e-discovery 2.0

The world of eDiscovery appears to be revolving around a trifecta of issues that are important to both clients and counsel. A discovery-focused conversation with litigants and lawyers in 2013 will almost invariably turn to some combination of this eDiscovery trinity: Spoliation sanctions, keyword searches and predictive coding. This should not come as a surprise since all three of these issues can have a strong impact on the cost, duration and disposition of a lawsuit. Indeed, the near universal desire among parties to minimize discovery costs and thereby further the resolution of cases on the merits has driven the Civil Rules Advisory Committee to explore ways to address the eDiscovery trinity in draft amendments to the Federal Rules.

While the proposed amendments may or may not succeed in reducing discovery expenses, the examples of how the eDiscovery trinity is playing out in litigation are instructive. These cases – bereft of the additional guidance being developed by the Advisory Committee – provide valuable insight on how courts, counsel and clients are handling the convergence of these issues. One such example is a recent decision from the DuPont v. Kolon Industries case.

Spoliation, Keywords and a $4.5 Million Sanction

In DuPont, the court awarded the plaintiff manufacturer $4.5 million in fees and costs that it incurred as part of its effort to address Kolon’s spoliation of ESI. In an attempt to stave off the award, Kolon argued that DuPont’s fees were not justified due to “inefficiencies” associated with DuPont’s review of Kolon’s document productions. In particular, Kolon complained about the extensive list of search terms that DuPont developed to comb through the ESI Kolon produced. According to Kolon, DuPont’s search methodology was “recklessly inefficient”:

Kolon then asserted that the “reckless inefficiency” of the search methodology was “fairly attributable to the fact that DuPont ran insipid keywords like ‘other,’ ‘news,’ and ‘mail.’” Had DuPont been more precise with its keywords searches, argued Kolon, it “would have saved vast amounts of time and money.”

Before addressing the merits of Kolon’s arguments, the court observed how important search terms had become in discovery:

After doing so, the court rejected Kolon’s arguments, finding instead that DuPont’s search methodology was reasonable under the circumstances. The court based its decision on the source of those search terms (derived from Kolon documents suggesting that ESI had been deleted), the “considerable volume” of Kolon’s productions and the nature of DuPont’s search (an investigation for deleted evidence).

The Impact of Predictive Coding on DuPont’s Search Efficiency

While DuPont considered the issues of spoliation and keywords in connection with the imposition of attorney fees and costs, it was silent on the impact that predictive coding might have had on the fee award. Indeed, neither the court’s order, nor the parties’ briefing considered whether the proper application of machine learning technology could have raised the success rate of DuPont’s searches for documents relevant to Kolon’s spoliation above the ten percent (10%) figure cited by Kolon.

On the one hand, many eDiscovery cognoscenti would likely assert that a properly applied predictive coding solution could have produced the same corpus of relevant documents at a fraction of the cost and effort. Others, however, might argue that predictive coding perhaps would not yield the results that DuPont obtained through keyword searches given that DuPont was looking for evidence of deleted ESI. Still others would contend that the issue is moot since DuPont was fully within its right to determine how it should conduct the search of Kolon’s document productions.

Whether predictive coding could have made a difference in DuPont is entirely speculative. Regardless, the debate over keyword searches versus machine learning technology will likely continue unabated. As it stands, the DuPont case, together with the recent decision from Apple v. Samsung, confirm that keywords may be an acceptable method for conducting searches for relevant ESI. The issue, as the DuPont court observed, turns on “the selection of the search terms.”

Nevertheless, the promise of predictive coding cannot be ignored, particularly if the technology that is used could ultimately reduce the costs and duration of discovery. Given that this debate is far from settled, these issues, along with spoliation sanctions, will likely continue to dominate the eDiscovery airwaves for the foreseeable future.

Apple obtained a narrow discovery victory yesterday in its long running legal battle against fellow technology titan Samsung. In Apple Inc. v. Samsung Electronics Co. Ltd, the court ordered non-party Google to turn over the search terms and custodians that it used to produce documents in response to an Apple subpoena.

According to the court’s order, Apple argued for the production of Google’s search terms and custodians in order “to know how Google created the universe from which it produced documents.” The court noted that Apple sought such information “to evaluate the adequacy of Google’s search, and if it finds that search wanting, it then will pursue other courses of action to obtain responsive discovery.”

Google countered that argument by defending the extent of its production and the burdens that Apple’s request would place on Google as a non-party to Apple’s dispute with Samsung. Google complained that Apple’s demands were essentially a gateway to additional discovery from Google, which would arguably be excessive given Google’s non-party status.

Sensitive to the concerns of both parties, the court struck a middle ground in its order. On the one hand, the court ordered Google to produce the search terms and custodians since that “will aid in uncovering the sufficiency of Google’s production and serves greater purposes of transparency in discovery.” But on the other hand, the court preserved Google’s right to object to any further discovery efforts by Apple: “The court notes that its order does not speak to the sufficiency of Google’s production nor to any arguments Google may make regarding undue burden in producing any further discovery.”

This latest opinion from the Apple v. Samsung series of lawsuits is noteworthy for two reasons. First, the decision is instructive regarding the eDiscovery burdens that non-parties must shoulder in litigation. While the disclosure of a non-party’s underlying search methodology (in this instance, search terms and custodians) may not be unduly burdensome, further efforts to obtain non-party documents could exceed the boundaries of reasonableness that courts have designed to protect non-parties from the vicissitudes of discovery. For as the court in this case observed, a non-party “should not be required to ‘subsidize’ litigation to which it is not a party.”

Second, the decision illustrates that the use of search terms remains a viable method for searching and producing responsive ESI. Despite the increasing popularity of predictive coding technology, it is noteworthy that neither the court nor Apple took issue with Google’s use of search terms in connection with its production process. Indeed, the intelligent use of keyword searches is still an acceptable eDiscovery approach for most courts, particularly where the parties agree on the terms. That other forms of technology assisted review, such as predictive coding, could arguably be more efficient and cost effective in identifying responsive documents does not impugn the use of keyword searches in eDiscovery. Only time will tell whether the use of keyword searches as the primary means for responding to document requests will give way to more flexible approaches that include the use of multiple technology tools.

Much of the writing in the eDiscovery community focuses on the consequences of a party failing to adequately accomplish one of the nine boxes of the Electronic Discovery Reference Model. Breaking news posts frequently report on how spoliation and sanctions are typically issued for failure to suspend auto-deletion or to properly circulate a written litigation hold notices. This begs the question, aside from becoming perfectly adept in all nine boxes of the EDRM, how else can an organization protect themselves from discovery wars and sanctions?

One way is explore the possibilities Alternative Dispute Resolution (ADR) has to offer. While there is no substitute for the proper implementation of information governance processes, technology, and the people who manage them; there are alternative and creative ways to minimize exposure. This is not to say that ESI is less discoverable in ADR, but it is to say with the proper agreements in place, the way ESI is handled in the event of a dispute can be addressed proactively.  That is because although parties are free to use the Federal Rules of Civil Procedure in ADR proceedings, they are not constricted by them. In other words, ADR proceedings can provide parties with the flexibility to negotiate and tailor their own discovery rules to address the specific matter and issues at hand.

Arbitration is a practical and preferred way to resolve disputes because it is quick, relatively inexpensive and commonly binding. With enough foresight, parties can preemptively limit the scope of discovery in their agreements to ensure the just and speedy resolution of a matter. Practitioners who are well versed in electronic discovery will be the best positioned to counsel clients in the formation of their agreements upfront, obviating protracted discovery. While a similar type of agreement can be reached and protection can be achieved with the Meet and Confer Conference in civil litigation, ADR offers a more private forum giving the parties more contractual power and less unwanted surprises.

For example, JAMS includes this phrase in one of their model recommendations:

JAMS recognizes that there is significant potential for dealing with time and other limitations on discovery in the arbitration clauses of commercial contracts. An advantage of such drafting is that it is much easier for parties to agree on such limitations before a dispute has arisen. A drawback, however, is the difficulty of rationally providing for how best to arbitrate a dispute that has not yet surfaced. Thus, the use of such clauses may be most productive in circumstances in which parties have a good idea from the outset as to the nature and scope of disputes that might thereafter arise.

Thus, arbitration is an attractive option for symmetrical litigation where the merits of the case are high stakes and neither party wants to delve into a discovery war. A fair amount of early case assessment would be necessary as well, so parties have a full appreciation about what they are agreeing to include or not include in the way of ESI.  Absent a provision to use specific rules (American Arbitration Association or Federal Arbitration Act), the agreement between parties is the determining factor as to how extensive the scope of discovery will be.

In Mitsubishi Motors v. Soler Chrysler-Plymouth, Inc., 473 U.S. 614, 625 (1985), the U.S. Supreme Court has explained that the “liberal federal policy favoring arbitration agreements’…is at bottom a policy guaranteeing the enforcement of private contractual agreements. As such, assuming an equal bargaining position or, at least an informed judgment, courts will enforce stipulations regarding discovery, given the policy of enforcing arbitration agreements by their terms.” Please also see an excellent explanation of Discovery in Arbitration by Joseph L. Forstadt for more information.

Cooperation amongst litigants in discovery has long been a principle of the revered Sedona Conference. ADR practitioners facing complex discovery questions are looking to Sedona’s Cooperation Proclamation for guidance with an eye toward negotiation by educating themselves on ways to further minimize distractions and costs in discovery.  An example of one such event is at The Center for Negotiation and Dispute Resolution at UC Hastings, where they are conducting a mock Meet and Confer on May 16, 2013. The event highlights the need for all practitioners, whether it be the 26 (f) conference for litigation or the preliminary hearing in the case of arbitration, to assess electronic discovery issues with the same weight they do claims and damages early on in the dispute.

It is also very important that arbitrators, especially given the power they have over a matter, to understand the consequences of their rulings. Discovery is typically under the sole control of the arbitrator in a dispute, and only in very select circumstances can relief be granted by the court. An arbitrator that knows nothing about eDiscovery could miss something material and affect the entire outcome adversely. For parties that have identified and addressed these issues proactively, there is more protection and certainty in arbitration. Typically, the primary focus of an arbitrator is enforcing the contract between parties, not to be an eDiscovery expert.

It is also important to caution against revoking rights to discovery by entering into mutual agreements to unreasonably limit discovery.  This approach is somewhat reminiscent of the days when lawyers would agree not to conduct discovery, because neither knew how. Now, while efficiency and cost savings are a priority, we must guard against a potential similar paradigm emerging as we may know too much about how to shield relevant ESI.

As we look to the future, especially for serial litigants, one can imagine a perfect world in arbitration for predictive coding. In the Federal courts, we have seen over the past two years or so an emergence of the use of predictive coding technologies. However, even when the parties agree, which they don’t always, they still struggle with achieving a meeting of the minds on the protocol. These disputes have at times overshadowed the advantage of using predictive coding because discovery disputes and attorney’s fees have overtaken any savings. In ADR there is a real opportunity for similarly situated parties to agree via contract, upfront on tools, methodologies and scope. Once these contracts are in place, both parties are bound to the same rules and a just and speedy resolution of a matter can take place.

References to the “Sedona Bubble” are overheard more and more commonly at conferences dealing with cutting edge topics like the use of predictive coding technology in eDiscovery. The “Sedona Bubble” refers to a small number of lawyers and judges (most of whom are members of The Sedona Conference) that are fully engaged in discussions about issues that influence the evolution of modern discovery practice. Let’s face it. The fact that only a small percentage of judges and lawyers drive important eDiscovery policy decisions is more than just a belief, it is reality.

This reality stems largely from the fact that litigators are a busy lot. So busy in fact, that they are often forced to operate reactively instead of proactively because putting out unexpected fires comes with the territory in litigation practice. As a result, the Sedona Bubble has a tremendous impact on cutting edge eDiscovery issues that include topics spanning everything from cross-border litigation and cloud computing, to social media and bring your own device to work (BYOD) issues. Recognizing the heavy time demands facing most litigators is what compelled me to provide more insight into the Sedona Bubble. That is why I am sharing my top three observations about the current state of predictive coding – one of the hottest eDiscovery topics on the planet.

#1 – Plenty of confusion about TAR still exists

Technology-assisted review (TAR) is a term that often means different things to different people. Adding further confusion to the discussion is the fact that the acronym TAR is commonly used interchangeably with other terms like computer-assisted review (CAR) and predictive coding. Many believe confusion about TAR is largely the result of misinformation spread by eDiscovery providers eager to capitalize on current marketplace momentum. Regardless of the reason, many in the industry remain confused about the key differences between predictive coding and other kinds of TAR tools.

What is important to remember is that most people are referring to predictive coding technology when they use any of the aforementioned terms. Predictive coding is a type of supervised machine learning technology that relies on human input to “train” a computer to classify documents. That does not mean attorneys are abdicating their responsibility to review and classify documents during discovery. t means that attorneys can review a fraction of the documents at a fraction of the cost by training the computer system.

In recent months, more litigators and judges are beginning to understand that there are many kinds of TAR tools to choose from in the litigator’s toolbelt.™ Predictive coding is one of the tools that falls underneath the broader TAR umbrella and is arguably the most important tool in the toolbelt™ if used properly. All the tools can be helpful, however, TAR tools such as keyword searching, concept searching, clustering, email threading, and de-duplication are not supervised machine learning tools and therefore are not predictive coding tools. The rule of thumb for those being courted by predictive coding is caveat emptor.  Make sure the providers clarify what they mean when they use terms like TAR, CAR, or predictive coding.

#2 – Momentum is building

In 2013, more and more attorneys and judges are dipping their toes into predictive coding waters – waters that were largely perceived as too frigid to enter only last year. One explanation for the increased usage of predictive coding technologies is the corresponding increase in judicial guidance. In the beginning of 2012, there were no known cases addressing the use of predictive coding technology. Since then, at least six different judges have addressed the use of predictive coding technology. (Moore v. Publicis Group; Kleen Products v. Packaging Corporation of America; Global Aerospace v. Landow Aviation; In re: Actos Product Liability Litigation; EOHRB v. HOA Holdings; Gabriel Technologies v. Qualcomm). Taken as whole, the court decisions are either supportive of the technology or remain neutral on the issue. In fact, an order in a new case named In Re Biomet was reported only a few days ago and continues the general trend toward judicial awareness and support of the technology.

In addition to the growing number of judicial opinions, conference attendees are sharing experiences related to the use of these technologies far more than was the case at conferences in 2012. This data point suggests that usage far exceeds the number of reported predictive coding cases. Further evidence of this momentum, and possibly even greater momentum to come, are discussions about adding comments to the proposed FRCP amendments that would encourage the use of predictive coding technology. Newly proposed amendments to the Federal Rules of Civil Procedure are expected to be published for comment in August, and predictive coding will almost certainly be part of the discussion.

#3 – Skeptics remain

Despite a significant uptick in predictive coding usage since early 2012, the technology is not without skeptics. Those less bullish cite concerns about the multitude of new predictive coding offerings that have recently come onto the market. Most realize that all predictive coding technologies are not created equally and the vast majority of tools on the market lack transparency. A key concern on this front is the lack of visibility into the underlying statistical methodology that many tools and their providers apply. Since statistics are the backbone of a viable predictive coding process, a lack of transparency into statistical methodologies by most providers has left some to perceive all predictive coding tools as “black boxes.” In reality, different tools provide different levels of transparency, but a general lack of transparency in the industry has perpetuated a “throw the baby out with the bathwater” mentality in some circles. Rumblings about the applicability of Daubert and/or Rule 702 in vetting these tools and the methodologies they rely upon are likely to gain steam.

The issue of transparency is also a common area of debate in the context of an issue known as the “discard pile.” The discard pile generally refers to documents classified as non-responsive that are used to train the predictive coding solution. The protocol established in Da Silva Moore and other cases requires the producing party to reveal the discard pile to the propounding party as part of the predictive coding training process. Proponents argue that this additional level of cooperation invites scrutiny by both parties that will help insure that training documents are properly classified. The rationale in support of this approach is that predictive coding tools are garbage-in garbage-out devices so improperly classifying training documents will lead to erroneous downstream results. The pushback by producing parties varies, but one common theme is predominant and can be summarized as follows: “I will share my non-responsive documents to the other side when they are pried from my cold, dead fingers.”

Conclusion

Although some barriers to widespread predictive coding adoption remain, it is clear that the future of predictive coding is now. Eventually best practices for using these technologies will rise to the surface and the tools themselves will improve. For example, most tools today require complex statistical calculations to be made manually. That means hiring consultants and/or statisticians to crunch the numbers in order to ensure a defensible process which increases costs. The tools themselves can also be costly because most providers charge a premium to use predictive coding solutions. However, price pressure is already afoot and some providers offer their predictive coding technology at no additional cost. In short, despite some early challenges, most of those within the Sedona Bubble believe predictive coding is here to stay.   

In Part One of “How Good is Your Predictive Coding Poker Face?” we shared video footage of Maura R. Grossman, Craig Ball, Ralph C. Losey and myself (Matthew Nelson) discussing similarities between predictive coding technology and the popular poker game Texas Hold ‘em during a panel discussion at Legal Tech New York in January. In particular we discussed how to “read your opponent” when devising predictive coding protocols and we tackled sensitive issues like whether or not parties should be required to show their “discard pile” (aka non-responsive files used to train the predictive coding system) to the other side.

In Part Two of our two part video series, the panel digs deeper into the parallels between poker and predictive coding by discussing a “middle ground” approach to predictive coding referred to as “splitting the pot.” The panel also explores interesting issues like the dwindling role of keyword search technology in eDiscovery, the importance of statistics, and the need for transparency. Listen in as the panel discusses these and other important eDiscovery issues and feel free to share your feedback.

Does “splitting the pot” make sense?

In poker, two or more players might end up splitting the money (the pot) when they have the same hand. In this situation, neither party wins the hand, but neither party loses. Instead, they live to play another day. Listen to the panelists explore how transparency could be the key to a “middle ground” approach where neither party completely wins the discard pile issue, but neither party loses. The panel also discusses the role judges or special masters can play in ensuring a fair protocol without sacrificing traditional notions of privilege.

http://bcove.me/jinlk2rx

Does using keyword search in conjunction with predictive coding tools result in a stacked deck?

Some predictive coding advocates believe keyword search is dying a slow death in eDiscovery while others believe the proper application of keyword searching has simply changed. When should keyword searches be used in conjunction with predictive coding technology, if at all? Is the deck stacked against parties that insist on keyword culling prior to using predictive coding technologies? Should other technology tools in the litigator’s technology toolbelt be incorporated into predictive coding protocols? Hear from Ralph Losey about a case where keyword searching tools didn’t quite stack up to predictive coding technology and listen to Maura Grossman explain how the high cost of many predictive coding solutions can slow adoption of better technology approaches.

http://bcove.me/dlnnebe5

Will ignoring statistics and transparency ruin your game?

Every good poker player understands the important role statistics play when making decisions like how much you should bet or whether or not you should call your opponent’s bet. Basic statistical calculations can help players estimate the likelihood of beating their opponent in certain situations, but miscalculations or ignoring statistics completely can result in costly errors. Listen to Maura Grossman discuss basic statistical approaches that can make or break your predictive coding protocol and hear Craig Ball’s final word on the importance of transparency for both parties.

http://bcove.me/y1jh1o8g

You may not understand everything there is to know about predictive coding technology after watching these short video clips. However, you will receive valuable tips from industry experts to help you avoid playing a rigged game with a stacked deck. Or as Kenny Rogers might say, you will know when to walk away from a bad predictive coding game and you will know when to run.

On a recent trip to South Africa, where Symantec sponsored an event with PricewaterhouseCoopers (PwC) entitled The Protection of Personal Information (POPI) Drives Information Governance, customers and partners shared important insights. One major concern the attendees had was how they will comply with the newly proposed privacy legislation set to pass any day now.

POPI is the first comprehensive body of law addressing privacy in the country. Personal data is defined as a natural person’s name, date of birth, national identification number, passport number, health or credit information and other personally identifiable information. The bill has eight principles, each of which addresses aspects of how data must be collected, stored, processed, secured, expired and how access may be granted. This bill will apply to both public and private organizations and is driving the need for archiving, classification, eDiscovery, and data loss prevention technology.

Interestingly, the main motivator for purchasing eDiscovery technology will be the need for organizations in Africa to be able to conduct internal investigations to detect fraud. South Africa’s recent POPI legislation was crafted in order to address the age of digital information and the risks associated with it, but also to instill a level of confidence from the global economy in South Africa as a safe place to do business. A recent survey by Compuscan found that South Africa and Nigeria have the highest number of reported fraud cases in Africa. In addition, fraud related crimes have cost African businesses and governments at least $10.9 billion in 2011-12. Of the 875 reported cases, 40% of fraud perpetrators were in upper management.

Archiving the email of top management is a recommended best practice to address this fraud because it ensures that there will be a record of electronic communications should an investigation or lawsuit be necessary. Similarly, leveraging in-house eDiscovery and data loss prevention (DLP) technology enables investigators within the organizations to collect and analyze these emails in conjunction with other pertinent information to detect and even prevent fraud. To date, the majority of organizations in South Africa lack this kind of capability because they have not invested in technology.

Because corruption and fraud have been impediments to doing business in South Africa in the past, businesses and the government are taking steps to address these issues. Having the ability to conduct internal investigations will be a huge advantage for organizations looking to gain control over their information and those who commit fraud. PwC Partner Kris Budnik noted at the conference, “Many times when clients call me for an emergency forensic investigation, about 50% of the time in South Africa I cannot help them.  The reason for this is that the clients are not keeping the appropriate information governance systems in place and not keeping log files. Many times when we go to collect evidence, none is there because it has truly been overwritten in the data environment due to poor information governance practices.”

Litigation does not appear to be the biggest factor for purchasing eDiscovery technologies and implementing workflows as one might expect. The reason for this is unclear, but may be related to a less aggressive litigation profile as compared to that of the U.S. Much of the discovery in South Africa that involves electronically stored information is printed, reviewed and produced in paper format. The concern over retaining relevant metadata and reviewing/producing data in the format data was originally created does not seem to be top of mind for litigators.

Litigators in South Africa are not taking advantage of the rich information in metadata to supplement their cases or to challenge opposing counsel’s claims/productions. Also of concern is the inability to deduplicate and sort data once metadata is removed. The reason for this is most likely because there have not been enough cases where lack of metadata has been challenged. With time, and as cross-border litigation increases, there will be more demand for eDiscovery technology in the traditional legal context.

The increase in privacy concerns and internal fraud investigations presents a compelling reason for investing in archiving, eDiscovery, and DLP technologies for businesses in South Africa. Many organizations are moving data to the cloud to streamline POPI related objectives faster and because outsourcing their infrastructure is very attractive to organizations that don’t want to own the responsibilities of managing their information on premise. The main business drivers for cloud archiving in South Africa are: email continuity, cost and compliance.

It is interesting to observe how different countries and economies respond to technology and what drives use cases. The legal frameworks in each jurisdiction around the world vary, but the great equalizer will be technology. This is because whether it is privacy, litigation or fraud driving the information governance plan, the technology is the same.

Check out this article for more information on privacy legislation in South Africa.

Available soon: please visit our eDiscovery passport page for more information the legal system, eDiscovery, privacy and data protection in South Africa and other countries.

At eDiscovery 2.0, we have consistently followed the reports that Gibson Dunn has released on the state of eDiscovery. This is for good reason given its reputation as an excellent source of information on the trends affecting individual organizations and the industry as a whole.

The recently released 2012 annual report is no different, except that the overall tone is more positive. Instead of spotlighting the continuing problem of sanctions, the report showcases predictive coding and rules reform as the key eDiscovery trends leading into 2013. Describing these trends as being potential game-changers, the report also notes that “many questions remain” and warns that the impact of these trends may affect organizations in unanticipated and perhaps troubling ways.

Predictive Coding

In its report, Gibson Dunn happily indicates that unlike previous years, predictive coding technology appears to be ready for prime time. With several decisions from 2012 expressly or tacitly approving the use of predictive coding, the report speculates that many organizations and their counsel could become adopters of the technology. As the technology becomes more widespread and additional court decisions provide judicial imprimatur to the technology, the prospects increase that predictive coding could “drastically alter the way in which documents are reviewed for production.”

Nevertheless, challenges remain before this gradually increasing trend becomes a fully blown industry norm. While the promise of predictive coding is in its potential for rapid and cost effective document review, the report questions whether the technology will live up to that hype. Indeed, Gibson Dunn asks whether predictive coding is like any other review tool: “is it merely the latest review technology that, while useful, neither obtains widespread adoption nor revolutionizes the landscape?” Such questions are particularly legitimate given the substantial costs associated with most of the reported predictive coding cases. Indeed, the recent case of Gabriel Technologies v. Qualcomm exemplifies this predictive coding cost paradox.

Rules Changes

Another positive industry development that is fraught with questions concerns potential changes to the Federal Rules of Civil Procedure. As the report indicates, the federal Civil Rules Advisory Committee has made significant progress on a proposed draft amendment to Rule 37(e). Designed to broaden the existing protection against sanctions, the proposal would theoretically safeguard an organization’s pre-litigation destruction of information from sanctions in most circumstances. The lone exceptions would include destruction that was “willful or in bad faith and caused substantial prejudice in the litigation” or that “irreparably deprived a party of any meaningful opportunity to present a claim or defense.” While such a rule undoubtedly could reduce the costs and risks associated with ESI preservation, ambiguities in the draft language, together with statements in the draft advisory committee note, could ultimately water down the proposal’s intended protections.

Another encouraging rule change being considered by rules-makers includes an effort to better emphasize proportionality limitations on the Rule 26(b)(1) permissible scope of discovery. While characterized by the report as an “underused” though “increasingly important” doctrine, a proportionality amendment to Rule 26(b)(1) could do much to bring the problematic costs and delays of eDiscovery under control.

The report also sounded a note of caution on the proportionality front. Referring to the Sedona Conference’s recently updated Commentary on Proportionality in Electronic Discovery, the report observes that technology will be a key aspect in any proportionality analysis. With that background, Gibson Dunn cautions against misusing the efficiencies of cutting edge eDiscovery technologies to increase the scope of production under the guise that such discovery will comport with proportionality principles:

Other Industry Trends

Gibson Dunn spotlighted several other key trends from 2012 in its 33-page report. Among them were developments in cross-border eDiscovery, the increasing importance of foreign data protection laws, congressional attempts to bolster domestic privacy regulations, the correlation between ESI preservation and courts sanctions, and discovery of information found on social networking sites. These and other industry trends confirm that “progress is being made in addressing e-discovery’s challenges[,] [t]he dense fog that often seems to surround e-discovery appears to have lifted somewhat, and the collective anxiety lowered a little.”

Despite such sanguine observations, risks remain for organizations and their lawyers. The report concludes by raising the specter of sanctions, which will likely continue to be an unpredictable hobgoblin unless meaningful rules reform takes place. Until that time, clients and counsel alike should be proactive in adopting industry best practices to better ensure compliance with existing rules and jurisprudence.

The 7^th Circuit Pilot Program sponsored an educational mock hearing and expert panel discussion in Chicago last May to tackle important issues related to the use of predictive coding technology. The long awaited video footage of the event is finally here and available for review courtesy of Symantec.

The event begins with U.S. Chief Judge for the Northern District of Illinois, James F. Holderman, welcoming a courtroom packed full of people eager to learn more about novel issues presented by increased usage of predictive coding technology in litigation. National Archives Director of Litigation, Jason R. Baron, follows with opening remarks about the role of information retrieval in eDiscovery to set the stage for a lively mock hearing and panel discussion about a number of hot topics related to the use of predictive coding technology. Notable speakers include Maura R. Grossman, Counsel at Wachtell, Lipton, Rosen & Katz; Dr. David Lewis, co-founder of the TREC Legal Track; Ralph Losey, Partner at Jackson Lewis; Matt Nelson, eDiscovery Counsel at Symantec; Jeff Sharer, Partner at Sidley Austin; and Martin T. Tully, Partner and National eDiscovery Practice Group Chair, Katten Muchin Rosenman LLP.

The hypothetical hearing centers on a dispute between parties to a patent litigation matter regarding the use of predictive coding technology. Plaintiffs argue defendants should use predictive coding technology to assist with the production and review of documents. Defendants counter that they have a process in place for responding to discovery requests that is sufficient and that includes the use of legal technology approaches like keyword search that are commonly used during discovery. The hearing participants take positions (not necessarily their own) about important issues such as the reliability of predictive coding technology, steps needed to establish a protocol that is fair to both parties, and cost shifting. Ralph Losey does an excellent job playing the role of “judge” and summarizes key arguments made by each party before ruling from the bench at the conclusion of the hearing.

Following the mock hearing, Losey and others debated important issues related to the use of predictive coding as part of a lively panel discussion.  The panel discussion covered a broad range of interesting issues, but some of the liveliest discussion related to the following topics:

•  Should parties be required to disclose their use of predictive coding technology?
• Is it appropriate to use keyword searches to cull electronically stored information (ESI) prior to using predictive coding technology?
• Could the misapplication of statistics be the downfall of predictive coding?

The mock argument and panel discussion are among several excellent resources practitioners should consider reviewing to help them navigate a rapidly shifting and sometimes confusing predictive coding technology landscape.  Please feel free to share your comments and feedback below and be sure to visit the 7^th Circuit Pilot Program’s homepage for more information about the group’s efforts to help clarify some of the most complex and important eDiscovery issues facing litigators today.

This post was co-authored by Symantec’s Allison Walton, eDiscovery Counsel

A federal judge for the Southern District of California rang in the month of February by ordering plaintiffs in a patent related case to pay a whopping $12 million in attorney fees. The award included more than $2.8 million in “computer assisted” review fees and to add insult to injury, the judge tacked on an additional $64,316.50 in Rule 11 sanctions against defendants’ local counsel. Plaintiffs filed a notice of appeal on February 13^th, but regardless of the final outcome, the case is chock-full of important lessons about patent litigation, eDiscovery and the use of predictive coding technology.

The Lawsuit

In Gabriel Technologies Corp. v. Qualcomm Inc., plaintiffs filed a lawsuit seeking over $1 billion in damages. Among its eleven causes of action were claims for patent infringement and misappropriation of trade secrets.  The Court eventually dismissed or granted summary judgment in defendants’ favor as to all of plaintiffs’ claims making defendants the prevailing party and prompting Defendants’ subsequent request for attorneys’ fees.

In response to defendants’ motion for attorney fees, U. S. District Judge Anthony J. Battaglia relied on plaintiffs’ repeated email references to “the utter lack of a case” and their inability to identify the alleged patent inventors to support his finding that their claims were brought in “subjective bad faith” and were “objectively baseless.” Given these findings, Judge Battaglia determined that an award of attorney fees was warranted.

The Attorney Fees Award

The judge then turned to the issue of whether or not defendants’ fee request for $13,465,331.01 was reasonable. He began by considering how defendants itemized their fees which were broken down as follows:

• $10,244,053 for its outside counsel Cooley LLP (“Cooley”);
• $391,928.91 for document review performed by Black Letter Discovery, Inc. (“Black Letter”); and
• $2,829,349.10 for a document review algorithm generated by outside vendor H5.

The court also considered defendants’ request that plaintiffs’ local counsel be held jointly and severally liable for the entire fee award based on the premise that local counsel is required to certify that all pleadings are legally tenable and “well-grounded in fact” under Federal Rule of Civil Procedure 11.

Following a brief analysis, Judge Battaglia found the overall request “reasonable,” but reduced the fee award by $1 million. In lieu of holding local counsel jointly liable, the court chose to sanction local counsel in the amount of $64,316.50 (identical to the amount of local counsel’s fees) for failing to “undertake a reasonable investigation into the merits of the case.”

Three Lessons Learned

The case is important on many fronts. First, the decision makes clear that filing baseless patent claims can lead to financial consequences more severe than many lawyers might expect. If reviewed and upheld on appeal, counsel in the Ninth Circuit accustomed to fending off unsubstantiated patent or misappropriation claims will be armed with an important new tool to ward off would-be patent trolls.

Second, Judge Battaglia’s decision to order Rule 11 sanctions should serve as a wake-up call for local counsel. The ruling reinforces the fact that merely rubber-stamping filings and passively monitoring cases is a risky proposition. Gabriel Technologies illustrates the importance of properly monitoring lead counsel and the consequences of not complying with the mandate of Rule 11 whether serving as lead or local counsel.

The final lesson relates to curbing the costs of eDiscovery and the importance of understanding tools like predictive coding technology. The court left the barn door wide open for plaintiffs to attack defendants’ predictive coding and other fees as “unreasonable,” but plaintiffs didn’t bite. In evaluating H5’s costs, the court determined that Cooley’s review fees were reasonable because Cooley used H5’s “computer-assisted” review services to apparently cull down 12 million documents to a more reasonable number of documents prior to manual review. Although one would expect this approach to be less expensive than paying attorneys to review all 12 million documents, $2,829,349.10 is still an extremely high price to pay for technology that is expected to help cut traditional document review costs by as much as 90 percent.

Plaintiffs were well-positioned to argue that predictive coding technology should be far less expensive because the technology allows a fraction of documents to be reviewed at a fraction of the cost compared to traditional manual review. These savings are possible because a computer is used to evaluate how human reviewers categorize a small subset of documents in order to construct and apply an algorithm that ranks the remaining documents by degree of responsiveness automatically. There are many tools on the market that vary drastically in quality and price, but a price tag approaching $3 million is extravagant and should certainly raise a few eyebrows in today’s predictive coding market. Whether or not plaintiffs missed an opportunity to challenge the reasonableness of defendants’ document review approach may never be known. Stay tuned to see if these and other arguments surface on appeal.

Q. On a daily basis I discuss the EDRM (Electronic Discovery Reference Model) with customers, of course attributing credit to you two, and I find it is a very actionable and helpful paradigm to describe eDiscovery and Information Governance. Not too long ago, the EDRM was supplemented by the IGRM- can you explain the history of both? 

Tom: IGRM (Information Governance Reference Model) is a response to the growing focus on addressing arguably the most significant issue facing organizations with respect to managing internal information – getting their electronic data house in order. Many of the cost components associated with electronic discovery can be lessened by effective information governance policies and processes in place. The IGRM provides a framework to help organizations develop and implement these policies and processes that reflect the active involvement of the key constituencies in an organization – Legal, IT, Records, Management, line of business entities, and Privacy /Security.

George: As Tom noted, organizations need to get their electronic houses in order.  To do that, they need the appropriate people at the table, all equipped with a something to use as the starting point for their discussions and as means of focusing their efforts.  Our IGRM framework is meant to be that something.  The new IGRM framework grew out of the first box in the EDRM diagram, Information Management.  That box started in 2005 as Records Management, but in 2007 we changed the name to Information Management to reflect that broader scope implicated by eDiscovery.  We launched the Information Management Reference Model (IMRM) project in 2009 and in 2011 changed the name from Management to Governance.

Q. Can you explain the appropriate use case for each? 

Tom: Best to review materials posted at EDRM.net on IGRM including the IGRM Guide, Using the IGRM Model, and the whitepaper developed jointly between IGRM and ARMA International.

George: Use cases abound for both the EDRM and the IGRM frameworks.  The intent with both frameworks is to provide a structure that allows for many different use cases. For example, one dealing with legal hold issues will approach the EDRM framework from a different perspective and use it in a different fashion than one concerned with choosing appropriate forms of production.

Q. I also see Privacy has been added as a key component to the IGRM. Many of the customers we talk to outside of the U.S. don’t have the litigation drivers that prompt U.S. companies to purchase eDiscovery software. However, when a privacy breach occurs- litigation typically follows, therefore they benefit from implementing aspects of eDiscovery in-house. Moreover, archiving and document classification can enable compliance with the requirements of data protection and privacy. Can you explain why you added Privacy from your perspective and your thoughts surrounding this issue? 

Tom: Privacy and Security were added this past year in recognition of the growing importance of this function within organizations.

George: With the growing attention paid to areas such as data transfer and data privacy, security and privacy issues have been receiving heightened scrutiny.  Were these likely to be short-term issues only, we would not have modified the IGRM diagram to add Privacy and Security.  Indications are that these issues will continue to grow in importance for many years to come, hence the addition.

Q. Finally, since you are both at the forefront of eDiscovery, please enlighten our readership on your 2013 predictions. 

Tom: Increasing focus on Information Governance, defensible disposition of ESI, increasing interest in Computer Assisted Review (EDRM recently published the Computer Assisted Review Reference Model (CARRM) to bring clarity and common understanding of basic principles involved), consolidation of providers in the EDiscovery space and the arrival of new providers  - a continuation of what we have seen in the past couple of years.

George: I suspect that we are going to see a dramatic growth in efforts to redirect “traditional” e-discovery tools and techniques, pointing them toward the much larger set of issues encompassed by information governance.  While some providers will continue to absorb others, the growth in new providers will continue to outstrip and decrease coming from acquisitions, mergers, or collapses.  Costs will remain of great concern, but with luck their will be a greater emphasis on how to make better use of the ever-growing array of tools and techniques to accomplish what we really need to do more of in the e-discovery arena – support and enhance the ability to build and tell a persuasive story.