August 17, 2016
The e-discovery landscape mid-2016 generally looks much better than it did several years ago.
The big development in e-discovery sanctions in 2016 has been courts’ implementation of amended Federal Rule of Civil Procedure 37(e), which became effective in December 2015. While it is still early since the amended rule became effective, courts mostly appear to be applying its key provisions in a manner consistent with its intent.
Courts also appear to be faithfully implementing the requirement of amended Rule 26(b)(1) that discovery must be both relevant and proportional, with courts repeatedly holding that merely establishing relevance but not proportionality is not enough. Although courts once implicitly permitted broad “fishing expeditions,” they are now explicitly prohibiting them.
New sources of potentially discoverable ESI, such as text messaging and social media, have created new risks and difficulties for identification, legal hold preservation and collection. Indeed, many of the sanctions decisions in the first half of 2016 have involved failures to preserve text messages on mobile devices. Additionally, while there was a relative dearth of technology assisted review decisions, the two that were issued were nevertheless significant.
Advances in e-discovery and database software promise to significantly reduce the costs and burdens to litigants. SaaS-based e-discovery solutions are making it more feasible for law firms and clients to handle e-discovery tasks without the involvement of a traditional e-discovery services provider in all but the largest matters.
As always, the e-discovery field is constantly and rapidly evolving. The good news, as reflected in the discussion below, is that most of the change during the first half of 2016 has been for the better.
Amended Federal Rule of Civil Procedure 37(e) appears to be having a substantial impact in the sanctions area. Key provisions of amended Rule 37(e)–i.e., the requirement of an intent to deprive for the most serious sanctions under Rule 37(e)(2), the Rule’s application only to ESI, the requirements of an actual loss of ESI and its unavailability from other sources, and the requirement of prejudice to the moving party–played out in a number of cases during the first half of 2016.
As former Magistrate Judge John Facciola recently observed, “the rule is ‘simple’ and relatively easy to apply” and, as a result, “opinions are getting shorter and more concise. With a single, national rule, “‘[t]here’s no longer the necessity for a lengthy account of each circuit’s understanding of how and when to sanction parties,’ according to Judge Facciola.”
The decision in Nuvasive, Inc. v. Madsen Medical, Inc. is a powerful example of the impact of the new rule. Before amended Rule 37(e) became effective, the court in Nuvasive had granted the sanction of an adverse inference jury instruction because the plaintiff had failed to preserve discoverable text messages. The court imposed the sanction even though it found that the plaintiff had not acted with the intent to deprive the defendant of the text messages. After the amended rule became effective, the plaintiff in Nuvasive moved to vacate the previous order. The court granted the motion, holding that an adverse inference instruction is not permitted under the amended rule without finding an intent to deprive.
In Best Payphones, Inc. v. New York, defendants argued that the plaintiff failed to preserve both hard copy documents and ESI. Because amended Rule 37(e) only applies to ESI, the court applied the Second Circuit’s existing sanctions standards to the lost hard copy documents and Rule 37(e) to the lost ESI. Although the court found that the plaintiff had been negligent in its failure to properly implement a legal hold, and that the lost information was relevant, it nevertheless declined to impose sanctions because defendants did not demonstrate that they suffered prejudice from the loss of the information.
In Living Color Enterprises, Inc. v. New Era Aquaculture, Ltd., the plaintiff sought case- terminating sanctions or, alternatively, an adverse inference jury instruction against an individual defendant because he failed to preserve discoverable text messages on his cell phone. The defendant contended that he always used the cell phone feature that automatically deletes text messages after 30 days and that he had neglected to disable the feature when the lawsuit was filed. The court found, however, that the vast majority of the defendant’s text messages were produced from other sources and those that remained lost were unimportant. Having found no loss of important text messages, and therefore no prejudice to the plaintiff, the court denied sanctions under Rule 37(e).
Similarly, in FiTeq Inc. v. Venture Corp., the court denied the plaintiff’s motion for an adverse inference jury instruction under amended Rule 37(e) because emails deleted by the defendant’s executive vice president were recovered from his old computer and also obtained and produced from the email accounts of others. The court relied on amended Rule 37(e)’s provisions that sanctions can only be imposed where the ESI “cannot be restored or replaced through additional discovery” and where there has been a “finding [of] prejudice to another party from the loss of the information.”
In Matthew Enterprise, Inc. v. Chrysler Group, the plaintiff failed to preserve emails and customer communications after its duty to preserve was triggered when it threatened litigation in a letter to the defendant. In particular, the plaintiff allowed all of the email communications on its email system to be deleted when it changed email vendors, and it failed to notify its customer communications vendor to suspend its auto-delete function. The court found that the plaintiff failed to take reasonable steps to preserve discoverable information, and that the defendant was prejudiced as a result. It granted curative measures–for example, that the defendant could present evidence and argument regarding the spoliation. But the court denied the requested sanction of evidence preclusion because it “would effectively decide the case,” finding that absent an intent to deprive this would be an excessive and inappropriate remedy.
In several cases in the first half of 2016, courts found that the sanctioned party intended to deprive other parties of discoverable ESI, warranting harsh sanctions under Rule 37(e)(2). In Brown Jordan International, Inc. v. Carmicle, the court found that the defendant–among other things–remotely wiped his company-owned iPad and locked his company-owned laptop computer, refusing to unlock it even up until trial. He also failed to preserve relevant data on his personal iPad, his personal laptop computer, his personal iPhone and on his wife’s laptop computer. Pursuant to amended Rule 37(e), the court allowed adverse evidentiary inferences to be made regarding the lost ESI, but declined to grant case terminating sanctions.
In GN Netcom, Inc. v. Plantronics, Inc., the court found that a senior executive intentionally deleted thousands of emails for the purpose of making them undiscoverable in the litigation, and that he ordered others to do so as well. Although the company had taken a number of steps to preserve documents–including distributing quarterly legal hold notices and conducting training sessions to ensure compliance–the court held that, because of the executive’s intentional spoliation, it could not be deemed to have acted “reasonably” in preserving documents, which would have resulted in no relief pursuant to the terms of amended Rule 37(e).
The court agreed with the plaintiff that the defendant company’s “reliance on these actions to excuse the intentional, destructive behavior of Mr. Houston requires a ‘perverse interpretation’ of Rule 37(e), one which would set a dangerous precedent for future spoliators.” The court ordered an array of sanctions against the defendant company under Rule 37(e)(2), including an adverse inference jury instruction, monetary sanctions in the form of reasonable fees and costs to the plaintiff, a “punitive” sanction of $3 million, and possible evidentiary sanctions to be determined before trial.
In InternMatch, Inc. v. Nxtbigthing, LLC, the court found that the defendant falsely claimed that a power surge had destroyed computers containing relevant evidence. The company failed to make any effort to determine if ESI was salvageable from the computers before discarding them. Additionally, eight days before the claimed power surge, the company’s owner had called his insurer to find out how a power surge would affect an insurance payout. Pursuant to amended Rule 37(e), the court granted an evidence preclusion order, an adverse inference instruction, and attorneys’ fees, but declined to impose case terminating sanctions.
In O’Berry v. Turner, which involved an accident between a big-rig truck and a car, the court found that the defendant’s preservation failures constituted an intent to deprive the plaintiff of relevant ESI in the litigation. Although the plaintiff sent a preservation letter, and the defendant sent a response that acknowledged receipt and stated that it would take the necessary measures to preserve evidence, the defendant failed to preserve relevant ESI (in this case, an electronic driver’s log and data from an application that recorded information about the truck). Rather, the defendant printed out a single, hard copy of the data, which it subsequently could not locate. The court found that, in failing to take reasonable steps to preserve ESI, the defendant acted with an intent to deprive the plaintiff of its use in the litigation, and it imposed the severe sanction of an adverse inference jury instruction.
Finally, at least two courts have taken somewhat inconsistent approaches to whether they can still issue sanctions based upon their inherent authority following the enactment of amended Rule 37(e). The Rules Advisory Committee expressly intended that the amended rule would preclude inherent authority, stating in the Advisory Committee’s notes that “[n]ew Rule 37(e) . . . authorizes and specifies measures a court may employ if information that should have been preserved is lost, and specifies the findings necessary to justify these measures. It therefore forecloses reliance on inherent authority or state law to determine when certain measures should be used.”
Relying on the Advisory Committee note, the court in FiTeq Inc. v. Venture Corp. denied the plaintiff’s motion seeking sanctions under inherent authority. The court in in CAT3, LLC v. Black Lineage, Inc., however, held that it had inherent authority to impose sanctions despite the language in the Advisory Committee note, relying on Supreme Court precedent stating that “‘the inherent power of a court can be invoked even if procedural rules exist which sanction the same conduct.'”
Whether courts’ inherent powers to impose spoliation sanctions persist despite the statement to the contrary in the Advisory Committee’s note to the 2015 amendment will likely be a hot topic in the future. And the outcome of this issue may very well determine whether amended Rule 37(e) succeeds in correcting many of the shortcomings of prior e-discovery sanctions jurisprudence.
Amended Rule 26(b)(1) and Proportionality
The other big development arising out of the December 1, 2015 FRCP amendments is how much traction proportionality has gained in the first half of this year. Amended Rule 26(b)(1) incorporates in the scope of discovery an explicit requirement that discovery must be “proportional to the needs of the case.”
For years, courts have grappled with ways to address the burden of costly electronic discovery that often fall to larger litigants, such as through fee shifting. Although proportionality was previously contained within the rules, this amendment to Rule 26(b)(1) puts it front and center in describing what is discoverable. The amendment makes discoverable only information that is proportional and relevant to any party’s claim or defense, rather than the prior standard that made discoverable any information relevant to any party’s claim or defense (and, upon a court order for good cause, relevant to the subject matter).
On its face, this change to Rule 26(b)(1) may mark a sea change in the way litigants and courts approach discovery disputes–a shift away from broad, amorphous inquiries into subject matter relevance to a more practical approach that balances the need for and likely use of the information against the burden of collecting and producing it. And the amendment appears to have garnered a good deal of traction with the courts to date.
In Noble Roman’s, Inc. v. Hattenhauer Distributing Company, for example, the defendant served document and deposition subpoenas upon a “major shareholder” of the plaintiff, seeking production of 23 “wide-ranging” categories of documents, essentially asking for “every document and every piece of information it has … about every aspect of [Plaintiff’s] business operations, finances, marketing plans, and management structure.” The plaintiff asserted that the discovery was an “improper fishing expedition” and “outside the proper bounds of discovery.”
The court invoked the proportionality requirement of amended Rule 26(b)(1) and stated that the defendant “beats the drum of ‘relevancy'” but “never attempts to demonstrate that the discovery is in any way proportional to the needs of the case, considering such things as the amount in controversy, the importance of the information in resolving contested issues, whether the burden of the discovery outweighs its likely benefits, whether the information can be obtained from other and more convenient sources, or whether the information is cumulative to other discovery[.]” The court stated “[t]hat’s not good enough.”
The court observed that while the proportionality amendment to Rule 26(b)(1) was a “structural and linguistic alteration” of the already-existing proportionality provision in Rule 26, it was “designed to protect against over-discovery and to emphasize judicial management of the discovery process.” The court held that, in this case, the broad document and deposition subpoenas constituted “discovery run amok” and that they “fail[ed] the proportionality test” of amended Rule 26(b)(1).
Gilead Sciences v. Merck is another case in which the court invoked the proportionality requirement of amended Rule 26(b)(1) in limiting discovery. In finding certain document requests were disproportionate to the needs of the case, the Court stated that under the amended rule “[n]o longer is it good enough to hope that the information sought might lead to the discovery of admissible evidence. In fact, the old language to that effect is gone. Instead, a party seeking discovery of relevant, non-privileged information must show, before anything else, that the discovery sought is proportional to the needs of the case.”
While the amended rule provides that discovery must be proportional and provides a list of factors to be considered in the proportionality analysis, courts still have broad discretion in interpreting and applying those factors. Some courts have stated that the amendments do nothing to change their evaluation of discovery disputes, and merely emphasize that courts should apply the same balancing test that they already applied.
Amended Rule 26(b)(1) does not specify how the proportionality factors are to be balanced, or whether any factor is outcome determinative. Given the relative infancy of the amendment (now only 8 months old), courts have only just begun to navigate their application. Over the next months and years, we expect to see more case law interpreting and applying the proportionality factors, giving courts and litigants additional insight into how amended Rule 26(b)(1) will be applied in the future.
Text Messaging and Mobile Devices
A plethora of decisions involving failures to produce text and instant messages, and other data from mobile devices, reflects that these data sources are becoming increasingly important in e-discovery. See Vera Nackovic & Anita Lam, “OMG! My text messages could be discoverable?”, Inside Counsel (Nov. 30, 2015).
Employees are regularly using text messaging for business purposes. One survey commissioned by RingCentral found that 80 percent of people use texting for business, and 15 percent of people surveyed said that over half of their text messages were business-related. Making matters more complex, many of these communications are not occurring on company-issued phones. The percentage of businesses with bring-your-own-device (“BYOD”) programs increased from 58 percent to 73 percent in 2015.
These trends make it increasingly important for companies to think about how to manage and handle employee text messages and mobile data. Nevertheless, more than two thirds of businesses report that they have no archiving or supervision solution in place for text messages used for business communications.
With text messages increasingly used in business communications, the trend of courts requiring parties to produce text messages along with other phone communications continued in the first half of 2016. In fact, many of the sanctions decisions this year have involved text messages.
For example, in NuVasive, Inc. v. Madsen Med., Inc., the court awarded relief under amended Rule 37(e) for the defendants’ failure to preserve four individuals’ text messages. The court had previously awarded an adverse inference jury instruction. It reduced the sanction to allowing the parties to present evidence to the jury regarding the loss of the text messages and their relevance because there was no evidence that the defendants’ had failed to preserve them with the requisite intent to deprive the plaintiff of their use in the litigation.
In Living Color Enterprises v. New Era Aquaculture, the defendant failed to suspend a feature on his smart phone that automatically deleted text messages after 30 days. The court held that the defendant “clearly had an obligation to retain the relevant text messages after this lawsuit was initiated.” Nevertheless, it declined to award an adverse inference jury instruction or other harsh sanctions because it found that there was not intent to deprive. “Defendant is an individual who appears to be a relatively unsophisticated litigant. At worst, his actions were negligent.”
In Stinson v. New York City, the court awarded an adverse inference instruction against the New York City for spoliating relevant text message evidence by failing to institute a litigation hold for the first three years of litigation and then failing to enforce the hold or collect relevant cell phone data once it was in place. In Timms v. LZM, LLC, the Fifth Circuit affirmed the trial court’s award of case terminating sanctions and attorneys’ fees after plaintiff produced her cell phone pursuant to a court order, but had removed relevant text messages and a messaging app from the device.
Similarly, the court in First Financial Security, Inc. granted an adverse inference sanction and attorneys’ fees after the defendants failed to produce text messages in compliance with the court’s discovery order. Rejecting the defendants’ claim that they had failed to comply with the order because of their own lack of sophistication, the court stated that “[i]t defies belief that experienced users such as defendants would have trouble locating text messages on their own phones (or asking someone to help them do so).”
The growing importance of text and mobile data in discovery makes companies’ policies even more critical for effective preservation. Companies may mitigate the risks associated with discovery of text messaging by establishing BYOD policies, implementing “sandboxes” (separate spaces for work apps) on employees’ mobile devices, and requiring the use of enterprise texting apps that are journaled or backed up onto company servers. We expect that preservation of relevant text messaging and mobile data will continue to be increasingly important for the rest of 2016 and beyond. See Gareth Evans & Veeral Gosalia, The Coming Storm: Companies Must Be Prepared to Deal With Text Messages on Employee Mobile Devices, 15 Digital Discovery & e-Evidence (Bloomberg BNA, June 25, 2015.
Technology Assisted Review
TAR Case Law
To date this year, the landscape has been relatively quiet with respect to predictive coding and other forms of technology assisted review (“TAR”). In the U.S., there have been only two reported or publicized decisions of which we are aware. Nevertheless, these decisions–Hyles v. New York and Dynamo Holdings II–are both important. And, demonstrating that TAR is becoming global, courts in England and Ireland have approved its use.
Hyles v. New York
In Hyles v. New York, Magistrate Judge Andrew Peck of the Southern District of New York recently added a third decision regarding predictive coding to his previous decisions in Da Silva Moore (the first case to approve the use of predictive coding, decided in 2012) and Rio Tinto (clarifying that there are alternatives to sharing training sets to ensure a reasonable predictive coding process, decided in 2015).
In Hyles, an employment discrimination case, Judge Peck concluded that defendant New York City could not be compelled to use predictive coding against its wishes. The plaintiff sought to require New York City to use predictive coding to review the documents of nine custodians initially (with the number potentially expanding to 15). In addition to concerns that the parties would not be able to agree upon a predictive coding protocol, New York City cited high vendor pricing for predictive coding, as well as its own low review costs (it would use in-house salaried staff of the City’s Corporation Counsel’s Office for document review), as reasons why it preferred to use search terms and human review.
Judge Peck agreed with the plaintiff that, “in general, TAR is cheaper, more efficient and superior to keyword searching.” In addressing whether the plaintiff could compel its use, Judge Peck referenced his statement in Rio Tinto that “the case law has developed to the point that it is now black letter law that where the producing party wants to utilize TAR for document review, courts will permit it.” Judge Peck also referenced his statement in Rio Tinto that, “in contrast, where the requesting party has sought to force the producing party to use TAR, the courts have refused.”
Judge Peck noted, however, that in the cases refusing to force the producing party to use TAR, the producing party had already “spent over $1 million using keyword search (in Kleen) or keyword culling followed by TAR (in Biomet)[.]” By contrast, in Hyles, the producing party had not yet invested much, if anything, on a document search and review, thus raising the issue of whether a court can force a responding party at the outset of litigation to use TAR. Judge Peck declared that “[t]he short answer is a decisive ‘NO,'” although he suggested that the answer could be different if the production is later found to be deficient.
Judge Peck explained that, although “[i]t is certainly fair to say that I am a judicial advocate for the use of TAR in appropriate cases[,] I also am a firm believer in the Sedona Principles, particularly Sedona Principle 6[.]” Sedona Principle 6 provides that “[r]esponding parties are best situated to evaluate the procedures, methodologies, and technologies appropriate for preserving and producing their own electronically stored information.” Furthermore, although Judge Peck “believe[s] that parties should cooperate in discovery,” he nevertheless found that “[c]ooperation principles . . . do not give the requesting party, or the Court, the power to force cooperation or to force the responding party to use TAR.”
Accordingly, Judge Peck reasoned that, “[u]nder Sedona Principle 6, the City as the responding party is best situated to decide how to search for and produce ESI responsive to Hyles’ document requests.” Although the City might have to re-do its search if the plaintiff later demonstrates deficiencies in the City’s production, he nevertheless concluded, “that is not a basis for Court intervention at this stage of the case.”
Judge Peck concluded that “it is not up to the Court, or the requesting party (Hyles), to force the City as the responding party to use TAR when it prefers to use keyword searching. While Hyles may well be correct that production using keywords may not be as complete as it would if TAR were used, the standard is not perfection, or using the ‘best’ tool, but whether the search results are reasonable and proportional.”
Dynamo Holdings II
The decision in Hyles dovetails with the other TAR decision to date this year, Dynamo Holdings II, in which the tax court addressed a post-production challenge to the sufficiency of a TAR process.
In Dynamo Holdings I, the tax court had previously authorized the petitioners’ use of predictive coding over the objection of respondent Commissioner of Internal Revenue, which had characterized it as an “unproven technology”–a label that the court rejected. The court in Dynamo Holdings I reasoned that it “is not normally in the business of dictating to parties the process that they should use when responding to discovery” and observed that “the technology industry now considers predictive coding to be widely accepted for limiting e-discovery to relevant documents and effecting discovery of ESI without an undue burden.” Although it permitted the use of TAR, the court stated that, “[i]f, after reviewing the results, respondent believes that the response to the discovery request is incomplete, he may file a motion to compel at that time.”
The Commissioner of Internal Revenue subsequently filed a motion to compel contending that the petitioners’ production was incomplete, which the tax court ruled upon in Dynamo II. Specifically, the Commissioner argued that the production was missing 1,353 documents identified when the petitioners ran against the document population 76 search terms that the Commissioner had provided to the petitioner to use in the training process. The petitioner largely debunked this claim, showing that only 765 such documents hitting the search terms were not produced and, of those, sampling showed that many fell outside the relevant date range of the case.
Nevertheless, the tax court stated that “[w]e will assume that [the predictive coding response] was flawed, but the question remains whether any relief should be afforded.” The court decided in the negative, holding that the petitioners made a “reasonable inquiry” in using predictive coding and producing documents “that the algorithm determined was responsive[.]”
In particular, the court reasoned that the respondent’s motion was “predicated on two myths.” The first, according to the court, is the “myth of human review,” i.e., “that manual review by humans of large amounts of information is as accurate and complete as possible–perhaps even perfect–and constitutes the gold standard by which all searches should be measured.” Citing studies, the court stated that “research shows that human review is far from perfect.”
The second myth, according to the court, “is the myth of a perfect response.” The court stated that the respondent was “seeking a perfect response, but our Rules do not require a perfect response.” Specifically, in response to discovery requests, Tax Court Rule 70(f)–which is analogous to Federal Rule of Civil Procedure 26(g)–“requires the attorney to certify, to the best of their knowledge formed after a ‘reasonable inquiry,’ that the response is consistent with our Rules, not made for an improper purpose, and not unreasonable or unduly burdensome given the needs of the case.” The court stated that “when the responding party is signing the response to a discovery demand, he is not certifying that he turned over everything, he is certifying that he made a reasonable inquiry and to the best of his knowledge, his response is complete.”
The court quoted Judge Peck’s decision in Rio Tinto that “it is inappropriate to hold TAR to a higher standard than keywords or manual review.” It concluded that “there is no question that petitioners satisfied our Rules when they responded using predictive coding.”
England and Ireland
Demonstrating that TAR is becoming accepted globally, courts in England and Ireland this year have approved the use of predictive coding for document search and review in litigation.
In Pyrrho Investments Ltd. v. MWB Property Ltd., the parties jointly sought the approval of the English High Court of Justice to use TAR. The court approved the request, identifying ten factors favoring the use of predictive coding and none against its use.
Among those ten factors were the following: (1) that “[e]xperience in other jurisdictions, whilst so far limited, has been that predictive coding software can be useful in appropriate cases;” (2) “[t]here is no evidence to show that the use of predictive coding software leads to less accurate disclosure being given than, say, manual review alone or keyword searches and manual review combined, and indeed there is some evidence . . . to the contrary;” and (3) “there will be greater consistency in using the computer to apply the approach of a senior lawyer towards the initial sample (as refined) to the whole document set, than in using dozens, perhaps hundreds, of lower-grade fee-earners, each seeking independently to apply the relevant criteria in relation to individual documents.”
The court in Pyrrho also cited the large volume of electronic documents to be reviewed (more than 3 million) and the high estimated cost of manual review (“several million pounds at least,” compared with estimates for predictive coding between approximately 200,00 to 500,000 pounds).
In Brown v. BCA Trading,  the English High Court of Justice approved the use of predictive coding over the objection of the requesting party. In doing so, the court found that most of the same 10 factors cited in Pyrrho Investments mitigated in favor of using predictive coding and there were no factors against its use. The court noted that the respondents estimated that the cost of using predictive coding would be at least approximately half that of using key word searches. Additionally, the court indicated that some deference to the respondents’ preferred approach was appropriate: “it is relevant to take into account when considering the Respondents’ assertion, presented from their own view and on advice received professionally, that they think predictive coding will be the most reasonable and proportionate method of disclosure.”
Finally, in February 2016, the Irish Court of Appeal in Irish Bank Resolution Corp. v. Quinn upheld the 2015 decision of the Irish High Court granted a responding party’s motion to use TAR over the requesting party’s objection. In Quinn, the Irish High Court rejected the requesting party’s assertion that the responding party was required to produce 100% of the relevant documents. Rather, it found that no method is guaranteed to find all relevant documents, but studies presented to the court showed that predictive coding performed better than key word search and manual review.
The requesting party repeated that argument on appeal, and the Irish Court of Appeal also rejected it. According to one observer, “[Ms Justice Finlay Geoghegan] said that the courts have addressed the necessity to use technology given the challenges presented by large discoveries. The Court was satisfied that what was approved by the High Court was fair, proportionate and took account of the increasing need for the courts to ensure that discovery is complied with in a cost effective manner.”
The trend of low, but steadily increasing, utilization of TAR in civil litigation that has existed for some time appears to be continuing. By contrast, TAR is very frequently used in responding to Federal Trade Commission or Justice Department Antitrust Division “second request” inquiries regarding merger transactions.
There are likely a number of reasons for the low take-up rate of predictive coding in civil litigation: Lack of familiarity with the technology; lawyers’ comfort with the usual playbook of using search terms; concerns regarding expensive disputes and motion practice regarding the TAR protocol; the prospect of potentially having to share training and validation sets (particularly otherwise undiscoverable documents in the training and validation sets); and the additional up-front expense of using TAR software. We suspect that the first two factors–lack of familiarity with TAR and lawyers’ comfort with search term methodologies–are the primary drivers.
Whether any development will drive the use of TAR to a tipping point remains to be seen. It could be developments in the technology; the case law; a new generation of lawyers open to departing from the “tried and true” methodology of search terms combined with human review; or more clients willing to accept the potential risks associated with predictive coding to achieve potentially significant time and cost savings (and potentially greater accuracy). In Hyles, Judge Peck stated that “[t]here may come a time when TAR is so widely used that it might be unreasonable for a party to decline to use TAR. We are not there yet.” Indeed, such a time appears to be, at a minimum, a long way off. As we are experiencing in many areas involving technology, however, change can come quickly and unexpectedly.
Below the surface, an evolution has been occurring on the TAR software front. Last year, the software vendor with the most widely-used “off the shelf” predictive coding software was acquired, with plans to repurpose the technology for non-ediscovery uses. During the first half of this year, a second prominent vendor of single-purpose predictive coding software was acquired, with plans to repurpose the technology as well.
Consequently, a perception among many e-discovery service providers has arisen that off-the-shelf predictive coding software that is generally available for licensing will either no longer be available or will not be adequately developed and supported in the future. In response, more service providers have begun to develop and offer their own predictive coding applications, rather than rely upon off-the-shelf software. Indeed, a gap may develop in the future among e-discovery service providers between “haves” and “have-nots” of predictive coding software if other software vendors do not step in to offer TAR applications available for license. Service providers that are unable to develop their own TAR applications may be left with an inferior TAR offering or no offering at all.
Meanwhile, predictive coding software technology continues to develop slowly. Continuous active learning-based software has received favorable mention (including in Rio Tinto and Hyles) for reportedly training the predictive coding algorithm more quickly, yielding more accurate results, and helping to moot disputes over sharing training sets. Yet the number of vendors offering this technology can be counted on just a few fingers (some would say just a couple). The same can be said for visual analytics, which allow visibility into documents regarding the same and similar topics and concepts. Unless and until these tools become more widely available, those offering them should experience a competitive advantage in cases where the technology can make a significant difference.
E-Discovery Vendor and Technology Developments
Continuing consolidation has been the most apparent trend in the e-discovery vendor space. In the first half of 2015, there were more blockbuster combinations of large e-discovery service providers and software vendors.
Changes in the market are squeezing service providers, in particular. On small and medium-sized matters, corporations and law firms–aided by the growing availability of cloud-based e-discovery software tools and data hosting–increasingly are able to handle e-discovery services themselves. E-discovery service providers therefore must focus on larger matters, which require not only greater depth and resources but also, very often, global coverage.
The search and review platform of one software provider has become dominant, and most other basic search and review platforms appear to be on their way to all-but disappearing. With the market largely coalescing around one basic search and review tool, many service providers have little other than pricing to differentiate themselves from competitors. The result is that service providers increasingly must compete on price, particularly if they cannot differentiate themselves with unique add-on software solutions such as visual analytics.
At the same time, a number of small, cloud-based providers of e-discovery software as a service (SaaS) have entered the market in recent years, largely targeting small and medium-sized law firms and companies with “do-it-yourself” e-discovery solutions. This business model is a reflection of the fact that e-discovery is becoming ubiquitous in all forms of litigation, even small cases. Additionally, a new generation of lawyers is more accustomed to and comfortable with undertaking themselves tasks involving the use of technology.
Some of the smaller SaaS providers are beginning to target larger matters and are venturing into providing predictive coding and analytics capabilities. Additionally, several large e-discovery service providers are now providing SaaS offerings too. The risk of such offerings in larger matters, however, is that the end-users may not have sufficient expertise to do so effectively in a large, complex matter.
Advances in database technology using “NoSQL” software are also making dramatic improvements in handling large volumes of data possible. The technology allows horizontal scaling of databases to clusters of servers, which is difficult for traditional relational databases, and significantly improved speeds in performing various operations, such as data ingestion, processing, and especially search–all at lower cost. It also simplifies database management, which should make it less labor-intensive.
Accordingly, handling “big data” in e-discovery at reasonable costs is becoming a real possibility for providers with the right technology. The use of NoSQL databases should also drive down pricing on a volume basis for ingestion, processing, hosting and database management in all types of matters.
Perhaps most importantly, NoSQL approaches enable very significant improvements in understanding large data sets through data mining and visual analytics. Service providers that offer big data analytics tools in addition to the latest database technology will therefore have a significant advantage over those that do not.
E-discovery software vendors–and service providers with their own, cutting-edge proprietary software–appear to be the big potential winners. Those offering SaaS products that provide corporations and law firms with a do-it-yourself solution will increasingly take market share away from small and medium-sized service providers. Smaller and medium-sized providers also will find it increasingly difficult to compete if they cannot provide the latest technologies. Even large vendors, if they cannot offer powerful database and analytics software, will likely struggle.
There are still unmet needs in the e-discovery software area, however. As companies increasingly look to implement e-discovery as a business process, the “holy grail” for them is a single end-to-end software solution incorporating legal hold notice, preservation, collection, processing, search, review and production functionalities to handle most day-to-day matters. Software vendors are increasingly able to effectively combine several aspects of the e-discovery process in their products–for example, combining legal hold notice management, preservation and collection; or, alternatively, combining processing, search, review and production.
Nevertheless, no single software vendor appears to be providing a solution combining what the market considers to be “best in breed” technologies in all of these areas. We expect, however, that, as you read this, software developers are out there working furiously to provide such a solution.
Overall, the year is shaping up to be a good one for e-discovery. A more rational framework for addressing lost or destroyed ESI and the requirement that discovery must be proportional are already having meaningful, positive impacts. And advances in e-discovery software promise to make the process simpler, faster and less expensive. Of course, it will likely take years for these trends to fully develop. But we are certainly off to a good start in 2016.
 Tera Brostoff, Facciola Optimistic About eDiscovery’s New Faces, Rules,” 16 Digital Discovery & e-Evidence 280 (Jun. 14, 2016).
 CAT3, LLC v. Black Lineage, Inc., 2016 WL 154116 , at *7 (S.D.N.Y. Jan. 12, 2016) (quoting Chambers v. NASCO, Inc., 501 U.S. 32, 42-43 (1991); see also id., quoting Haeger v. Goodyear Tire & Rubber Co., 793 F.3d 1122, 1131-32 (9th Cir. 2015) (“This inherent power is not limited by overlapping statutes or rules.”) .
 Noble Roman’s, Inc. v. Hattenhauer Distrib. Co., 314 F.R.D. 304, 311-12 (S.D. Ind. 2016).
 See id., 314 F.R.D. at 311.
 See id., 314 F.R.D. at 308.
 Gilead Sciences, Inc. v. Merck & Co., Inc., Case No. 5:13-cv-04057, 2016 WL 146574 (N.D. Cal. Jan. 13, 2016).
 See, e.g, Vay v. Huston, 2016 WL 1408116 (W.D. Pa. Apr. 11, 2016) (finding that discovery was proportional because defendant had greater resources).
 See, e.g., Hahn v. Hunt, 2016 WL 1587405 at *2 (E.D. La. Apr. 20, 2016).
 NuVasive, Inc. v. Madsen Med., Inc., 2016 WL 305096, at *1 (S.D. Cal. Jan. 26, 2016)
 Id. at *2.
 Living Color Enters. v. New Era Aquaculture, Ltd., 2016 WL 1105297, at *2 (S.D. Fla. Mar. 22, 2016).
 Id. at *6.
 Stinson v. City of New York, 2016 WL 54684, at *8 (S.D.N.Y. Jan. 5, 2016).
 Timms v. LZM, LLC, 2016 WL 3611550, at *2-*3 (5th Cir. Jul. 05, 2016).
 First Fin. Sec. Inc. v. Lee, 2016 WL 881003, at *7 (D. Minn. Mar. 8, 2016).
 Id. at *5.
 Id. (quoting The Sedona Principles: Second Edition, Best Practices Recommendations & Principles for Addressing Electronic Document Production, Principle 6 (available at www.TheSedonaConference.org).
 Id., slip op. at 7-8 (citing Nicholas M. Pace & Laura Zakaras, RAND Corp., Where the Money Goes: Understanding Litigant Expenditures for Producing Electronic Discovery (2012) at 55, 58 (summarizing studies)).
 Pyrrho Inv. Ltd. v. MWB Prop. Ltd.,  EWHC (Ch) 256 (Eng.).
 Irish Bank Resol. Corp. v. Quinn,  IEHC 175 (H. Ct.) (Ir.), upheld by the Irish Court of Appeal (see Court of Appeal Approves use of TAR for Discovery, McCann Fitzgerald (2016), http://www.mccannfitzgerald.com/McfgFiles/knowledge/6802-Court%20of%20Appeal%20Approves%20Use%20of%20Tar%20For%20Discovery.pdf.
 Court of Appeal Approves use of TAR for Discovery, McCann Fitzgerald (2016), http://www.mccannfitzgerald.com/McfgFiles/knowledge/6802-Court%20of%20Appeal%20Approves%20Use%20of%20Tar%20For%20Discovery.pdf.
The following Gibson Dunn lawyers assisted in the preparation of this client update: Gareth Evans, Jennifer Rearden, Heather Richardson, Matthew Kahn and Katherine Warren.
Gibson Dunn’s lawyers are available to assist in addressing any questions you may have regarding the issues discussed in this update. The Electronic Discovery and Information Law Practice Group brings together lawyers with extensive knowledge of electronic discovery and information law. The group is comprised of seasoned litigators with a breadth of experience who have assisted clients in various industries and in jurisdictions around the world. The group’s lawyers work closely with the firm’s technical specialists to provide cutting-edge legal advice and guidance in this complex and evolving area of law. For further information, please contact the Gibson Dunn lawyer with whom you usually work or the following leaders of the Electronic Discovery and Information Law Practice Group:
© 2016 Gibson, Dunn & Crutcher LLP
Attorney Advertising: The enclosed materials have been prepared for general informational purposes only and are not intended as legal advice.