What is the Purpose of an ESI Protocol? Court Addresses Seven Key Issues

The purpose of an ESI Protocol “is to encourage reasonable electronic discovery with the goal of limiting the cost, burden and time spent, while ensuring that information subject to discovery is preserved and produced to allow for fair adjudication of the merits.” Leon v. American Honda Motor Co. Inc., 2025 WL 3255005, at *3 (C.D. Cal. Nov. 7, 2025)(citation omitted).  “ESI Protocols “also promote, when ripe, the early resolution of disputes regarding the discovery of ESI without Court intervention.” Id. “Indeed, ‘a robust ESI protocol will aid the parties in identifying discoverable ESI, thus helping [them] complete discovery in this case by [the applicable] discovery deadline.’” Id. (citations omitted).

In Leon, the plaintiffs filed a “Motion to Compel re Discovery Disputes on ESI Protocol.” While both parties agreed that an ESI Protocol was necessary, the parties filed a stipulation describing seven issues on which they could not agree. The case was a putative nationwide class action alleging an undisclosed allegedly defective sliding door sensor.

This blog discusses six of the seven ESI Protocol issues resolved in Leon:

• Issue 1: Production of Known Responsive Documents Outside of the Search Terms
• Issue 3: Keyword Construction and Relevant Technology
• Issue 4: Hit Reports
• Issue 5: Validation of Search Methodology
• Issue 6: Family-Inclusive Production
• Issue 7: Relevancy Redactions

Initially, however, the Leon court drew a distinction–one that I have also suggested–between an “ESI Protocol” and a “Discovery Plan.” Id. at *3 (stating that a Rule 26(f)(3)(C) discovery plan must include “any issues about disclosure, discovery, or preservation of electronically stored information, including the form or forms in which it should be produced”); An “ESI Protocol” is Not a Rule 26(f) “Discovery Plan” (Nov. 24, 2025); “ESI Protocol” v. “Discovery Plan” (Jan. 2, 2024); Parties Ordered to Cooperate in Good Faith and Develop a “Discovery Plan” (May 25, 2024).

Issue 1: Production of Known Responsive Documents Outside of the Search Terms

Here, plaintiffs said (in my less elegant words), “if you know of a responsive document that the search terms do not hit, you need to produce it even though the terms did not retrieve it.” Defendant responded that this suggestion improperly added terms to the ESI Protocol.

The court quickly overruled defendant’s objection, writing: “But Defendant has not articulated any burden or prejudice from having this provision in the Protocol. And this language reinforces the purpose of an ESI Protocol—ensuring that all responsive and relevant ESI is produced without Court intervention.”

For more on the question of “if a litigant has actual knowledge that an opponent’s proposed search terms will not ‘hit’ on responsive, proportional ESI, does the litigant have a duty to tell the opponent,” see Is There a Duty to “’Fess Up?” – Part II (Nov. 12, 2022).  I suggest, and the Leon court held, that the answer is “yes.” The theory is that an ESI Protocol and the litigant’s obligations under the rules of procedure work in tandem. One does not replace the other.

Often, the converse issue is present, i.e., does a party have to produce every unprivileged document on which a search term hits, regardless of responsiveness?  The answer is generally “no.” Does Every “Hit” on a “Hit Report” Have to be Produced? (Nov. 13, 2024); How to Avoid Contentious “Hit Report” Problems – Part III (Nov. 15, 2023); 8 Lessons Learned – Part II – “Hit” Report Provision of ESI Protocol (May 7, 2024); cf. Agreement in ESI Protocol to Produce All “Hits,” Without Review (Mar. 4, 2025).

Issue 3: Keyword Construction and Relevant Technology

Issue 3(A): Use of “Or” Connectors

Plaintiff contended that use of an “OR” connector obscured the search terms’ individual performance.

They argue that if Defendant is allowed to compound “multiple ‘OR’ terms into a single string …—for example, ‘defect AND (door OR Honda)’—it becomes impossible to tell whether the hits are being driven by ‘door,’ ‘Honda,’ or both.”

Defendant objected and asserted there was “no good reason to limit the types of search terms the parties can agree to in the ESI protocol,” but admitted there was  only a “marginal” increase in burden.  Its objection was overruled.

Issue 3(B): Disclosure of Internal Terminology

The court described the dispute:

Plaintiffs request that Defendant identify the internal acronyms, abbreviations, nicknames, and project code names it uses to refer to the subject matter of this litigation…. They argue that their language “ensures that keyword searches—one of the agreed proxies for relevance—are run in a transparent, reliable, and complete manner.” ….

Defendant agrees that its counsel will need to consult with Defendant’s employees and cooperate with Plaintiffs’ counsel to come up with appropriate keywords to use when searching for responsive documents…. But Defendant asserts requiring its counsel “to prepare such a list would violate the work product doctrine because it would disclose … counsel’s mental impressions and determinations of what items are important to the case.”

The Leon court agreed with defendant: “Cooperation among counsel to determine keywords is materially different than requiring Defendant’s counsel to unilaterally identify keywords.” Defendant’s objection was granted.

For more information, see Golden Oldie #6 – Four Historical Search Decisions (Apr. 21, 2025). In Wm. A. Gross Construction Assoc., Inc. v. Amer. Mfrs. Mut. Ins. Co., 256 F.R.D. 134 (S.D.N.Y. 2009), the Hon. Andrew Peck (ret.)  issued a wake-up call to the Bar and stated that lawyers cannot design searches by the seat of their pants.  Judge Peck pointed to the need for quality control to test for false negatives and false positives.  His Honor also pointed to the need to speak with the authors of the ESI to design search terms.

Issue 4: Hit Reports

“Plaintiffs’ proposed language requires that when a party objects to a proposed search term as overbroad or disproportionate, it must produce a basic ‘hit report’ showing how many documents that term retrieves.”

Defendant responded that it need only provide hit reports on request. It argued: “That is a reasonable procedure that avoids the expense of sweeping efforts to gather, import and store data for every single search term.”

The court wrote that Defendant failed to quantify the cost of having to run a hit report.  In my experience, the cost is nominal and perhaps that explains the failure to quantify.

The court found that failure to be fatal, writing: “Thus, any objection to a proposed search term as overbroad or disproportionate must be accompanied by a hit report so that the other party—and the Court—can assess the burden. Making such an objection without the hit report invites disputes and Court intervention, which an ESI Protocol is designed to avoid.” Defendant’s objection was overruled.

For more information on “hit reports,” see Does Every “Hit” on a “Hit Report” Have to be Produced? (Nov. 13, 2024); Keyword Hits + Overbroad Terms ≠ Duty to Produce (Sep. 9, 2024); 8 Lessons Learned – Part II – “Hit” Report Provision of ESI Protocol (May 7, 2024); How to Avoid Contentious “Hit Report” Problems – Part III (Nov. 15, 2023).

Issue 5: Validation of Search Methodology

Plaintiff demanded that quality control measures be applied to the result of keyword searches.  Specifically, it demanded a random sampling of the null set. In my experience, that process is well-accepted – in fact, it is generally required as a matter of prudence.

Defendant agreed that “quality control measures, including random sampling of null sets, should [ ] be used to evaluate keyword searches.”  However, it made what I view as a curious objection in an age of cooperation and transparency:

But it “objects to Plaintiff’s attempt to take over the process.” … At the hearing, counsel argued that requiring Defendant to turn over nonprivileged null set documents violated Rule 26 because it necessarily includes documents nonresponsive to the search terms.

The court overruled that objection, writing: “But the point of the null set evaluation is to ensure that the search terms are capturing all responsive, relevant information. And Defendant does not quantify the burden of producing a small random sample of the null set documents to Plaintiffs.”  As an aside, I suggest that a null set evaluation need not require production of nonresponsive documents – disclosure of metrics may suffice.

Validation protocols are necessary.  Court Resolves Disputes Over Number of Custodians and Validation Protocol (Aug. 6, 2024); see generally Historical ESI Highlights – Part VI – Victor Stanley and Keyword Searching (May 29, 2022).  Cases citing the need for quality control are discussed in Golden Oldie #6 – Four Historical Search Decisions (Apr. 21, 2025).

Issue 6: Family-Inclusive Production

Plaintiff suggested that entire “families” must be produced together, subject to withholding of privileged information.  “Defendant objects that the [proposed ESI Protocol] language would require production of irrelevant documents.”

“While the Court appreciates that the case law is split on this issue, the Court finds that the efficient and fair approach is to treat emails and their attachments as a single document.”  That will allow “Plaintiffs—and the Court when appropriate—to review the document family as it was created and maintained.”  Further, breaking up families would, it wrote, be an impermissible “relevance redaction.”

The parent-child question may also have impacts on privilege decisions.  See “Attachments to Attorney-Client Communications May Be Withheld as Privileged, Without an Independent Basis for Privilege,” But…. (Aug. 8, 2024).

I have written several blogs on the question of “what is a document” in the digital age. E.g., What is a “Document?”: Interior Email Omitted from Email Chain – Sanctions Follow for Lack of Candor to Court (Sep. 11, 2025). When it comes to ESI, that is a good question to ask.  See:

• Sedona Conference Commentary on Discovery of Collaboration Platforms – What is a Document? – E-Discovery LLC;
• Uber Technologies – Another Hyperlink Decision – E-Discovery LLC;
• “Modern Attachments” or “Pointers”- What is a Document? (Part IV) – E-Discovery LLC;
• What is a Document? (Part III) – E-Discovery LLC;
• What is a Document? (Part II) – E-Discovery LLC.

The parent-child issue in Leon does not appear to have been one of “modern attachments,” or hyperlinks. See, e.g., Uber Technologies – Another Hyperlink Decision (Mar. 6, 2025); 8 Lessons Learned – Part I – The “Humpty Dumpty” Clause of ESI Protocol (May 6, 2024); More on “Modern Attachments,” “Pointers,” or Hyperlinked Documents – Humpty Dumpty and “Usability” (Nov. 13, 2023); ESI Protocol Dispute – “Modern Attachments” and the “Humpty Dumpty Issue” (Apr. 29, 2024); “Modern Attachments,” ESI Protocols, & Second Chances (Jun. 8, 2023);  What Hath Noom Wrought? (Apr. 25, 2023);  “Modern Attachments” or “Pointers”- What is a Document? (Part IV) (Aug. 12, 2022).

Issue 7: Relevancy Redactions

Defendant sought the right to redact parts of documents that contain irrelevant sensitive business information.  The Leon court wrote: “But courts in this Circuit typically conclude that the unilateral redaction of documents on the ground that the producing party deems the information irrelevant is generally improper.” It explained the reasons for that rule. Id. at *6.

I have written several blogs on relevance redactions.  E.g., ESI Protocol Should Define “Documents” and Address Redaction Based on Irrelevancy; (May 14, 2025); Relevance Redactions Prohibited – Again (May 17, 2024); 8 Lessons Learned – Part III – Redaction Provision of ESI Protocol (May 8, 2024). They are generally impermissible.

Leon addresses a number of important ESI Protocol issues.