Maine Evidence and Electronic Discovery

Another Rare Appellate Decision Vacating E-Discovery

It's always nice, although rare, to see an appellate court reverse a lower court's decision to permit some form of electronic discovery.  It doesn't happen often at all, and usually when it does it's because the trial court clearly went overboard.  An essential ingredient in getting an appellate court to vacate an electronic discovery order is to show that the producing party has taken reasonable steps to produce likely sources of relevant documents already, and that the requesting party has failed to make a showing that more e-discovery (in the form ordered by the lower court) is likely to lead to relevant documents for which the need is significant enough to overcome the burdens being imposed.

The Texas Supreme Court recently issued such a decision under its state court rules that mirror the federal e-discovery rules.  In In re Weekley Homes, L.P., No. 08-0836 (Tex. Aug. 28, 2009), a mandamus proceeding, the court held that the trial court abused its discretion by ordering four of defendant’s employees to turn over their computer hard drives for a forensic examination.  In that case, the plaintiff served discovery requests that included requests for a variety of emails.  The defendant produced only thirty-one responsive emails, which the plaintiff did not believe was a sufficient or complete production.  The plaintiff filed a motion to compel additional searching, based on its assumption that more emails must have existed.  In response, the defendant asserted to the trial court that its employees’ inboxes were of limited size and were emptied by the users manually deleting messages once the inboxes became full.  The defendant reported that emails were saved only if manually saved by an individual employee to his personal hard drive.  All deleted emails, in contrast, were saved on computer backup tapes for only thirty days.  Based on that evidence, and in the absence of any evidence to the contrary or any indication that emails had been intentionally deleted to thwart discovery, the trial court initially denied plaintiff's motion.

The plaintiff then moved for "limited access” to certain of defendant's employees’ hard drives specifically to search for deleted emails.  The plaintiff proposed a search methodology under which a forensic expert would search for deleted emails on the hard drives, extract any located deleted emails, and provide all of those emails to defendant for its review and production to plaintiff.  not surprisingly, the defendant objected to that proposal, arguing that the proposed method was intrusive, would take the hard drives out of commission when needed by the employees, and that plaintiff had not even made a showing about either the need or the feasibility of obtaining the deleted data.  Nevertheless, the second time around the trial court granted plaintiff's motion.  The defendant initially sought mandamus relief from the court of appeals, which was denied, then seeking mandamus review in the state supreme court.

In analyzing the issue, the state court relied heavily on reported federal caselaw involving e-discovery, as there were few reported state court decisions under the analogous rules.  The court noted that, under the rules, a party ordinarily is only required to produce responsive electronic information that is “reasonably available to the responding party in its ordinary course of business.”  A party may be ordered to produce information that is not reasonably available “upon a showing … that the benefits of ordering the production outweigh the costs.” 

In this instance, the court concluded that the plaintiff had failed to meet its burden of showing that the benefits of the intrusive search it proposed outweighed the costs and burdens involved for the defendant.  As the court observed, the plaintiff's argument boiled down to a complaint that the defendant had only produced "a handful of emails," which the plaintiff assumed must have been incomplete.  The court agreed that the trial court "could have concluded that [plaintiff] made a showing that [defendant] did not search for relevant deleted emails that [plaintiff] requested.  But it does not follow that a search of the Employees’ hard drives would likely reveal deleted emails or, if it would, that they would be reasonably capable of recovery."

The court found that the plaintiff's "conclusory statements that the deleted emails it seeks 'must exist' and that deleted emails are in some cases recoverable is not enough to justify the highly intrusive method of discovery the trial court ordered, which afforded the forensic experts 'complete access to all data stored on [the Employees’] computers.'"

The missing step is a demonstration that the particularities of [defendant's] electronic information storage methodology will allow retrieval of emails that have been deleted or overwritten, and what that retrieval will entail.  A complicating factor is the some two-and-a-half years that passed between the time any responsive emails would have been created and the time [plaintiff] requested them.  Under these circumstances, it is impossible to determine whether the benefit of the forensic examination the trial court ordered outweighs the burden that such an invasive method of discovery imposed.

Therefore the court held that the trial court abused its discretion in ordering the intrusive discovery.

Recent Massachusetts Decision on Production of Metadata and Cost Shifting for E-Discovery

Recently a senior district judge in Massachusetts waded into the electronic discovery waters, providing guidance in a jurisdiction in which there are not yet many reported e-discovery decisions.

In Dahl v. Bain Capital Partners, LLC, Civil Action No. 07-12388-EFH (D. Mass. June 22, 2009) (click here), Judge Harrington noted that "the Local Rules of this court have yet to provide any guidance on
electronic discovery."  Id.at 1.  The court ruled on a discovery motion that addressed, among other things, issues concerning production of metadata and fee-shifting for discovery.  Regarding the latter, the court began by stating its presumption that "parties must satisfy their own costs in replying to discovery requests," although "electronic discovery costs can be shifted if the responding party identifies the source of the requested documents as not reasonably accessible because of an undue burden or cost on the responding party."  Id. at 2.  Because the producing party neglected "to identify which documents are inaccessible or the nature of that inaccessibility" or to "explain how they will sustain an undue burden or cost in producing documents," the court declined to shift the costs.  Id.  However, the court did rule that the responding parties "do not have to pay to change the format of their responsive documents" into a format desired by the requesting party, where the documents were otherwise "reasonably usable" in their ordinary format. Id. at 3.  Therefore, the court shifted the costs onto the requesting party "to pay: (1) for scanning and optical character recognition (“OCR”) for paper documents; and (2) for OCR for those electronic documents without text search capabilities."  Id.  At the same time, the court ruled that, if the producing parties OCR'd or otherwise changed the format of any of the produced documents for their own use in the case, they "should offer the [requesting party] access to these altered documents."  Id.

Regarding metadata, the court noted that the requesting parties "seek all of the metadata associated with emails and word documents produced," while the producing parties "refuse to produce all metadata and instead offer to provide 12 fields of metadata."  Id.at 4.  The court adopted the producing parties' limited approach, observing that "case law shows wariness about metadata’s value in litigation" and "producing diverse types of electronically stored information in the same format would be costly, burdensome, and ultimately fruitless." Id.  "Rather than a sweeping request for metadata, the [requestors] should tailor their requests to specific word documents, specific emails or specific sets of email, an arrangement that, according to their memorandum, suits" the producing parties.  Id.

All in all, the decision seems to reflect a balanced approach toward e-discovery in the context of this particular dispute.

If Your Discovery Compliance Stinks as Much as Your Case Does, the Ultimate Sanction Is Probably in Order

So we've talked before about the virtual impossibility of getting an electronic discovery-related order overturned on appeal.  Every once in a while, though, a case comes along in which it happens.

The California Court of Appeals recently issued a decision in just such a case.  See Doppes v. Bently Motors, Inc., -- Cal. Rptr. 3d --, 2009 WL 1578384 (Cal. App. 4 Dist. June 8, 2009) (click here).  In Doppes, a consumer brought various breach of warranty claims arising out of a foul odor emanating from his new Bentley automobile. 

The court made the "extraordinary, yet justified, determination that the trial court abused its discretion by failing to impose terminating sanctions against defendant for misuse of the discovery process. The record demonstrates defendant engaged in repeated and egregious violations of the discovery laws that not only impaired plaintiff's rights, but threatened the integrity of the judicial process."  Id. at *1.  The opinion chronicles a lengthy history of ongoing obstruction by the defendant in refusing to produce documents, intentionally moving emails to locations where they would not be captured in the electronic searches, failing to abide by court-ordered discovery, witnesses misrepresenting facts and general cover-up of reams of relevant and discoverable information, including entire file cabinets worth of documents. 

Prior to trial, the trial court appointed a discovery referee to oversee the contentious discovery battles.  The referee ultimately rejected plaintiff's request for entry of default against the defendant based on the electronic discovery abuses, but recommended that the court give an adverse inference instruction at trial.  The trial court did so.  In the midst of trial, the plaintiff was given leave to conduct additional depositions, which disclosed even more discovery violations.  Still the trial court declined to enter the ultimate sanction fo default against the defendant.  The plaintiff ended up prevailing on most of his claims at trial, but lost on a his fraud claims.

On appeal, the court noted that the discovery statutes in California "evince an incremental approach to discovery sanctions, starting with monetary sanctions and ending with the ultimate sanction of termination."  Id. at *17.  The court concluded that, with respect to the trial court's decision to give an adverse inference instruction based on the discovery violations prior to trial, the appellate court "cannot say ... the trial court abused its discretion initially in approving the discovery referee's recommendation ... of imposing the 'more narrow' sanction of giving the jury instruction."  Id. at *18.  Yet by the time of trial, the uncovering of additional serious violations mid-trial demonstrated that the defendant had "flagrantly engaged in such further discovery abuses so as to compel the trial court to impose the next level of sanctions-terminating sanctions."  Id.  Accordingly, "Terminating sanctions against [defendant]were imperative."  Id. Therefore, the court of appeals "reverse[d] the judgment on the fraud cause of action, with directions to the trial court to strike [defendant's] answer, enter a default and default judgment against [defendant] on that cause of action," and directed the trial court to conduct further proceedings to establish plaintiff's entitlement to damages and penalties.  Id. at *20.

The Format and Scope of E-Discovery Productions

A recent case addressed two important issues concerning the format for production of electronic documents and the showing necessary for a party to subject another party to intrusive tests by a third-party vendor to assess the completeness of a prior production.

Ford Motor Co. v. Edgewood Properties, Inc., -- F. Supp. 2d --, 2009 WL 1416223 (D.N.J. May 19, 2009) (click here), arose out of the demolition of an automotive assembly plant and the distribution of contaminated concrete from the plant.  Among other things, the defendant sought an order compelling plaintiff to reproduce its entire ESI production in native format, rather than in the format previously produced.  Defendant also sought to “confirm the adequacy of Ford's manual document collection process by using a third-party vendor to perform keyword searches on documents not in the existing repository of ESI, but instead, documents within the possession of certain ... custodians."  Id. at *1.

With respect to the format of the production, the defendant requested in its initial document request that plaintiff produce documents in native format; however, in responding, the plaintiff stated instead that it would produce electronically stored information in TIFF format with accompanying searchable text, but not in native format.  See id. at *5.  The parties apparently never reached an agreement as to how the documents would be produced, so plaintiff "unilaterally adopted its own objection and produced them in TIFF format."  Id.  Referencing the Sedona Principles, the court looked to the fact that the defendant failed to object to the TIFF format production until about eight months after the production - a delay that the court found "patently unreasonable," particularly in light of the fact that the production was long since complete.  Id. at *7.  The court therefore denied the defendant's request on the grounds that it would be "unduly burdensome to a party months after production to require that party to reconstitute their entire production to appease a late objection."  Id. at *8.

As to the second issue, regarding defendant's desire to require the plaintiff to permit a third-party vendor to perform random searches in various custodian electronic files, the court denied that request, as well.  The defendant argued that there was a suspicious absence of certain types of documents in the production made by the plaintiff, and posited that the document collection method used by the plaintiff must have been flawed.  To test that hypothesis, the defendant proposed having a vendor randomly search through electronic files for various employees of the plaintiff to see if responsive documents existed that were not produced.  Looking again to the Sedona Princples, the court noted that the producing party ordinarily "is the best position to determine the method by which they will collect documents," and that "the choice is clearly within the producing party's sound discretion."  Id.at *9.  Because the defendant's suspicions about missing documents was nothing more than "conclusory allegation premised on nefarious speculation ," id., the court denied the request.  "The notion that a document production is insufficient based on a belief that documents must exist simply is not enough to grant a motion to compel that would require Ford to go back to square one and begin its document collection efforts anew."  Id.  "To countenance such a holding would unreasonably put the shoe on the other foot and require a producing party to go to herculean and costly lengths (especially in a document-heavy case such as this) in the face of mere accusation to rebut a claim of withholding. This scenario is not contemplated by the Federal Rules."  Id.

So the lessons here are (1) if you're going to insist on a particular format for production of documents by the other side, you'd better actually insist upon it from the start, and (2) although the electronic discovery rules typically extend the volume of discovery in many cases, they are not a license to conduct fishing expeditions based on nothing more than a hunch that there's dirt that has not yet been uncovered.

Another Reminder to Cooperate in Conducting Email Searches

So I've written about the importance of collaboration among parties on the parameters of email searches as part of an electronic discovery protocol.  Absent collaboration, the requesting party demands too much and the producing party offers too little.  An even greater problem arises when the party being asked to produced simply offers no suggestions of its own.  That was one of the issues in the In re:  Fannie Mae decision I posted about last month.

Another court issued a reminder lesson recently in William A. Gross Constr. Assocs., Inc. v. American Manuf. Mut. Ins. Co., -- F.R.D. --, 2009 WL 724954 (S.D.N.Y. Mar. 19, 2009) (click here).  In that case, the court issued a "a wake-up call to the Bar in this District about the need for careful thought, quality control, testing, and cooperation with opposing counsel in designing search terms or 'keywords' to be used to produce emails or other electronically stored information ('ESI')."  Id. at *1. 

The case arose out of a dispute over alleged defects and delay in the construction of the Bronx County Hall of Justice.  At issue was the production of emails from the construction manager, a non-party to the lawsuit.  The owner of the project proposed search terms to be used by the construction manager in searching for responsive emails.  The "other parties requested the use of thousands of additional search terms," id., which essentially would have required "production of the entire [construction manager] email database," including all emails having nothing to do with this particular project.  Id.  Oddly, the construction manager declined to offer any suggestions for search terms of its own, instead only observing that the owner's "search terms were probably too narrow but the other parties' terms were overbroad."  Id.  As the court observed, the construction manager "was in the best position to explain to the parties and the Court what nomenclature its employees used in emails," but did not do so, "perhaps because, as a non-party, it wanted to have as little involvement in the case as possible."  Id. That unfortunately "left the Court in the uncomfortable position of having to craft a keyword search methodology for the parties, without adequate information from the parties."  Id.  The court therefore proposed using the owner's proposed search terms, which everyone seemed to agree were probably too narrow, plus additional search terms involving the names of the parties' personnel involved in the Bronx construction.  The court candidly recognized that "this result is less than perfect, and that there is a risk that as information later comes out at depositions of the [the construction manager's] employees, another search may have to be done."  Id. at *1 n.2.

The court expressed its frustration that this case was "just the latest example of lawyers designing keyword searches in the dark, by the seat of the pants, without adequate (indeed, here, apparently without any) discussion with those who wrote the emails."  Id.at *2.  The court declined to address at this stage "whether expert testimony is required" to expound upon how the ESI search should be conducted," but concluded that "what is required is something other than a lawyer's guesses, without client input, and without any quality control testing to see if the search terms produce reasonably all the responsive ESI and limited 'false positives.'”  Id. at *2 n.3.  "It is time that the Bar - even those lawyers who did not come of age in the computer era - understand this."  Id. at *3.

The All Important Litigation Hold

Another court issued a reminder recently on the importance of implementing a litigation hold at the onset of a case.

In Acorn  v. County of Nassau, 2009 WL 605859 (E.D.N.Y. Mar. 9, 2009) (click here), the plaintiff sued Nassau County.  During discovery, did not instruct any of the departments named in the complaint to preserve documents until roughly 15 months after it was served with the complaint.  The County further informed the plaintiff in response to other electronic discovery requests that it “does not possess either the resources or the technology to locate and access non-e-mail electronic documents created by such programs as Word, Excel or PowerPoint.”  Id. at *1.  The plaintiff then sought sanctions, arguing that electronic documents and evidence had not been preserved, so the plaintiff was entitled to sanctions including an adverse inference instruction.

In response, the County claimed that the county attorney immediately provided a copy of the complaint to the county executive and imposed a "verbal hold" on two county departments.  There was no evidence that the county attorney, or anyone else, or anyone else in her office, conducted any follow-up to ensure that the "verbal hold" instructions were being followed.  A formal litigation hold was not implemented until 15 months after the complaint was served on the County.

The County sought to explain its failure to implement a formal hold for 15 months by pointing to the motion to dismiss that it had filed, and its position that it was not until the motion to dismiss was denied and discovery commenced that it was able to identify the key players.

The court found the County's argument unpersuasive.  "The court cannot credit the County's argument that it was unable to glean from the complaint which particular agencies would likely possess discoverable information. If, as the County alleges, Ms. Miller was able to perform such an analysis leading to her “verbal hold” on several departments, then there is no apparent reason why the same analysis could not have been used to prevent the routine destruction of documents in those departments."  Id.at *3.  As the court concluded, "No matter how it is dressed up or characterized-as confusion or inconvenience or indifference or incompetence-the County's position is that it had no obligation to preserve discoverable materials until after the motionto dismiss was decided. This position is indefensible and is in direct contradiction with clearly established law."  Id. at *3.

The court further concluded that the County's failure to implement a legal hold at the outset of litigation amounted to gross negligence, sufficient to establish the culpability prong of the adverse inference test.

Luckily for the County, the court concluded that the plaintiff failed to establish sufficient relevance of the destroyed documents to meet the third step of the adverse inference inquiry, essentially concluding that there was no way to know whether there was anything of value in what was destroyed, and based on the limited value of the electronic documents that were produced, the likelihood of destruction of relevant materials was minimal.

"While it is certainly theoretically possible that some electronic data may have been deleted, there is no reason to believe that those materials would have provided any additional support for plaintiffs' claims. The court finds that the plaintiffs have not sufficiently demonstrated that any destroyed or lost mate-rials would have been favorable to them, and thus, to the extent plaintiffs' motion seeks an adverse inference instruction, it is denied."  Id. at *6.

Accordingly, the court awarded the plaintiff costs, but not an adverse inference instruction.

Regarding the County's prior statement that it lacked the resources to search for electronic documents beyond email, the County clarified its position to note that it actually did have the key players manually search their hard drives for responsive information.  The court disagreed with the plaintiff that those steps were insufficient.  "Plaintiffs have not provided any case law suggesting that a “manual” search of electronic files does not satisfy a party's discovery obligations."  Id. at *7.

A Party May Be Forced to OCR its Documents Before Producing Them

In most big document cases, having documents either produced or converted into electronic, searchable format is essential to facilitate review, organization and understanding of the content of what has been produced.  Sifting through rooms full of paper documents, and then trying to organize them in a coherent way for future access and use throughout the case, often is not feasible.  A recent decision demonstrates that, even where one party claims it has no interest in having the documents available in a searchable format, the court may order it anyhow after a cost-benefit analysis.

In Proctor & Gamble Co. v. S.C. Johnson & Son, Inc., 2009 WL 440543 (E.D. Tex. Feb. 19, 2009) (click here), the parties agreed that all documents produced in the case would be produced electronically in TIFF format, but disagreed whether the producing party also should have to perform the Optical Character Recognition (“OCR”) process on the documents prior to production.  Defendant did not want to do so, arguing that the total cost of conversion would run around $200,000, and that it "does not itself seek to use the OCR process, and any extra expense would be incurred on its behalf solely for Plaintiff's convenience."  Id. at *1 n.1.  The plaintiff disputed the $200,000 estimate, noting that the cost of converting a document into OCR format typically is in the range of 3 cents per page, and of course would only apply to documents maintained only in hard copy format.

The court did not find the defendant's cost estimate to be well supported.  Moreover, the court observed that the defendants did not dispute that providing OCR formatting for the documents would "make documents easier to examine and evaluate, thus reducing costs for attorney time."  Id. at *1.  After conducting the cost-shifting analysis under the seven-factor criteria set forth in Zubulake, the court concluded that the factors did not support shifting the OCR costs to the plaintiff.  The court referenced the parties' case management plan, noting that "the parties' respective litigation budgets in this case are estimated to be several million dollars apiece, so the parties have adequate resources."  Id. at *2.  As the court concluded:

OCR, while perhaps not absolutely necessary to litigation, is a tool that greatly decreases the time and effort counsel must invest in searching and examining documents. Presumably, each party would perform the OCR process in a cost-effective manner to minimize their costs. Requiring the parties to incur this cost, when the OCR process is likely to streamline the discovery process and reduce the chance that either side will employ tactics designed to hide relevant information in a mountain of difficult-to-search documents is neither unreasonable nor burdensome. Therefore, after review of the Zubulake factors, the court concludes that cost shifting is not appropriate in this case.

Id.  At bottom, perhaps the court simply found the defendant's argument that it had no intention of OCRing the documents for its own use unpersuasive, in light of the scope of the litigation and the resources available to both sides.

Third Party Electronic Discovery: Another Lesson on Email Searches

Following up on my post of February 24, 2009, discussing e-discovery email search terms, a recent D.C. Circuit decision highlights the extreme costs (both literally and figuratively) that can be imposed even on third parties who are served with electronic discovery requests that target email.

In In re:  Fannie Mae Securities Litig., 552 F.3d 814 (D.C. Cir. 2009), various plaintiffs sued Fannie Mae and several of its executives in actions consolidated into multidistrict litigation in the district court in the District of Columbia, alleging that Fannie Mae manipulated its earnings in order to inflate executive compensation.  A federal agency, the Office of Federal Housing Enterprise Oversight ("OFHEO") conducted its own investigation into the allegations and issued a preliminary investigative report.  In connection with the multidistrict litigation, several of the defendant Fannie Mae executives served subpoenas on OFHEO, seeking, among other things, copies of emails.

OFHEO initially moved to quash the subpoenas, and the district court entered an order denying the motion and directing the agency to comply with the requests. The agency later took the position that the court's order applied only to paper discovery, not electronic discovery.  The court again clarified that its order did, indeed, apply to electronic discovery, but at OFHEO's request it extended the deadline for the agency to comply with its production of both paper and electronic records. 

Months later, OFHEO informed the court that it had produced all documents requested by the executives, including all of the subject emails.  The executives were skeptical of that assertion, so obtained leave to conduct a Rule 30(b)(6) deposition of the agency.  During that deposition, the executives learned that the agency had not searched any of its off-site backup tapes for responsive emails.

The defendants then moved to hold the agency in contempt for failing to comply with the court's prior discovery order.  The court convened a contempt hearing, and following the first day of the hearing, the defendants and OFHEO entered into a stipulated order that stayed the proceedings on the contempt motions and required the agency to search its backup tapes for any responsive emails, produce any responsive emails and log any privileged materials within roughly 90 days.  Critically, the stipulated order contained the following language:

OFHEO will work with the Individual Defendants to provide the necessary information (without individual document review) to develop appropriate search terms. By October 19, 2007, the Individual Defendants will specify the search terms to be used.

Slip op. at 5.  The defendants than submitted more than 400 search terms which resulted in about 660,000 emails identified from the agency's backup tapes.  OFHEO objected, arguing that the stipulation did not give the defendants unfettered discretion to pick the search terms, but the district court agreed and ordered the agency to comply roughly 60 days before the deadline contained the stipulation expired.

In response, the agency hired "50 contract attorneys solely for that purpose" of complying with the order, and spent "over $6 million, more than 9 percent of the agency’s entire annual budget" on the search and review process.  Id.  Nevertheless, the agency still failed to complete the production of responsive emails and privilege logs by the stipulated deadline. 

The defendants then moved to hold the agency in contempt.  The district court was "cognizant of the large number of attorneys, contract attorneys, and OFHEO personnel working to comply with the subpoenas and the resulting costs of this compliance," but was concerned that the agency had "treated its Court-ordered deadlines as movable goal posts and has repeatedly miscalculated the efforts required for compliance and sought thereafter to move them."  Id.at 6.  In particular, the district court was disturbed by the fact that the agency had "only recently hired the necessary number of contract attorneys" to do what it had agreed to do several months before, which the district court found to be "too little too late."  Id. 

On appeal, OFHEO argued that the stipulation should not be read to permit the defendants unilaterally to select the search terms, and even if so, the district court should have considered the burdens and costs involved with the agency's compliance.  The court of appeals agreed that the agency "undertook extensive efforts to comply with the stipulated order."  Id.at 5.  Nevertheless, the court found the language in the stipulation to clearly allow the defendants to select the search terms.  Regarding the agency's arguments about cost-shifting and burden, the court found that OFHEO had "abandoned them by entering into the stipulated order" in the midst of the very hearing scheduled by the district court to consider the agency's objections to the scope of the subpoenas. 

And although the agency "undeniably made extensive efforts to produce the documents and privilege logs in accordance with the timetable set forth in the stipulated order," the court of appeals found those efforts insufficient to afford any "basis for concluding that the district court abused its discretion by finding it in contempt for failing to comply with the stipulated order’s deadlines."  Id.at 14-15.  That was so even though the D.C. Circuit noted that, "Were we deciding this matter in the first instance, we might not have held OFHEO in contempt."  Id. at 15.

So what are some of the important lessons here? 

First, there is no doubt that electronic discovery can be sought from, and ordered to be produced by, third parties who are not directly involved in the underlying litigation.

Second, before you agree to anything regarding e-discovery - whether representing a party or third party - it is essential to know what is likely to be involved in your client's end to comply with the requests and what the potential scope and costs may be.

Third, if you want to have some control over the scope of e-discovery, and email searches in particular, don't simply delegate that responsibility to the requesting party and hope that they will exercise restraint.  They probably will not, at least unless there is push-back. 

Fourth, although courts generally will be more mindful of costs and burdens for e-discovery productions when the target is a non-party, courts nevertheless will hold parties and non-parties alike to their own agreements.

And fifth, don't irritate the district judge.  Absent extreme results, an appellate court reviewing a discovery order from below is likely to be deferential.  Even in a case like this, where the costs were staggering and the court of appeals indicated it likely would have reached a different result, the lower court's findings and sanctions were upheld.

Parties Need to Work Together to Identify the Best E-Mail Search Terms

So I've written about the benefits of properly conferring about possible search terms in connection with email searches being performed as part of e-discovery.  See, e.g., this post.  Typically both sides want to dictate the terms that will be used when a company runs searches for emails, but the rules require a give-and-take process that should lead in most cases to neither side getting exactly what it wanted.  And as I've written about before, there is real benefit in a company offering its own suggestions on the types of searches that are likely to be most fruitful in the context of a given case, rather than resisting altogether.  The latter approach is more likely to lead to a court imposing requirements more onerous than the parties otherwise could have adopted.

The recent case of Lapin v. Goldman, Sachs & Co., 2009 WL 222788 (S.D.N.Y. Jan. 23, 2009), is instructive.  There, the plaintiff wanted the defendant to produce an entire collection fo about 300,000 emails; the defendant, in response, proposed to identify and produce from that universe only those emails that contained any of 75 company names included in plaintiff's document requests, plus 23 other search terms.  The plaintiff objected to using any search terms to narrow the universe of emails. 

In ruling on the discovery dispute, the court understandably looked at the defendant's offer and concluded that the "effort shows good faith."  Id. at *2.  But in spite of the plaintiff's intransigence, the court still permitted the plaintiff to have some say in identifying additional potential search terms.  The court direct plaintiff "to serve his own list of search terms, and then the attorneys and their computer experts must promptly hold a meeting with at least four hours of discussion about proposals for search terms.  Among other things, they should discuss (a) the estimated cost of the search and (b) the cost of a possible follow-up search with a supplemental list of search terms."  Id.

The court perhaps was more generous than it might have been; some courts when faced with a party that essentially has refused to participate cooperatively in the electronic discovery process will rubber stamp the other party's proposal, as there is nothing else on the table.  And the court also properly focused on the key issue in any e-discovery production:  the cost involved, balanced against the likely relevance of whatever might be produced.

Computers Can't Do Everything in Electronic Discovery

So I’ve talked before about the new Federal Rule of Evidence 502, which generally provides that disclosures of privileged materials don’t waive the privilege if they were inadvertent and reasonable steps were taken to prevent the disclosure.  (See my posts of September 9thand November 2nd.)  Well what kinds of steps are “reasonable?”

I know my inbox gets clogged these days with unsolicited offers from electronic discovery consultants” and vendors who claim to be experts in e-discovery issues, many of which are trying to peddle various kinds of software that can be used to screen and sort through emails in an attempt to identify privileged emails based on search terms and other parameters the user can define.  Some of the offers sound too good to be true.  And in some cases they might be.

The temptation to assume we can just push a button and let a computer do all the work can be great, particularly when faced with the prospect of a large, expensive and time consuming document review.  But the simplicity of the process might come back to haunt you.

In Rhoads Indus., Inc. v. Building Materials Corp. of America, Civil Action No. 07-4756 (E.D. Pa. Nov. 14, 2008) (click here) , the plaintiff retained an IT consultant in anticipation of litigation.  As part of that preparation, the plaintiff also purchased a piece of software Discovery Attender or “Sherpa” that identifies emails and screens them for privilege, based on data input by the user.  The plaintiff later filed its lawsuit, and after the defendants served discovery requests, the plaintiff ran searches using Sherpa that identified about 210,000 emails, about 2,000 of which Sherpa tagged as privileged.  The plaintiff ran an additional more refined search that narrowed the results from 210,000 down to about 78,000 emails.  The plaintiff then manually reviewed some of the email mailboxes from which those 78,000 emails came, and identified some additional privileged emails – beyond the 2,000 that had been flagged by Sherpa already – and pulled them and listed them on a log; the plaintiff did not, for some reason, list the 2,000 privileged emails identified by Sherpa on the log.  The plaintiff produced the balance of the 78,000 emails.

A month or so later, one of the defendants informed the plaintiff that it had identified several apparently privileged emails among the tens of thousands produced by the plaintiff.  The plaintiff only then reviewed the entire collection of the 78,000 emails, and identified 812 additional emails that it contended were privileged and which it included on a new privilege log.  The plaintiff argued that all 812 of those emails should be returned; the defendants not surprisingly argued that the privilege had been waived.

The district court began its analysis by referencing a 5-factor test for inadvertent disclosures in place in that district for over a decade.  Under that test, the court focused on the following:

(1) The reasonableness of the precautions taken to prevent inadvertent disclosure in view of the extent of the document production.

(2) The number of inadvertent disclosures.

(3) The extent of the disclosure.

(4) Any delay and measures taken to rectify the disclosure.

(5) Whether the overriding interests of justice would or would not be served by

relieving the party of its errors.

Id. at 4 (internal quotation omitted).  The court observed that “Keyword and name searches are frequently employed as an initial method to screen and sort documents, but they are not foolproof.  For instance, privileged communications frequently creep into e-mail 'chains,' but may appear only to some participants in the chain depending on a user’s application of the 'reply' and 'forward' commands.”  Id. at 5.

In weighing the 5 factors, the court concluded that the first 4 weighed in favor of finding a waiver of the privilege.  Id. at 19.  Nevertheless, the court found that the fifth factor – the interest of justice – favored the plaintiff because “[l]oss of the attorney-client privilege in a high-stakes, hard-fought litigation is a severe sanction and can lead to serious prejudice.”  Id. 

Regarding the plaintiff’s failure to log the 2,000 emails Sherpa had identified as privileged until six months after they were withheld, the court found the plaintiff’s conduct inexcusable, and in breach of the privilege log requirements of Fed. R. Civ. P. 26(b)(5).  See id. at 17.  Upon reconsideration, the court clarified its holding with respect to non-logged emails, holding that segments of those 2,000 emails that had separately been logged with the 812 emails could be redacted from email strings in which they also were found within the 2,000 emails.  However, the court rejected the plaintiff’s argument that it should be able to shield any emails within an entire string, simply by logging the most recent email in the string.  See Rhoads Indus., Inc. v. Building Materials Corp. of America, Civil Action No. 07-4756 (E.D. Pa. Nov. 26, 2008) (click here).

Simple isn’t always better, nor is it always even adequate.