What Hath Noom Wrought?

E-Discovery LLC logo

The “modern attachments” decision of Nichols v. Noom, Inc., 2021 WL 948646 (S.D.N.Y. Mar. 11, 2021), generated a lot of buzz when it was issued.  

At its core, Noom held that hyperlinked files are generally not part of a transmitting email.  Said differently, when a party produces an email that contains a pointer to a document on a storage platform such as OneDrive or Google Drive, under Noom, the target file generally need not be produced as a family with the email.  

Noom has been followed; however, it has also been criticized on procedural, evidentiary, logical, and technological grounds.  I pointed to evidentiary concerns in What is a “Document?” Semantic and other issues were discussed in “Modern Attachments” or “Pointers”- What is a Document? (Part IV).  

Recently, however, an email from a colleague prompted me to look at both the precedents and progeny of Noom.  That leads me to suggest that there is no “one size fits all” solution to hyperlinks, “modern attachments,” or pointers.  

Instead, production of “modern attachments” is a topic that merits pre-production discussions that consider several factors.  First, of course, proportionality – cost and time – is always a consideration.  Second, custodians and the people who generated potentially responsive ESI should be interviewed and asked how they used their ESI so that discussions are informed by facts.[1]  Third, technology – both how ESI is created and stored, and also how it can be collected and produced – must be considered.  Fourth, context and the needs of the case should be discussed.  For example, if the target file is not produced as a family with the parent, will the context of a significant communication become unavailable? Fifth, one specific question is, if the targeted link is not produced as a family, can a foundation can be laid so that either the parent or the target, or both, will be admissible?  In the words of the Hon. Paul W. Grimm, “considering the significant costs associated with discovery of ESI, it makes little sense to go to all the bother and expense to get electronic information only to have it excluded from evidence or rejected from consideration during summary judgment because the proponent cannot lay a sufficient foundation to get it admitted.”  Lorraine v. Markel Amer. Ins. Co., 241 F.R.D. 534, 538 (D. Md. 2007).[2] 

Noom ended in a settlement on July 12, 2022, according to the 519th docket entry.  Dkt. 1:20-CV-03677.

A SHORT DISCUSSION OF NOOM

In Noom, the defendant used Gmail.  During discovery, the plaintiff learned that hyperlinks to Google Drive were used in lieu of attachments to many emails.  Plaintiff asked the court to order re-collection using MetaSpike’s Forensic Evidence Collector in order to pull the hyperlinked documents as part of the “family.”[3]  The defendant responded that: (1) the linked documents are not attachments; (2) it was collecting the responsive documents separately; and, (3) the requested tool was unworkable, costly, and would lead to delay. 

The court first looked at the parties’ negotiated ESI protocol.  It wrote that the protocol: (1) did not state that hyperlinked documents were part of “family groups”; and, (2) did not define “attachments.”  The Noom court added: “It is clear to this Court that there was no meeting of the minds on whether hyperlinks were attachments….”  Plaintiff, however, stated that neither party anticipated that there would be thousands of linked Noom documents.

The Noom court wrote that the dispute “highlights the changing nature of how documents are stored and should be collected.”  The court permitted the use of Google Vault as a collection tool, even though it did not collect the linked files.  The court did “not agree that a hyperlinked document is an attachment.”  Id. at *4 (emphasis added).  It wrote that the hyperlinked document “may or may not be necessary to the communication.”  As an example, it wrote that hyperlinked cases in a legal memo are not part of the memorandum.  Further, a hyperlink may point to an entire folder, a tracking site for packages, or a phone number.  In my words, the court was considering “context.”

The court implied that plaintiff may have asked for too much, writing: “Moreover, Plaintiffs have not made a showing that the value of obtaining the metadata establishing the linkages for all hyperlinked documents is proportional to the needs of the case.” Id. at *5 (emphasis in original).  

A more focused request may have led to a different result.  The court wrote that it “held that if there were certain documents discovered in the production containing hyperlinks for which the corresponding hyperlinked document could not be located or identified, Plaintiffs could raise the issue with Noom and Noom would be required to provide the document or Bates number.”  The court explained that “Noom was, from the start, willing to do this for a reasonable number of documents.”

The Noom court relied on three prior decisions.  None addressed hyperlinks. 

SOME DECISIONS THAT PRECEDED NOOM ARE CONTRADICTORY

The Noom decision did not cite IQVIA Inc. v. Veeva Systems, Inc., 2019 WL 3069203 (D.N.J. Jul. 11, 2019)(order and opinion of special master), which reached the opposite conclusion.  

In pertinent part, the IQVIA court wrote that “IQVIA explains that Veeva has also produced over 2,200 documents that purport to link to or otherwise reference documents from the Google Drive, without actually attaching the documents from the Google Drive….  IQVIA argues that this is akin to producing emails without attachments, which federal courts have disallowed, even when relinking an email to its attachment involves a tedious manual process.” 

The producing party was Veeva. The court explained:

With respect to linked documents, Veeva argues that IQVIA’s contentions that Veeva must produce all documents referenced in emails has no legal basis and misunderstands how email and the internet work. Veeva argues that the law does not require it to produce the targets of Google Drive links contained in already produced emails. Veeva maintains that courts require that attachments to emails be produced with their corresponding emails because that is how they are kept in the usual course of business. Veeva argues that that scenario is distinguishable and not akin to producing emails with Google Drive links that outsiders cannot access. Veeva explains that Veeva’s email servers do not archive a copy of the target of a Google Drive link in the email and that doing so may be impossible as Google Drive links point not only to documents but also to folder structures.

Id. at *3 (emphasis added).

The IQVIA court described the discovering party’s responsive argument:

IQVIA points to Pom Wonderful LLC v. Coca-Cola Co., 2009 WL 10655335 (C.D. Cal. Nov. 30, 2009), to argue that courts require emails and attachments to be produced together even if that is not how they are stored in the usual course of business. IQVIA argues that it does not have the ability to identify which documents from [Veeva’s] Google Drive belong with the emails linking to such documents, only Veeva is able to link the emails to the documents from Google Drive.  

Id. at *4.

The special master’s order in IQVIA stated:

With respect to the 2,200 linked Google Drive documents referenced in emails, the Special Master finds Pom Wonderful LLC v. Coca-Cola Co.2009 WL 10655335 (C.D. Cal. Nov. 30, 2009) instructive. Like in Pom Wonderful LLC, there is no dispute that the linked documents are relevant to the claims and or defenses of this action. While, as Veeva argues, the linked documents are not stored with emails in the ordinary course of business, IQVIA has no way to link the documents, only Veeva is capable of linking the emails to the Google Drive documents. The Special Master is not convinced that relinking these 2,200 documents is unduly burdensome in light of the issues at stake in this matter, the resources of the parties, and the amount in controversy. Accordingly, to the extent possible, Veeva is ordered to link the Google Dive [sic] documents referenced in emails within thirty days of the date of this Order.

Id. at *5 (emphasis added).[4]

The Pom Wonderful decision cited in IQVIA is also instructive.  There, Coca-Cola produced relevant attachments for 138 emails that were not linked to an email.  It “claims that it has no legal obligation to re-link the emails and their attachments because it produced them in the usual course of business.” 2009 WL 10655335, at *2.  It asserted that re-linking was a “tedious manual process.” Id. at n. 1.  

The Pom Wonderful court wrote that “plaintiff must have the ability to identify which attachments belong to which emails. There is no dispute that plaintiff does not have that ability.”

The Pom Wonderful court looked to two precedents.  

First, “given the Sedona Conference’s view of an email and attachment as one document or message unit, it is clear that [Coca-Cola] is required to provide plaintiff with either the ability to re-link the emails with the attachments or re-produce the emails with their attachments.”  Id. at *3.

Second, the court looked at a series of cases that prohibit dumbing down documents.  One case stated that the producing party’s software configuration “does not provide Defendants with an excuse to produce the e-mails and attachments in a jumbled, disorganized fashion.”  Another refused to permit a party to shield itself from discovery by using “a system of record-keeping which conceals rather than discloses relevant records, or makes it unduly difficult to identify or locate them, thus rendering the production of documents an excessively burdensome and costly expenditure.”  Several others refused to permit use of “an outmoded filing system” as a reason for inadequate responses.  

Finding that undue burden had not been demonstrated by the generalized objection, the Pom Wonderful court held that Coca-Cola “cannot seek to preclude plaintiff from pursuing discovery based on a record-keeping system that is plainly inadequate.”

Many years ago, I litigated a commercial case on behalf of a large corporation.  Faced with discovery requests, I told a Magistrate Judge that the company’s documents were stored in a multi-floor building that was the size of a square block.  I said that the client did not have an index to its documents.  The Judge replied with smile:  “There will be one now.”

SOME POST-NOOM DECISIONS

Noom was followed, and IQVIA was distinguished, in Porter v. Equinox Holdings, Inc., 2022 WL 887242 (Cal. Super. Mar. 17, 2022).  The court wrote: “The parties take opposing positions on the issue of whether Equinox must produce linked documents in family relationships with the emails that reference those linked documents.”  

Defendant Equinox used SharePoint and Google Drive.  Plaintiffs argued that linked documents are analogous to traditional attachments and they requested that defendant “produce in family relationships all linked documents referenced in relevant emails….”  Plaintiffs cited IQVIA.

Unsurprisingly, citing Noom, the defendant took the exact opposite position.

The Porter court explained that discoverability is distinct from whether documents must be produced as linked.  In essence, it viewed the scope of discovery as different from the form of production.

Porter then looked at the technology:  “For a party responding to discovery like Equinox in this instance, producing linked documents in family relationships requires Equinox to [identify] the linked document at issue, determine the cloud site where the document is stored, and then match up the linked document with the communication—be it an email, a text message, or a chat string—that references the linked document.”  

In Porter, the California trial court continued, writing that, while “[a]t first glance” that re-linking process may appear “fairly straightforward,” it presents “unique challenges” that differentiate it from email attachments.  It wrote:

For example, a responding party may not be able to collect the precise linked document referenced in a message if the document has been modified or deleted. In some instances, users have made linked documents inaccessible by revoking access to files, thus disabling the responding party’s ability to collect that information.  Finally, linked documents can also create review challenges and inefficiencies given the complexity of connecting those documents to the communications which reference them.  For these reasons, the Court finds that IQVIA, Inc. v. Veeva Systems—which essentially held that linked documents were comparable to attachments—is inapposite in this instance.

The Porter court then issued a narrow order:

  • “Given the potential technological challenges with producing linked documents in family relationships, the Court declines to order at this time that Equinox produce all linked documents referenced in relevant communications in family relationships.” [emphasis in original]; and,
  • “Instead, the Court will follow the procedure from Nichols v. Noom to address Plaintiffs’ concerns regarding the linked documents. In particular, if there are key linked documents referenced in Equinox-produced communications that Plaintiffs cannot find, Plaintiffs shall notify Equinox accordingly and identify the particular communications referencing the linked documents in question. In response, Equinox shall identify the linked document at issue by Bates number or, if it has not been produced, Equinox shall produce the linked document without delay.”

Disputes were to be referred to the “Expert Advisor” and the court retained jurisdiction to modify the order for good cause shown.

In another recent case, Stitch Editing, the parties seem to have listened to Noom’s admonition to discuss linked documents and focus their request.  That process led to an order to produce the linked documents.  In their memorandum, plaintiffs wrote that they “provided a comprehensive chart showing the relevance of the limited number of links to the development of the Stitch Editing Feature.”  They added: “It is unclear how reviewing a few dozen links is ‘unduly burdensome’ to a billion dollar corporation.”  And, they asserted that:  “The ESI protocol allows for the production of links, and the links meet the threshold of FRCP 26(b)(1).”  See Plaintiff’s Brief in Stitch Editing Ltd. v. TikTok, Inc., 2022 WL 18937342, at §B (emphasis added).

The Stitch Editing court apparently agreed. It ordered that: “Documents referenced in hyperlinks from all productions—existing and future—must be produced together with the source document containing the hyperlinks so that the association between parent document and hyperlinked document is maintained.”  Stitch Editing Ltd. v. TikTok, Inc.2022 WL 17363054, at *1 (C.D. Cal. Aug. 31, 2022).[5]

For another example of a post-Noom ESI protocol addressing the Noom issue, see In re Acetaminophen – ASD-ADHD Prod. Liab. Litig., 2023 WL 196157, at *5 (S.D.N.Y. Jan. 17, 2023)(“Email attachments and embedded files or ‘modern attachments’ (i.e., hyperlinks pointing to files stored in the cloud or a shared repository such as SharePoint and other types of collaborative data sources, instead of being directly attached to a message as has been historically common with email communications) shall be collected and produced with the parent message.”).

An attempt to expand Noom was briefed in another case that settled in October 2022.  ECF No. 186, Bedford, Freeman & Worth Publishing Grp., LL v. Shopify, Inc., No. 1:21-CV-01340 (E.D. Va.).  There, at least according to the briefing, the issue was a bit different.  It did not involve hyperlinked emails.  

According to the defendant, the plaintiff sought documents hyperlinked in defendant’s policies and manuals.  See Deft’s Memo, 2021 WL 10396835 at ¶2.  The defendant argued that “a wholesale request for ‘all’ hyperlinked documents is facially overbroad, unduly burdensome, and disproportionate.”  Id.  In that motion, 346 documents were at issue.  Defendant Shopify asserted that it “allowed Plaintiffs to identify hyperlinks that appear to be relevant by highlighting the policy documents, so that Shopify could produce those documents, or meet-and-confer in the event the parties disagreed as to relevance.”  The case appears to have been settled before a ruling was issued.[6]

SOME LESSONS LEARNED

Fed.R.Civ.P. 34 permits a party to specify the form or forms of production.  A requesting party should consider the hyperlink issue in drafting its requests.

Rule 34 should be used in combination with a Fed.R.Civ.P. 26(f) conference of the parties: “If there is a master blueprint to e-discovery under the Federal Rules, it lies in amended Rule 26(f)….”  Steven S. Gensler, “Some Thoughts on the Lawyer’s E-Volving Duties in Discovery,” 36 N. Ky. L. Rev. 521, 522 (2009).  That is particularly apropos in this context.See What is a Document? (Part III).

Parties should consider putting hyperlinks on their list of topics to be discussed in a Fed.R.Civ.P. 26(f) conference of the parties.  Before finalizing a Rule 26(f) discovery plan or ESI protocol, it may be prudent to ask an opposing litigant, for example, how the opponent’s email system handles attachments and links.  

Further, considering Bedford, Freeman & Worth Publishing Grp., and Shopify, it may make sense to discuss hyperlinks in other types of ESI.  For example, if this blog were to be produced in litigation, it seems unlikely that every hyperlinked document would be considered to be part of it.  However, one or more links could be important and that may also be an issue for a Rule 26(f) conference.

I have suggested that context is key to determining an appropriate scope of production and that would appear to apply with equal force to “modern attachments.”  See What is a “Document?”   The text message decision of  Sandoz v. Un. Therapeutics Corp., 2021 WL 2453142 (D. N.J. Jun. 16, 2021), may provide a conceptual benchmark for negotiations.  There, the issue was whether a party could produce only the text bubble that hit on a search term or whether preceding and following contextual bubbles must also be produced.  The order required production of “context-related text messages…, including the preceding text message or responding text message, if they exist. If they do not exist, [the producing party] must so state, and provide an explanation based on information available to it why they do not exist.”  Id. at *2.  In short, context was an important consideration in determining the scope of a “document.”

One option would be to agree that hyperlinked documents are defined to be part of the email transmitting them.  Depending on the facts, systems used, nature of the case, and proportionality, requesting “all” hyperlinked documents may or may not be a bridge too far.

However, phased[7] discovery may provide a more proportional approach, much like the Noom middle ground.  For example, in Phase I, the target file would not be produced as an email family.  However, in Phase II, a recipient would have the right to make a focused request for important familial emails (or other ESI) and the producing party could either comply or object, with all rights of each party reserved in Phase II, and none waived by Phase I.  

Understanding information governance issues will be important in any such discussion.  In The Annotated ESI Protocol ver. 20230107_Final.pdf (craigball.com), 11, Craig Ball’s exemplar requires production of the hyperlinked file and he added:  “The parties must appreciate what this obligation entails in the context of their messaging environment. Some Cloud systems (e.g., Microsoft 365) make it easy to collect documents transmitted as hyperlinked files versus embedded attachments, whereas others may demand manual collection with attendant uncertainty as to whether the item collected remains faithful to the item transmitted. As phrased, the operative distinction is whether the hyperlink in the transmittal points to a resource readily available to anyone with the link (that is, ‘documents merely referenced’) or whether the modern attachment item is unavailable to the requesting party if not produced with the transmittal.”

Porter and IQVIA point to another information governance issue and there may be unrecognized risks to the producing party if that party uses “modern attachments.”  If “modern attachments” are used, it may later become impossible for the custodian to determine what was sent with an email.  For example, the Porter court explained that “a responding party may not be able to collect the precise linked document referenced in a message if the document has been modified or deleted.”  This could create a problem for that custodial party:

[A]ssume that two years after an email with a pointer was sent, the email is used to question the recipient in a deposition.  The deponent may be unable to answer unless the “attachment” or “target” is provided.  However, because the target is in collaboration software, the “attachment” as sent may no longer exist.

See “Modern Attachments” or “Pointers”- What is a Document? (Part IV).  Not only does that present an issue to the discovering party, but it may also deprive the custodian of a factual defense.  Businesses using “modern attachments” may not be retaining prior drafts of important business records.

There may be additional evidentiary concerns for either the producing party or the recipient.  When only part of a “document” is produced under Noom, a foundational problem may arise.  Specifically, can part of a “document” be used to question a witness in depositions and at trial?  What is a “Document?”accord M. Berman, “Use of ESI in Depositions,” Chap. 23 in M. Berman, et al., eds., “Managing E-Discovery and ESI” (ABA 2011).  

Each of these issues, and likely others, may need to be discussed in a Fed.R.Civ.P. 26(f) conference of the parties.[8]


[1] See What Do Judge Peck’s 2009 Wm. Gross Opinion and “Zoomers” Have to Do With Each Other? and William A. Gross Const. Assocs., Inc. v. Am. Mfrs. Mut. Ins. Co., 256 F.R.D. 134 (S.D.N.Y. 2009)(discussing the need for custodian interviews)(Peck, J.).

[2] For a more specific example of the potential foundational problem in the context of ESI, see What is a Document? (Part II).

[3] MetaSpike’s webpage states: “Forensic Email Collector can acquire Google Drive attachments and their revisions during Gmail / Google Workspace acquisitions.”

[4] Interestingly, in a later decision in IQVIA, 2022 WL 111059 at *2 (D.N.J. Jan. 12, 2022), the court wrote that “after the Special Master granted IQVIA’s motion to compel in March 2018, it became clear that many documents ordered to be produced had been deleted.”

[5] After a subsequent summary judgment decision,  in which “the Court observes that the parties have made its task more difficult through their overly aggressive litigation,” an appeal was noted to the Ninth Circuit on April 14, 2023.

[6] Plaintiffs’ reply memorandum presents a different picture: “Shopify does its best to paint Plaintiffs’ request as overbroad. But it is not. Plaintiffs did not argue that all embedded material hyperlinked in every document Shopify produced must be turned over. Plaintiffs moved only on a critical group of documents: those that describe repeat infringement policies and provide guidance or instructions explaining how to implement those policies….  The materials referenced or embedded therein form a key part of those documents.”  2022 WL 19335696, at ¶I.

[7] Fed.R.Civ.P 26 (f)(3)((B) mandates that a discovery plan state the parties’ views and proposals on phased discovery.

[8] The prior blogs, hyperlinked above, cited to authors who addressed the technological issues presented.  See, also, e.g., The Modern Attachment: How to Handle Hyperlinked Documents in Emails | Kilpatrick Townsend & Stockton LLP – JDSupra.

Author

  • Michael D. Berman

    Mike is the owner of E-Discovery, LLC, and of counsel at Rifkin Weiner Livingston LLC, in Baltimore. He concentrates on commercial litigation and offers mediation services. He was the primary editor of Electronically Stored Information in Maryland Courts (Md. State Bar Ass’n. 2020), and he co-edited M. Berman, C. Barton, and P. Grimm, eds., Managing E-Discovery and ESI: From Pre-Litigation Through Trial (ABA 2011), and J. Baron, R. Losey, and M. Berman, eds., Perspectives on Predictive Coding (ABA 2016). Mike has litigated a number of cases in the trial and appellate courts in Maryland. He is an Adjunct Professor at the University of Baltimore School of Law where he co-teaches a three-credit discovery workshop that focuses on e-discovery. He has lectured at the Maryland Judicial College and he chaired the Bar committee that drafted the proposed ESI Principles for the District of Maryland. He is a past: co-chair of the Federal District Court Committee of the Maryland State and Federal Bar Associations; chair of the Litigation Section Council, Maryland State Bar Association; and, co-chair of the American Bar Association Litigation Section Book Publishing Board. He graduated from the University of Maryland School of Law and is also an Army veteran. He is admitted to the Maryland bar. The opinions expressed in this blog are not necessarily those of Rifkin Weiner Livingston LLC.