LawShelf courses have been evaluated and recommended for college credit by the National College Credit Recommendation Service (NCCRS), and may be transferred to over 1,500 colleges and universities.

We also have established a growing list of partner colleges that guarantee LawShelf credit transfers, including Excelsior College, Thomas Edison State University, University of Maryland Global Campus, Purdue University Global, and Touro University Worldwide.

For a limited time: Purchase a course multi-pack for yourself or a friend!

Proportionality and Admissibility - Module 4 of 5

Module 4: Proportionality and Admissibility

This module will cover some facets of the cost of discovery from the point of view of the court, and how to determine the admissibility of discovered ESI.


“Proportionality” is a cost-benefit analysis applied to the cost of electronic discovery to a party against other cost considerations like the potential value of the case and the probative value of the potentially discoverable evidence. The bigger the case, the more courts will allow discovery costs to escalate.

Proportionality has been a struggle for courts since it was first used as the basis for a series of decisions in Zubulake v. UBS Warburg early in the 21st Century. Zubulake consists of five pretrial decisions by Southern District of New York Judge Shira A. Scheindlin between 2002 and 2004. These are referred to as Zubulake 1 through 5.[1]  

The case involved the confluence of three facts that made it groundbreaking. First, the judge was one of the first jurists to take an interest in e-discovery and was on the ABA committee writing early e-discovery rules. Second, the plaintiff was very adept at technology and kept copious records. Third, the defendant hid and destroyed emails in obvious ways and lied about it.

During the pretrial process, the defendant claimed that it had 100 pages of emails related to the case, but the plaintiff had collected over 400 pages. Then the defendant claimed to have the missing emails on a backup tape, but that it would take approximately $300,000 to get them from the tape. At that point, the judge began one of the first proportionality analyses in e-discovery history in Zubulake III when she divided the cost of the discovery of those tapes between the parties — 75% to the defendant and 25% to the plaintiff. The court used the 5-category breakdown of where the data was located that we discussed in the previous module to conduct her analysis.

The defendant continued to stall and hide evidence.

By Zubulake V, the judge had lost patience with the defendants, and ruled that their stalling and obfuscating created an adverse inference - that the stalling could be taken to imply that the defendant had evidence that it was hiding. The jury eventually returned a verdict for the plaintiff of almost $30 million, including over $20 million in punitive damages. The parties settled, so there was no appeal.[2]

This and other adverse inference cases led to the creation in 2015 of new Federal Rule 37(e), which limits adverse inference findings to deliberate acts of data sabotage.[3]  

The 2015 rules added specific proportionality tests under FRCP 26(b)(1), which we will discuss below. The 2015 updates also included the rewritten “Cost Allocation Rule,” which imposed more severe limits on discovery costs by tightening up the proportionality test.[4]

Nevertheless, Zubulake remains a landmark case in e-discovery. In fact, the developers of the Electronic Discovery Reference Model point to this case as the one that propelled e-discovery into the forefront of legal thinking. 

The 2015 Amendments to the Federal Rules

Proportionality considerations are listed in Federal Rule 26(b). They are: (1) the importance of the issues at stake, (2) the amount in controversy, (3) the parties’ relative access to information, (4) the parties’ resources, (5) the importance of the discovery in resolving the issues, and (6) whether the burden or expense of discovery outweighs its likely benefits.[5]

Objecting to a discovery request based on these considerations requires a motion for a protective order under Rule 26(c). A protective order can be issued by a court on motion to stop a discovery request based on “annoyance, embarrassment, oppression, or undue burden or expense…” The rules then list several characteristics of protective orders, which can specify “terms, including time and place or the allocation of expenses, for the disclosure or discovery.”[6]

Decisions under the 2015 Rule Amendments

As we can see by the lengthy list of factors limiting discovery by cost, each court will decide whether to issue a protective order on a case-by-case basis. Here are some decisions based on the 2015 rule changes.

1.   The first federal court of appeals to cite proportionality under new Rule 26(b)(1) did so in the context of reversing the denial of a motion to amend a complaint.[7] In Niagara Risk Management v. Folino, the Eastern District of Pennsylvania court cited to proportionality as guidance for avoiding excessive expense and difficulty during potentially complex discovery.

2.    Not all courts have used the six-part analysis from the rule. In FTC v. DirectTV,[8] the Northern District of California federal court relied on the “spirit” of proportionality rather than the six-factor analysis to limit discovery against the Federal Trade Commission. It limited ESI production to a “sampling” rather than full-scale production of all requested documents. The court held that the parties and the court have a collective responsibility to consider the factors bearing on proportionality before propounding discovery requests, issuing responses and objections or raising discovery disputes before the courts.

3.    Some cases may be too complicated, or the data may be too spread out to even try to analyze the cost, especially when there already has been sufficient discovery. In In Re Baird,[9] an Arizona federal court held that wide-ranging discovery of communications with foreign regulators was too burdensome and therefore not proportional in light of extensive discovery already conducted involving communications with US regulators.

As of 2018, it is too early to assess the impact of the 2015 rules changes on state laws, but some states and the District of Columbia have adopted some of these changes. What is changing over time is the emphasis of e-discovery software vendors, who are developing “tiered” discovery platforms to limit costs as the discovery process unfolds. As in every facet of electronic discovery, the rest of the story is developing every day in the field and in the courts.

Admissibility of Electronically Stored Information at Trial

Once the data is secured by the parties, it still must be admitted at trial to be most useful. Evidence may be stipulated as admissible by the parties or there may be objections to admission at trial. If stipulated, which means that the parties agree on its admissibility, the evidence is simply admitted when proffered by any party at any time—often before the trial begins.

If one party objects to the admission of ESI, that party must tell the court the reason for the objection. After discussion, the court will rule one way or the other on the admissibility of the evidence.

There are five main objections to admitting most forms of ESI in civil litigation [10]:

--  Relevance

--  Unfair Prejudice vs. Probative Value

--  Best Evidence (Original Writing) Rule

--  Authentication/ Chain of Custody

--  Hearsay

Let’s look at each one, in turn. 


The relevance of ESI is covered under Federal Rules of Evidence, Rule 401. Neither the Federal Rules of Criminal Procedure nor the Federal Rules of Evidence have separate ESI rules like the Federal Rules of Civil Procedure do, so arguments around the relevancy of ESI are the same as those around any forms of evidence. There may be admissibility discussions at the pretrial conferences or objections during trial.[11]

Unfair Prejudice vs. Probative Value

To be relevant, and thus admissible, evidence must tend to prove something important to the resolution of the case. It must have “probative” weight or value. Additionally, it cannot unduly prejudice the jury or unduly distract the jury from the relevant questions. For example, in one wrongful death action, the judge may allow one photo of the victim from a distance, but not a close-up showing the decedent’s body. One photo would be enough to show where and how he died. Anything more would may be seen treated as inflammatory and prejudicial. Other forms of prejudicial evidence include anything distracting, causing undue delay or confusion.

This type of objection[12] is not specifically tied to ESI. However, the form of the evidence may itself create undue prejudice. People write before they think in emails, texts, social media, etc. far more than in a thoughtful or business-oriented letter. Courts may be reticent to admit statements that contain the kind of highly emotional, offensive or derogatory language that people post habitually on social media because of their prejudicial nature (unless the language itself is at issue).

Courts have also ruled that computer animation may be inadmissible because of confusion—the jury might think that it is real.[13]

Summaries of ESI

Courts can allow brief summaries of the ESI collected[14]—after all, no jury is going to want to wade through 10 million pages of documents. But these summaries will be subject to a prejudice vs. probative value analysis by the court.

Authentication and the Original Documents Rule

The Best Evidence Rule, sometimes called the Original Documents Rule, is a very old rule in the law that states that an original document is preferable to a copy.[15] Copies are authenticated and admissible only under much more stringent rules than original documents. The problem with bringing that rule into modern times, of course, is that most original documents are not originally on paper, aren’t usually printed, are changed continually and sit not on a desk or in a file cabinet but on a hard drive on some device.


Does the party proffering that evidence have to provide the device itself? No! Under the rules, a printout or digital copy is generally sufficient to satisfy this rule.[16]


There are also many circumstances under which an original electronic document may not be able to be produced. Cell phones get destroyed, computers get their memories wiped or sent to the trash, emails get deleted, backup tapes are erased, data is unavailable by ordinary means or is so collateral to the case that it may not be discoverable by ordinary means, and so on. In those cases, “secondary evidence” may be used to authenticate a copy of the data.[17]

At least one court, however, has ruled that “cut-and-pasted” materials from a document was NOT admissible as the original document under these rules.[18]


The issue of authentication is simple on its surface: is the evidence what it purports to be? Federal Rules 901 and 902 cover authentication.[19] Those rules cover evidence that is authenticated by certain kinds of testimony, including direct witness testimony, comparison testimony of an expert witness, distinctive characteristics, and self-authentication (like the date on a newspaper article). Both direct and circumstantial evidence are subject to authentication.

ESI presents some very specific authentication challenges.

The first of these is authenticating data received through computer forensics. Third-party computer forensics companies have standard authentication and chain of custody metrics, which will be proffered to the court as authentication of the data to be presented as evidence.[20] Presenting this evidence often involves testimony from someone from the forensics company, third party or another expert and will involve the steps of the discovery and preservation of the data. 

Next is metadata, which shows the author(s) of a document and the document’s changes through time. It is often used as both authentication and evidence itself. Data can also contain various hashmarks, which are unique identifiers assigned to specific documents (not to be confused with hashtags). In addition, encryption and passwording can provide evidence of the trustworthiness of data.

Admitting Specific Types of Data


First, questions will arise about whether a printout, screenshot or however a website is presented is what the website looked like at the time. Sometimes the webmaster or, more often, an administrator can testify to that.[21] Next, authenticating a website requires proof of ownership of the site or that a specific person made the web posting.[22]


Emails are not only the most prevalent form of ESI but have the lowest authentication standard of any ESI. The general standard for admission of email data:

E-mails satisfy authentication requirements when they bear “distinctive characteristics” including the actual e-mail addresses containing the “@” symbol, email addresses containing name of the person connected to the address, signatures within e-mails linking e-mails to alleged sender, and discussions in the e-mails of personal and professional matters known to be associated with senders. [23]


The email itself must be submitted—not just information about the email.[24] If the email was generated during the regular course of a business, it may be authenticated as a business record.[25]


Text Messages


Text messages are authenticated similarly to emails. Some courts require evidence that the alleged text sender not only owned the texting source device but actually sent the text.[26] While the Best Evidence Rule may require the physical presence of the texting device, transcripts of the texts may be allowed if the source of the texts has been destroyed.[27] Texts may also be authenticated by expert testimony.

Online Chats, Social Media Chats and Instant Messages

Online chats, social media chats and instant messages have similar methods of authentication, which include:[28]

--  Evidence that individual used the screen name in question when participating in chat room conversations;


--  Evidence that when a meeting with the person using the screen name was arranged, the person showed up;


--  Evidence that the person using the screen name identified himself as the person in the chat conversation;


--  Evidence that the person had in his possession information given to the person using the screen name;


--  Evidence from the hard drive of the individual's computer.


Also, authentication of online chats can be accomplished through testimony of a person in the chat that could testify that the chats were fairly reproduced through direct personal knowledge.[29] A printout of the chat/ IM can satisfy the Best Evidence Rule.[30]


There are further authentication rules for social media chats including knowledge of the person’s user name and several other factors, which vary from case to case.




Authentication of voicemail consists of proof that the person’s recorded voice is authentic and unadulterated. It parallels the admissibility of any other voice recording. Authentication may need a contemporary witness to the recording, although high tech analysis including voice printing may be used. One particularly difficult piece of tech involves authenticating a voice message sent as an attachment to an email. This data could be subject to an expensive forensic study to determine authenticity.


Cloud Storage Data


There are no recognized legal or technical standards for the storage of data in the cloud.  Each piece of proffered evidence that comes from a third-party cloud storage service must be authenticated within the framework of its own chain of custody and its own technology, including metadata and the methods used to store and retrieve the data. One thing is clear, though: courts have found that the party that produced and stored specific cloud data is the owner of that data for authentication purposes.[31]




Hearsay objections under Article 8 of the Federal Rules of Evidence may be raised in circumstances such as overheard cell phone conversations, conversations about text messages or Facebook posts that a witness never read, and so on. Those are dealt with as in the case of any other case of hearsay.


In our final module, we will look at unique problems that arise in the context of e-discovery and potential solutions to these cutting edge and high tech issues.

[1] Zubulake v. UBS Warburg, 217 F.R.D. 309 (S.D.N.Y. 2003) (Zubulake I); Zubulake v. UBS Warburg, 230 F.R.D. 290 (S.D.N.Y. 2003) (Zubulake II); Zubulake v. UBS Warburg, 216 F.R.D. 280 (S.D.N.Y. 2003) (Zubulake III); Zubulake v. UBS Warburg, 220 F.R.D. 212 (S.D.N.Y. 2003) (Zubulake IV); Zubulake v. UBS Warburg, 229 F.R.D. 422 (S.D.N.Y.) (Zubulake V). Note Zubulake II does not relate to electronic discovery.

[4] The entire 2015 e-discovery update to the FRCP included:

(1) Cooperation (Rule 1)

(2) Case Management (Rules 4, 16, 26, 34)

(3) Scope of Discovery/Proportionality (Rule 26(b))

(4) Presumptive Limits (Rules 30, 31, 33)

(5) Cost Allocation (Rule 26(c))

(6) Production Requests/Objections (Rule 34, 37)

(7) Failure to Preserve/Spoliation (Rule 37(e))

[5] Niagara Risk Management v. Folino, 2016 U.S. Dist. LEXIS 106094, at *11 (E.D. Pa. Aug. 11, 2016).

[8] FTC v. DIRECTV, Inc., 2016 U.S. Dist. LEXIS 75480 *3-5 (N.D. Cal. June 9, 2016).

[9] In Re Baird, 317 F.R.D.562, 566 (Dist. Ariz. 2016).

[10] See generally Lorraine v. Markel Am. Ins. Co., 241 F.R.D. 534 (D. Md. 2007).

[13] Friend v. Time Mfg. Co., 2006 WL 2135807 at *7 (D. Ariz. July 28, 2006) (computer simulation of a traffic accident).

[18] United States v. Jackson, 488 F. Supp. 2d 866, 870-71 (D. Neb. 2007).

[20] United States v. Meienberg, 263 F.3d 1177, 1180 (10th Cir. 2001); see also Peter Coons, How to Document Your Chain of Custody and Why It’s Important, D4 Discovery, (July 1, 2015),http://www.d4discovery.com/discover-more/how-to-document-your-chain-of-custody-and-why-its-important#sthash.GtsAa9CX.dpbs.

[23] United States v. Safavian, 435 F. Supp. 2d 36, 40 (Dist. D.C. 2006).

[24] Jimena v. UBS AG Bank, Inc., 2011 WL 2551413 at *11 (E.D. Cal. June 27, 2011) (same defendant as Zubulake).

[26] Pennsylvania v. Koch, 39 A.3d 996, 1005 (Pa. 2011).

[27] State v. Espiritu, 176 P.3d 885, 892-93 (Haw. 2008).

[28] See generally United States v. Barlow, 568 F.3d 215, 220 (5th Cir. 2009); see also Hon. Paul W. Grimm & Gregory P. Joseph, Best Practices for Authenticating Digital Evidence 13-14 (West 2016).

[30] United States v. Nobrega, 2011 WL 2116991 at * 8 (D. Me. 2011).

[31] See United States v. Ganias, 824 F.3d 199 (2d Cir. 2016).