The Basics of Electronic Discovery: Part 2 of 2

In Part One of this article, we took a broad look at what e-discovery really is, defined the key terminology, and examined some of the technology impacting this field. Now, in this final Part Two, we will explore the technology in more detail, address the defensibility of the technology, and conclude with how all of this affects you, as a practitioner.


The technology used for e-discovery is constantly changing and improving. Well, if it is evolving at such a rate, then why do you need to obtain some technology to assist you now? Because, like your phone and computer, it has become an unavoidable component of the practice of law. New phones come out all the time, but you don’t likely change yours every three months. Similarly, new advances improve the scope and speed of e-discovery tools all the time, however, you will not need to change your technology regularly, but merely to have some reliable and proven technology upon which to rely to assist you in this regard.

The technology or software to assist you in your e-discovery practice can be hosted on your computer or reside elsewhere, such as on the server of a vendor or in the cloud. Each of these options has their own security concerns for protecting client data. Naturally, the decision of which method to utilize is one that only you can make after exploring the options with vendors. Many vendors offer multiple versions of the same technology for those who wish to self-host or those who wish to have remote access to the data.

While technology exists to aid you in sorting data collected or provided from your client and third-parties, you must also be aware that technology exists to assist your clients in preserving, maintaining, and producing ESI. To complement the software, it is not too early to begin discussing a litigation hold policy with your clients and even developing internal training strategies for their employees and contractors to avoid the inadvertent loss of important ESI.

Whatever route you take and whatever technology you ultimately choose to assist you in handling and producing ESI, you will be faced with a decision of how much control to release to the computer/software “brain”. At present, this is a hot button topic in the federal courts and has been addressed in several recent federal opinions. Software that assists us in searching voluminous data that allows the software to make decisions as to relevance has been referred to as “computer assisted research”, “computer assisted coding”, and “predictive coding”. This is in contrast to technology that utilizes a different method of searching, such as keywords, whereby the software merely identities those pre-defined keywords in documents and leaves the ultimate decision on responsiveness to the practitioner. Obviously, the first method is arguably more sophisticated programming, but its process must be reliable in order to be defended to a Court later if the issue arises, which leads us to the issue of defensibility.


Although, at present, our Florida State Courts may be in somewhat uncharted waters, it is foreseeable from parallel developments in federal cases that the method by which ESI was processed may be called into question in a particular case. While one argument may be that the culling method is protected work product, it may also be appropriate for a Court to inquire into the method used in a particular case to determine, for example, whether the imposition of sanctions is appropriate or in balancing which party should bear the costs of production from the party or a third-party.

In selecting software to assist you in e-discovery or a third-party vendor to perform this task, you will want to inquire as to the methodology employed in the software and its track record to avoid a potentially costly situation later where you are called upon to defend the methodology used in production. That is not to imply that computer assisted review or predictive coding is somehow inappropriate or untested technology, but simply that you must be aware of the technology you are using no different than understanding the methodology of an expert you would employ.


This brings us to the all-important question of why should you care about any of this. E-discovery will become an integral component to civil discovery and with it the real potential for a wide range of sanctions.

In recent months there have been several highly visible cases in which Courts imposed sanctions based on e-discovery violations. Sanctions can range from the exclusion of evidence and adverse jury instructions to post-trial findings that impact a party’s appeal.

Sanctions can also be monetary and can run against a party or even, in some cases, directly against the attorney that aids in secreting a client’s electronic data. Similarly, the improper production of privileged materials in e-discovery may be considered legal malpractice. Thus, the lesson to be taken is that all attorneys now have a new obligation to counsel their clients on ESI and data management, therefore, the attorneys must understand this area of the law, as well.

This “brave new world” of e-discovery is here to stay and must be embraced and integrated into the practices of all Florida civil litigation attorneys. It is conceptually not that different from traditional discovery. Disregarding it, however, can be perilous and costly, while mastering its nuances can make you more efficient, cost effective, and provide a better service to your clients.

The Evolution of E-Discovery and Computer Forensics, Part 2: Zubulake V. UBS Warburg

The field of electronic discovery and digital forensics is rapidly evolving. In the early years of this millennium, discovery rules dealt primarily with paper, but with the advent of the computer age, documents are drafted electronically and important rules regarding Electronically Stored Information still needed to be invented. This series looks at a few of the major cases, opinions and outcomes that have informed this evolution.

Judge Shira Scheindlin issued precedent-setting (and often-cited) opinions in the important case of Zubulake v. UBS Warburg – 2003-2005.

The Back Story:

Laura Zubulake worked in New York and Connecticut for the Swiss-based firm, UBS Warburg, Europe’s largest bank at the time. She was an extremely successful equities manager, earning more than $650,000 a year selling Asian equities to institutional investors for a decade. In 2000, she was passed over for a promotion that she had been promised, to take over as senior sales manager of the Asian desk in the U.S. when her superior left that position.

Instead, a Matthew Chapin was given the position, whereupon, according to the plaintiff, he “belittled her in front of co-workers and denied her important accounts… During the trial, a former UBS sales assistant testified that she overheard Chapin call Zubulake “old” and “ugly.” (“UBS Must Pay Ex-Saleswoman $29.3 Mln in Sex Bias Case (Update5)” – Bloomberg, April 6, 2005).

In August 2001, Zubulake filed a complaint with the employment commission, and in October, Chapin fired her, in the process (as the court found) falsifying emails, records and complaints. Zubulake sued UBS for gender discrimination, failure to promote, and retaliation under federal, state, and city law.

UBS argued that Chapin wasn’t abrasive because of sexual discrimination, but rather that he was abrasive to everyone, including male employees. A remarkable argument! The bank maintained that she was fired for insubordination.

When the time came for producing documents in discovery, UBS produced just 100 total emails, yet Zubulake herself was able to produce 450 relevant emails of communications between company staffers. UBS was under litigation hold obligations but still had apparently made hundreds of emails disappear in spite of their obligation to retain them. Furthermore, UBS produced additional emails that appeared to be falsely generated.

When discovery was requested for archival data and backup tapes, UBS asserted that to go after such data would be an undue expense and a burden on UBS. It cited the case of Rowe v. William Morris and asked the court to shift the expense of production to Zubulake based on the “Rowe test,” a set of weighting factors used to determine cost-shifting that derived from the Rowe case.

Judge Shira Scheindlin of the New York Southern District produced five evolving opinions regarding who should pay for production/discovery, to what extent discovery and production of ESI is allowable, and how to determine a party’s duty to preserve evidence. She found that just because data is electronically stored (ESI) doesn’t necessarily make its production an undue expense. In fact, due to the ability to perform machine/computer searches, costs can actually be less than equivalent human searches of paper documents.

The burden of cost is increased with decreasing accessibility, as determined by the type of media on which the ESI is stored. There were determined to be five categories of electronic repositories: online data (such as hard disks), near-line data (such as CDs and other optical disks), offline storage (such as magnetic tapes), backup tapes, and fragmented, erased and damaged data. Backup tapes and fragmented/damaged data were considered to be most inaccessible and therefore most subject to cost-shifting.

The court ordered sampling of the data by having 5 backup tapes restored to determine whether there was a likelihood that the remaining 70+ tapes would produce relevant data. They did produce 600 responsive messages. Judge Schendlin designed a new seven-factor test to determine whether cost-shifting was in order.

The first two factors are considered to be of the most importance.

1: Is the request tailored to discover relevant info? (Fishing expeditions frowned upon).

2: Is information available from other sources? (The parties should get the information from the most readily available sources, such as company reports or public information instead of having to dig though old backup tapes, for instance).

The next three factors are considered to be of secondary importance.

3: Total cost of production v. the amount in controversy (the cost of discovery should be considerably less than the potential winnings in the case).

4: Total cost v. resources available to each party (it shouldn’t bankrupt anyone).

5: Relative ability, incentive to control costs (clearly the party paying for production has a strong incentive to control costs).

The final two factors are considered to be of lesser importance than the first five.

6: Importance of issues at stake in litigation (Will the case have an important impact on society? The Zubulake case had to do with gender discrimination, but was not a groundbreaking case in that area).

7: What are the relative benefits to parties of obtaining the requested information?

(It is generally assumed that the plaintiff aims to benefit and so this test is then rarely considered to be of great importance.)

Eventually, the court found that UBS had lost evidence (some monthly backup tapes were missing), carelessly destroyed evidence (some weekly tapes backfilled the monthly tapes), willfully withheld additional evidence, and even faked evidence. As a result, Judge Scheindlin issued an adverse inference instruction to the jury, “Because UBS’s spoliation was willful, the lost information is presumed to be relevant.” In other words, if data was missing, the jury could assume that UBS destroyed it on purpose because it might have hurt the bank’s case. A disaster for UBS.

Zubulake won more than she’d asked for: $29 million, including $9 million in compensatory damages and $20 million punitive damages. UBS had to pay for depositions and repeat depositions, the costs of the motion, and nearly all the cost of production.

The Zubulake case produced several milestones in the evolution of law around Electronic Discovery, and led to many of the 2006 Amendments to the Federal Rules of Civil Procedure (FRCP). The milestones include:

The parties have a duty to preserve ESI during litigation. – not only once there’s a litigation hold, but also if litigation is anticipated.

Lawyers have a duty to monitor their clients’ ESI compliance. This includes outside counsel! Sanctions don’t only affect the party and inside counsel.

Data sampling is allowed and encouraged. In the discovery process, grab data from a few tapes and hard drives first, for instance – to see if there’s likely to be anything on the rest – or even if it’s all available from a few (and possibly duplicated on the rest).

The disclosing party can shift the costs for less accessible data. If the requestor is looking for information that’s hard to dig up or produce, the cost of producing that may have to be shifted to the requestor.

There may be sanctions imposed for the spoliation of ESI.

The Zubulake case set out rules and tests that have informed later courts’ decisions as well as the 2006 amendments to the Federal Rules of Civil Procedure, and the 2009 California rules. They continue to help shape & inform law with regard to electronically stored information. As a result, the case also continues to change the shape of the computer forensic and electronic discovery industries.

Next in this series: the 2006 ESI Amendments to the Federal Rules of Civil Procedure.