• Start narrow and expand for good cause. A short list of strong, specific secrets beats a long list that invites an IDX or Sit-Up attack. Keep a record of when you learned each new fact so that you can show diligence if you later amend.
• Propose the protective order and the forensic protocol yourself. The party that drafts the first version frames the negotiation, and a protocol with real privacy protections is far more likely to be ordered quickly.
• Mind the forum. California state court applies § 2019.210 strictly; a federal court in the Ninth Circuit hearing only a DTSA claim may not require an up-front identification but will still manage discovery around one; the Seventh Circuit demands concrete secrets at summary judgment.

If you are the accused employee or the new employer

• Quarantine and document immediately. The new employer should isolate the employee from the relevant project, collect and preserve devices and accounts, and record what it did and when. Those steps are both a defense and the best evidence of good faith.
• Demand the identification, but do not overplay it. Ask for a written, numbered identification under a protective order before producing your own technology. Advanced Modular Sputtering and Quintara show that courts lose patience with defendants who use the demand as an indefinite stay.
• Build the independent-development record. Engineering notebooks, version-control history, design reviews, and the timeline of the accused product are the defense. Preserve them early; they are as much at risk of routine deletion as anything the plaintiff wants.
• Offer a privacy-protective forensic protocol. Hash-only searching by a neutral, with your counsel's privilege review first, is hard for a court to refuse and keeps the plaintiff out of the employee's personal life. Stonewalling invites an order like Genworth's, with the bill attached.
• Never touch the devices. Deleting, "cleaning," or reformatting after a demand letter turns a defensible case into Organik Kimya.
• Use the identification at summary judgment. Once discovery closes, the numbered list is the target: show, secret by secret, that each item is public, unused, or independently developed.

Costs and practical realities

Identification fights are expensive in a way that surprises clients, because they front-load expert work. In Advanced Modular Sputtering, three rounds of designations and dueling expert declarations consumed "thousands of dollars" in fees without a single document being produced. In Quintara, the identification dispute delayed discovery about five months. A well-drafted first identification, supported by a short declaration from someone who can explain it, is the cheapest insurance available.

Forensic work scales with the number of devices and accounts and with how adversarial the protocol is. Imaging a laptop and a phone and producing an artifact report is a modest expense; a neutral protocol covering multiple custodians, cloud accounts, and rounds of searching and privilege review can run into the tens or hundreds of thousands of dollars. Courts allocate those costs by agreement or by order, often shifting them to the party whose conduct made the examination necessary. Source code review adds its own costs: secure-room logistics, expert hourly rates, and printing disputes. The time horizon matters as much as the budget. The first thirty days decide whether the evidence survives; the identification fight usually occupies the next one to six months; and the case is often won or lost on the record built in those early months.

Common mistakes

1. Reimaging the departed employee's laptop. It is routine IT practice and the single most damaging thing a plaintiff can do to its own case.
2. Identifying by document dump. "The trade secrets are contained in the documents at JUN-000001 through JUN-013000" is Sit-Up, not an identification.
3. Describing functions instead of secrets. "Software that optimizes battery life" describes what the code does, the error the DeRubeis court flagged in 2007 and the Seventh Circuit punished in NEXT Payment Solutions.
4. Catchall language. "Including but not limited to" signals that the plaintiff does not know what it owns.
5. Filing the identification publicly. Beyond the obvious harm, it is evidence that the information is not secret (Loop AI Labs).
6. Claiming what you already published. Patents, papers, data sheets, and marketing claims are public; claiming them as secrets damages credibility on everything else.
7. Seeking an ex parte seizure by reflex. Courts expect a noticed TRO unless there is concrete evidence the defendant will evade an order.
8. Subpoenaing the email provider for contents. The Stored Communications Act will defeat it; use Rule 34 against the party.
9. Over-designating everything as attorneys' eyes only. It invites challenges, slows the case, and undermines the credibility of designations that matter.
10. For defendants, treating the identification demand as a stay. Courts increasingly let non-technical discovery proceed, and a defendant that delays may face a well-documented, carefully amended identification at summary judgment instead of a weak one.

A worked example: the departing engineer and the USB drive

Return to Juniper Labs. Its senior firmware engineer, here called Dana, left on a Friday for Kestrel Devices, a competitor, after plugging a 64 GB USB drive into her workstation; 2,900 files went with it.

Days 1 to 5. Juniper's general counsel stops IT from reimaging Dana's laptop and preserves the USB connection logs, the data-loss-prevention alerts, and the cloud sync history. Juniper's forensic examiner creates a hash list of the 2,900 files. Triage shows that about 2,100 are public datasheets, personal photos, or obsolete builds. The remainder cluster around three things: the secure-boot authentication sequence for Juniper's next controller, the power-management parameter tables tuned for a particular chipset, and two years of thermal test reports documenting approaches that failed. Juniper's counsel sends Dana and Kestrel a letter identifying the device and the categories of files, demanding preservation and return, and proposing a neutral forensic protocol. That letter fixes the latest date on which their preservation duty attached.

Days 6 to 20. Kestrel's counsel quarantines Dana from the controller project, collects her devices, and agrees to deliver them and the USB drive (if it can be found) to a neutral examiner. Because Kestrel cooperates, Juniper does not seek ex parte seizure; it files a DTSA and state UTSA complaint describing its secrets in general terms and moves for a stipulated order covering preservation, delivery of devices to the neutral, non-use of the copied files, and an interim two-tier protective order.

Day 30. Juniper serves its identification under the attorneys' eyes only tier: nine numbered secrets, each with a narrative description, the specific files that embody it identified by name and hash value, and no catchalls. Secret No. 3 is the negative know-how, identified as the specific heat-spreader materials and fan-curve approaches that failed Juniper's thermal tests, with the test reports cited. Secret No. 8, "the controller firmware architecture," is broader than the rest.

Days 45 to 90. The neutral's hash report finds 312 of the Juniper files on Dana's personal laptop, 41 of them opened after her start date at Kestrel, and a sync to a personal cloud account. Juniper does not subpoena the cloud provider for contents; it serves a Rule 34 request on Dana and obtains an order directing her to produce from the account. Kestrel moves to compel a more particular identification of Nos. 3 and 8. The court finds No. 3 adequate and orders No. 8 amended within twenty-one days. Juniper withdraws No. 8 rather than broaden its theory, and the list keeps its numbering.

Months 4 to 8. Under the source code protocol, Kestrel makes the relevant modules of its controller firmware available for review. Juniper's expert finds a power-management table whose structure and values match Juniper's Secret No. 2. Kestrel's 30(b)(6) witness testifies about its development timeline and produces version-control history showing that its secure-boot sequence predates Dana's arrival, which persuades Juniper to drop No. 1. Discovery of Kestrel's commit history also reveals that Dana committed a calibration routine resembling one in her copied files; Juniper moves to add it as Secret No. 10, showing good cause because it learned the facts only from Kestrel's production.

Summary judgment. Kestrel moves on every remaining secret. The court grants the motion as to the thermal negative know-how for lack of evidence of use and denies it as to Nos. 2 and 10. The case settles on terms that include a royalty, a certified destruction protocol overseen by the neutral, and a two-year bar on Dana's work on power management.

The supplier variation. Change the facts: Halvorsen Foods shares its batter process with Brightline, a contract manufacturer, under a nondisclosure agreement, and two years later Brightline launches a competing private-label product. The identification problem is different in one important way: Brightline has its own long-standing process know-how and other customers' confidential processes in the same plant records. Halvorsen's identification must separate what it disclosed from what any competent co-packer already knows, and the protective order must protect third-party customer information, usually through redaction and an additional confidentiality designation. Brightline's best evidence will be its own dated batch records from before the relationship, which it should preserve the day the dispute surfaces.

Frequently asked questions

Do I have to put my trade secrets in the complaint? No. Courts agree that a complaint need not disclose the secrets in detail, because doing so would destroy them; it must describe them in general terms sufficient to give notice and plausibly allege a claim under Bell Atlantic Corp. v. Twombly, 550 U.S. 544 (2007). The detailed identification comes later, under a protective order.

Can the defendant refuse all discovery until I identify my secrets? In California state court, discovery "relating to the trade secret" waits for the § 2019.210 identification. In federal court, it depends on the judge; the FJC guide recommends against a general stay, and many courts let non-technical discovery proceed while protecting the defendant's technology until a sufficient identification is served.

Can I add trade secrets later? Usually, if you show good cause: diligence, new facts learned in discovery, and no undue prejudice to the defendant. The later you ask, the harder it gets.

Will the employee I am suing get to see my identification? Possibly. Courts balance the employee's ability to defend against the risk of re-exposure, and often allow supervised access to the specific secrets the employee is accused of taking, without copies.

Can I subpoena Google or Microsoft for my former employee's personal email? Not for the contents. The Stored Communications Act bars providers from producing message contents in response to a civil subpoena. Serve a Rule 34 request on the employee, who controls the account.

Is an ex parte seizure order worth seeking? Rarely. It requires extraordinary circumstances and detailed findings, and a wrongful seizure exposes the applicant to damages. A noticed TRO with preservation, device turnover, and a neutral protocol usually achieves the same result.

What if the defendant wiped the laptop? Rule 37(e) allows curative measures upon prejudice and severe sanctions, up to default, upon a finding of intent to deprive. Prove the wiping forensically, but keep building the merits; an adverse inference supplements proof rather than replacing it.

Practical takeaways

• The identification is the spine of a trade secret case. Draft it early, number it, describe secrets rather than functions, and protect it under a two-tier order.
• Reasonable particularity means enough detail for the defendant to search its own files and for the court to set the boundaries of discovery, not proof of secrecy and not a patent claim.
• In federal court, expect sequencing but not a merits gate; Quintara bars striking secrets at the threshold, while NEXT Payment Solutions shows the price of vagueness at summary judgment.
• Amend for good cause, never renumber, and drop weak secrets freely.
• Preserve first, image second, and search by hash with privacy protections; use Rule 34 rather than provider subpoenas for personal accounts.
• Seek seizure only when evasion is shown; seek spoliation sanctions when deletion is proven, but do not ask an adverse inference to carry the whole case.

This document is general information, not legal advice for any particular situation; the law changes, and the facts matter.

Related Documents

Related Articles

• Trade Secret Damages and Remedies: Unjust Enrichment, Reasonable Royalties, Injunctions, and Exemplary Awards: how the numbered identification drives apportionment, head-start periods, and injunction scope.
• Expert Depositions: Taking Apart the Opinion, Protecting Your Own Expert, and Building the Daubert Record: deposing the technical and forensic experts who will testify about each secret.
• Protective Orders, Confidentiality Designations, and Sealing: Secrecy and Its Limits in Federal Court: the doctrine behind two-tier orders and sealing motions.
• Litigation Holds, Spoliation, and Rule 37(e): Preserving Electronic Evidence Before It Costs You the Case: preservation duties and sanctions in depth.
• Trade Secrets in the Age of Remote Work and Cloud Computing: where the evidence lives when employees work from personal devices and cloud accounts.
• Building a Trade Secret Protection Program from Scratch: the internal inventory that makes a litigation identification possible.

Related Guides

• Managing Discovery in a Trade Secret Case: A Practical Guide: the step-by-step companion to this article.
• Trade Secret Misappropriation Litigation Under the Defend Trade Secrets Act: A Practical Guide: the full life cycle of a DTSA case.
• Negotiating a Protective Order and Filing Under Seal: A Practical Guide: drafting and defending the order.
• Negotiating an ESI Protocol and Managing E-Discovery Costs: the companion protocol for document and data discovery.
• Preparing and Defending a Rule 30(b)(6) Deposition: topics, preparation, and objections for corporate witnesses.

Related Checklists

• Trade Secret Identification and Discovery Checklist: A Practical Checklist: every step from intake to summary judgment in checklist form.
• Employee Departure Trade Secret Protection Checklist: A Practical Checklist: what to do on the employee's last day so you have evidence later.
• Litigation Hold and Evidence Preservation Checklist: A Practical Checklist: preservation steps for both sides.
• Trade Secret Audit Checklist: A Practical Checklist: building the inventory before a dispute arises.

Related Toolkits

• Trade Secret Litigation Toolkit: Identification Statements, Two-Tier Protective Orders, and Forensic Protocols: model identification statements, protective order provisions, and a neutral forensic protocol.
• Trade Secret Remedies Toolkit: Damages Models, Head Start Analyses, and Injunction Briefing: templates for the remedies phase.
• Confidentiality Toolkit: Protective Orders, Designation Challenges, and Sealing Motions: sealing motions and designation challenges.
• Trade Secret Protection Toolkit: Building and Defending a Confidentiality Program: the reasonable-measures record that every identification depends on.