Document Correlation Under Fed.R.Civ.P. 34(b)(2)(E)

[EDRM Editor’s Note: The opinions and positions are those of Michael Berman.]

The long simmering “document correlation” dispute was addressed in Sievert Elec. Svc. and Sales Co. v. Storako, 2025 WL 2084229 (N.D. Ill. Jul. 24, 2025).  The court followed the majority rule. 

While there is a clear split of authority, I respectfully disagree with Sievert and suggest that there is a better mousetrap.  Several years ago, I wrote a blog titled Does Rule 34(b)(2)(E)(i) Mandate “Document Correlation” When ESI is Produced and, If So, Does Metadata Provide a “Work-Around” to Avoid the Costs of That Correlation Process? – E-Discovery LLC (Jul. 20, 2021)(the “2021 Blog”):

“Document correlation” is the requirement that a producing party label produced documents to correspond to categories in the request for production of documents.  Doing so with voluminous ESI can be exceedingly costly. [Emphasis added].

Michael Berman, Does Rule 34(b)(2)(E)(i) Mandate “Document Correlation” When ESI is Produced and, If So, Does Metadata Provide a “Work-Around” to Avoid the Costs of That Correlation Process? – E-Discovery LLC (Jul. 20, 2021).

The starting point is Fed.R.Civ.P. 34(b)(2)(E).  In pertinent part, Rule 34(b)(2)(E) provides:

(E) Producing the Documents or Electronically Stored Information. Unless otherwise stipulated or ordered by the court, these procedures apply to producing documents or electronically stored information:

(i)  A party must produce documents as they are kept in the usual course of business or must organize and label them to correspond to the categories in the request;

(ii)  If a request does not specify a form for producing electronically stored information, a party must produce it in a form or forms in which it is ordinarily maintained or in a reasonably usable form or forms…. [Emphasis added].

Fed.R.Civ.P. 34(b)(2)(E).

The Sievert court described the two schools of thought regarding (E)(i) and (E)(ii):

• “The majority view is that Rule 34(b)(2)(E)(i)’s disjunctive organization requirement exists alongside Rule 34(b)(2)(E)(ii)’s separate ESI formatting requirements and applies to both ESI and hard copy documents.”
• “But some courts have held that the organizational requirements of Rule 34(b)(2)(E)(i) do not apply to ESI productions, and a road map is not required for ESI productions under Rule 34.”¹

The Sievert court adopted the majority view, writing:

Such a reading comports with the Advisory Committee’s goal in rewriting Rule 34(b): to enact protections—like those adopted in 1980 for paper productions—that prevent parties from producing ESI in a way that creates obstacles for the receiving party. See Fed. R. Civ. P. 34(b), Advisory Committee’s note to 2006 amendment (“[ESI production] should be subject to comparable requirements to protect against deliberate or inadvertent production in ways that raise unnecessary obstacles for the requesting party. Rule 34(b) is amended to ensure similar protection for [ESI].”). It is axiomatic that discovery responses themselves have significant evidentiary value as they constitute party admissions. But without a clear connection between the discovery requests and the corresponding responses, the discovery responses cannot be deemed reliable.

Sievert Elec. Svc. and Sales Co. v. Storako, 2025 WL 2084229 (N.D. Ill. Jul. 24, 2025).

The Sievert court rejected the producing party’s “it wasn’t requested” argument:

Integrated’s secondary argument that Sievert is not entitled to this organizational information because it did not request it in its RTP instructions is without merit. Rule 34(b)(2)(E)(i) and (ii) sets forth organizational and formatting requirements. A litigant need not instruct the opposing side to comply with the Federal Rules of Civil Procedure when responding to discovery requests. Accordingly, Integrated must ensure its ESI production corresponds to specific document requests.

Id.

I agree that there is merit to the Sievert analysis.  After all, many courts have adopted it.

Sievert pointed to a rationale for the minority rule – it is that ESI is treated differently because it is generally text-searchable.  Id. at *3.  I suggest that this is an oversimplification.

In the 2021 Blog, I suggested some additional concepts.

Rule 34(b)(2)(E)(“[u]nless otherwise stipulated”) permits stipulations altering its terms, and under Fed.R.Civ.P. 29, the parties can stipulate out of the Rule 34(b)(2)(E) requirements in a discovery plan or ESI Protocol.²

There may be good reasons to do so.  As authorities cited in the 2021 Blog point out, ESI is not stored like paper documents.  Thus, it may be difficult to produce ESI as it is kept in the usual course of business.  In fact, standard near-de-duplication changes how ESI is kept in the usual course.  And, it may be unduly burdensome and disproportionate to require document correlation for thousands of emails, text messages, or chats.

I suggest a different approach.  Fed.R.Civ.P. 34(b)(1)(C) permits a requesting party to specify the form or forms of production.  A requesting party may specify requested metadata fields.  If there is a dispute, the producing party must state how it “intends” to produce the ESI, and discussions should follow.

In my opinion, as I wrote in the 2021 Blog:

If ESI is produced with appropriate metadata, it likely makes no difference whether subsection (E)(i) or (E)(ii) governs.  The produced ESI can be loaded into a litigation review platform, such as Digital War Room, Relativity, Reveal, or others.  At that point, the receiving party will have the information and ability to determine how the documents were kept “in the usual course of business” by sorting by custodian, folder, file path, sender, recipients, or dates. 

Michael Berman, Does Rule 34(b)(2)(E)(i) Mandate “Document Correlation” When ESI is Produced and, If So, Does Metadata Provide a “Work-Around” to Avoid the Costs of That Correlation Process? – E-Discovery LLC (Jul. 20, 2021).

In short, I suggest that production of typical metadata fields is the efficient, effective, and proportionate solution and it renders document correlation unnecessary.

For example, if appropriate metadata is provided, the receiving party can determine what was stored in Jane Smith’s folders named “Widget Sale to ABC” or “Hot Docs”  It can easily ascertain what documents were authored, or email sent by or to, John Jones.  And, it can quickly locate all documents stored in c:\users\jones\Documents\Transaction_w_ABC Corp.  That easily available functionality produces the “documents as they are kept in the usual course of business,” for purposes of (E)(i), and “produce[s] it in a form or forms in which it is ordinarily maintained,” for purposes of (E)(ii).

Michael Berman, Does Rule 34(b)(2)(E)(i) Mandate “Document Correlation” When ESI is Produced and, If So, Does Metadata Provide a “Work-Around” to Avoid the Costs of That Correlation Process? – E-Discovery LLC (Jul. 20, 2021).

In fact, one court wrote that: “The nimbleness of current search functionality with ESI software is the very reason the Advisory Committee found it unnecessary to make (E)(i)’s organization guarantees applicable to ESI.” Id.

In short and in my opinion, producing standard metadata fields such as custodian and folder information, is producing the ESI as it was stored and correlation is a disproportionate, redundant exercise.  In my template ESI Protocol/Discovery Plan, I add: “The parties opt out of document correlation under Fed.R.Civ.P. 34(b)(2)(E)(i) based on production of agreed metadata fields, including without limitation the custodian field.”

The Sievert court apparently viewed issue tagging as possibly sufficient to provide document correlation: “Integrated asks the court to compel Sievert to produce its supplemental production in a manner that identifies the documents responsive to each request as required under Rule 34(b)(2)(E)(i) and (ii)—the same relief Sievert seeks in its motion against Integrated—…. In response, Sievert represents that the ‘relevancy reason’ tags, which had been ‘inadvertently excluded,’ were included in each document… after Integrated filed its motion…. Accordingly, these issues appear to be moot. To the extent Sievert’s supplemental production does not correspond to each request, Sievert is ordered to cure this deficiency.”³

Notes

1. The court’s extensive citations are omitted. ↩︎
2. For the differences between and “ESI Protocol” and a “discovery plan,” see “ESI Protocol” v. “Discovery Plan” (Jan. 2, 2024). ↩︎
3. On different issues, the Sievert court wrote that: (1) “The attorney-client privilege protects communications between a client and his lawyer—not documents related to an investigation a client orders a third-party computer forensics professional to perform—and does not apply here.”  The record was devoid of any indication that Sievert’s attorneys requested a forensic investigation in anticipation of litigation.  On the facts presented, it was not work product either; (2) while “discovery on discovery” is “generally disfavored,” it is “permitted where the requesting party proves a specific discovery deficiency in the responding party’s production”; and, (3) a mere allegation of “self-collection” is insufficient to show a discovery deficiency.  Id. at *5-6 (cleaned up).  The Sievert court ruled: “Because Integrated did not prove the existence of a discovery deficiency, the court will not order Sievert to disclose its ESI search protocol.”  However, it ordered a re-search by the producing party. ↩︎

Assisted by GAI and LLM Technologies per EDRM GAI and LLM Policy.