Good, Better, Best: a Tale of Three Proportionality Cases – Part One

April 8, 2012

Three cases came out recently on proportionality, the key legal doctrine to discovery based on Federal Rules 26(b)(2)(C), 26(b)(2)(B)(iii), and 26(g)(1)(B). I-Med Pharma Inc. v. Biomatrix, 2011 WL 6140658 (D.N.J. Dec. 9, 2011) (GOOD); U.S. ex rel McBride v. Halliburton Co.,, 272 F.R.D. 235 (D.D.C. 2011) (BETTER); DCG Sys., Inc. v. Checkpoint Techs, LLC, 2011 WL 5244356 (N.D. Cal. Nov. 2, 2011) (BEST).  Since all three cases embody proportionality, they are all good. But some are better looking than others.

The quality of the application of the doctrine in these cases is directly tied to the parties timing. In the best case the issue was raised fast, even before discovery. It was raised in the 26(f) conference and 16(b) hearing. DCG Sys., Inc. v. Checkpoint Techs, LLC, 2011 WL 5244356 (N.D. Cal. Nov. 2, 2011). Counsel followed the Facebook good-hacker credo, as I explained in Impactful, Fast, Bold, Open, Values: Guidance of the “Hacker Way.”  They were fast, and their bold use of proportionality will likely have a big impact on the case. With the guidance of the wise judge supervising the parties’ discovery, U.S. Magistrate Judge Paule S. Grewal, the application of the proportionality doctrine should lead to open, yet cost-controlled discovery. This should allow both parties in the patent case to focus on  the merits. It should help avoid wasteful email reading endurance contests. Is that not a positive value we should all endorse? Answer: Yes! See Rule 1 and Judge Waxse’s article, Cooperation—What Is It and Why Do It?, as discussed in last week’s blog, Judge David Waxse on Cooperation and Lawyers Who Act Like Spoiled Children.

The Good Case: I-Med Pharma

Before I get into the California case that shows the right way to raise proportionality – at the beginning of the case – let’s look at the other two cases. The responding parties in both cases raised proportionality, but they came to the party a tad late. The worst of these good cases illustrates the saying better late than never. I-Med Pharma Inc. v. Biomatrix, 2011 WL 6140658 (D.N.J. Dec. 9, 2011). The party responding to discovery, the plaintiffs, finally raised the doctrine when their backs were against the wall fighting sanctions for failing to comply with a discovery order they had stipulated to.

Yup. You read that right. The plaintiff, I-Med Pharma Inc., was seeking relief on the basis of disproportionate burden from a discovery order they had agreed to. Kind of makes you wonder. But no explanation is provided in this opinion as to why plaintiff’s counsel in this case agreed to review and produce all non-privileged files that matched a ridiculously long list of keywords from opposing counsel. This was a contract dispute and the keyword list dreamed up by defense counsel, who apparently engaged in a rousing game of “Go Fish,” included such zingers as: contract*, loss, profit*, credit, refund, revenue, CL, HS*, return, claim, FDA, HA. I could go on, but you get the picture. (Side note: when I say keyword search sucks, as I did in Secrets of Search: Part One, this is the kind of search I am referring to: the blind guessing, Go Fish, linear kind with no quality controls. I am certainly not referring to the kind of iterative keyword search on steroids that we see in Kleen Products, which appears to be almost as good as predictive coding based hybrid multimodal methods. But I digress.)

In I-Med Pharma the attorneys not only used Go Fish keyword search, the kind that sucks, they agreed to have the search run by an outside forensic expert with no limits placed on target custodians. It was a search of the plaintiff’s entire corporate computer system. Not only that, there were no time limits placed on the search. To make matters worse, they not only agreed to search the active files with word matches, they agreed to search the slack space too. That is the so-called “unallocated space files” recovered by a  forensic exam of plaintiff’s computer system.

Yup. You read all that right again. No wonder the wise judge presented with this conundrum, Senior U.S. District Court Judge Dickinson R. Debevoise, began his opinion with these words:

This case highlights the dangers of carelessness and inattention in e-discovery.

Boy did Judge Debevoise get that right!

Plaintiff’s counsel finally woke up and discovered proportionality (here is where we get to the better late than never part), when the forensic expert searched the unallocated space of their client’s computer system and found 64,382,929 hits covering the equivalent of 95 Million pages of documents! Based on the complete failure to limit the search to custodians, or date, and the Go Fish type list of keywords, this result, in Judge Debevoise’s own words, should come as no surprise. Id. at pg. 5.

The opinion does not say how many pages of documents with hits were found in the allocated spaces of the system, the active files, but it was probably millions more. I-Med Pharma, Inc., apparently did not oppose the privilege review and production of these documents. No doubt they paid millions in vendor costs and attorney fees to comply with this portion of the stipulation.

Since plaintiff’s counsel by now probably had a pretty good idea of what a privilege review of another 95 Million pages of mostly gibberish from slack space might cost, and since at this point the client probably did not want to pay for more, plaintiff’s counsel said no. They asked defense counsel to give them a break on their prior agreement. Defense counsel said no to that, a deals a deal, and  perhaps sensing victory, refused to modify the prior stipulation. Then plaintiff moved for a protective order asking relief from the prior stipulation on discovery that had, as a matter of course, been converted to an order.

Here is where plaintiff raised the doctrine of proportionality and suggested that the costs and burdens to review 64,382,929 hits covering the equivalent of 95 Million pages of documents from slack space would exceed any possible benefit from that exercise. The Magistrate assigned to hear the dispute, Judge Shipp, quickly agreed, and the defendant, having little to lose (except perhaps credibility), appealed the decision to Judge Debevoise.

Judge Debevoise, of course, affirms his magistrate, which is why, after all, I call this comedy of errors a good decision. Judge Debevoise is a master of understatement and notes that:

A privilege review of 65 million documents is no small undertaking. Even if junior attorneys are engaged, heavily discounted rates are negotiated, and all parties work diligently and efficiently, even a cursory review of that many documents will consume large amounts of attorney time and cost millions of dollars.

Id. at pg. 10.

Judge Debevoise granted a hearing on plaintiff’s appeal, which I imagine would have been interesting to observe, since this judge really seems to have a good understanding of e-discovery. At the hearing defense counsel argued that plaintiff’s obligation to review 95 Million pages was not really that burdensome. As footnote six of the opinion explains, Judge Debevoise responded by asking defense counsel how they would do a privilege review of that many documents. That’s where it gets fun, as defendants’ counsel said they would simply run a search for the word privilege and only review the documents with that word. Uh, huh. Sure. Judge Debevoise observed in footnote six:

In spite of the answer given, it is difficult to believe that lawyers from xxxxxxx and xxxxxxxx regularly disclose large quantities of information from their client’s files without examining it.

Id. at pg. 10-11, FN 6. (self censored, hey it could happen to any otherwise good firm.)

So, Judge Debevoise let’s plaintiff’s counsel off the hook and relieves them of their prior e-discovery agreement. But he has some choice words for them too (and again I will not name plaintiff’s counsel), and provides good advice for all on a better way to do keyword search, a way that goes far beyond the simple guessing game the attorneys in this case had apparently been playing:

While the precise number of hits produced was not known in advance and Plaintiff argues that it could not have predicted the volume of material that the search would uncover, it should have exercised more diligence before stipulating to such broad search terms, particularly given the scope of the search. In evaluating whether a set of search terms are reasonable, a party should consider a variety of factors, including: (1) the scope of documents searched and whether the search is restricted to specific computers, file systems, or document custodians; (2) any date restrictions imposed on the search; (3) whether the search terms contain proper names, uncommon abbreviations, or other terms unlikely to occur in irrelevant documents; (4) whether operators such as “and”, “not”, or “near” are used to restrict the universe of possible results; (5) whether the number of results obtained could be practically reviewed given the economics of the case and the amount of money at issue.

While Plaintiff should have known better than to agree to the search terms used here, the interests of justice and basic fairness are little served by forcing Plaintiff to undertake an enormously expensive privilege review of material that is unlikely to contain non-duplicative evidence.

Id. at pgs. 11-12.

I-Med Pharma Inc. v. Biomatrix, is a good case, not only for proportionality, but also for search. It is very telling to note that even though the case embodies the doctrine of proportionality, the keyword itself is never used – even Rule 26(b)(2)(C) is never referred to. This once again demonstrates the limits of keyword search. I would not have found this case but for the genius of a non-linear, protodigital, artificial intelligence agent I know, a mobile Siri of e-discovery, Kenneth J. Withers. Thank you Ken. (Actually, truth be known, he found all three of these cases.)

The Better Case: U.S. ex rel McBride

The next case demonstrating proportionality is U.S. ex rel McBride v. Halliburton Co. , 272 F.R.D. 235 (D.D.C. 2011). It is a better case than I-Med-Pharma because proportionality was applied in a slightly more timely fashion. True, discovery had already closed, but at least protection was sought before stipulation to an order. It is also a good case because it was written by one of my favorite judges, a master wordsmith and  e-discovery guru, John M. Facciola. He’s got a proportional look too.

Judge Facciola is a strong advocate for proportionality. By the way, that does not mean the judge has prejudged anything, or favors responding parties over requesting parties, or anything like that. It is the same thing as saying a judge favors reasonability, or motherhood, or apple pie, or, dare I say it, modern efficient search and review methods.

Judge Facciola was one of my keynote speakers in a CLE event that I co-chaired in 2010 with Maura Grossman. The sole topic of that day-long event was proportionality. It was one of the best CLEs I have ever done for that reason. There was a thorough examination of one issue from some of the top minds in e-discovery, including especially the top judges who specialize in the subject. Judge Facciola started us off by speaking of proportionality in the law and art and music where the principle is called the golden mean or golden ratio. He even played a few lines of Bach that demonstrated the golden ratio in music. I followed his presentation with my own where I spoke and showed slides on proportionality in the law and art, paintings and design. I showed famous paintings, buildings (Parthenon) and other golden ratio based designs (iPods) and examined how this ratio would apply to e-discovery costs. It seemed like part two of what Judge Facciola had started.

Judge Facciola and I were both blown away by the eerie coincidence of our art based approaches to explaining proportionality in the law. I am sure everyone there thought this was a well orchestrated presentation. But the truth is, I had no idea he was going to use the same art based approach to proportionality. Neither of us had ever talked to the other about what we were going to say. (You don’t need to have prep talks with a pro like him.) The truth is, there was no need for us to talk and prepare in order to have harmonious presentations. There is a high degree of consensus and similar thinking among judges and lawyers who specialize in this area. We have all read each others writings and heard each other speak a number of times. We think about this stuff, a lot. We tend to reach the same conclusions. There is a general consensus on most issues. What’s wrong with that? But I digress.

Back to U.S. ex rel McBride v. Halliburton Co.. Although Judge Facciola is an expert and strong proponent of proportionality, my keyword search of the opinion shows that he never once uses the word in this opinion. He cites Rule 26(b)(2)(C) several times to be sure, but never says proportionality, proving once again the limits of that old keyword technology. For more on new search alternatives to 1940s era software methods, new technologies that favor the truth, not one side of the Bar or another, see (in reverse chronological order) my following recent articles (yes, I’m obsessed with this):

So Judge Facciola implements proportionality without ever saying the word, and grants the defendant’s motion for protective order. Here Halliburton in this qui tam action had already reviewed and produced relevant emails of 230 custodians. Yup, you read that right, 230! Discovery closed and what does the plaintiff want? They want Halliburton to search and produce still more email from an additional 35 custodians. These additional custodians were now targeted by the plaintiff, McBride, because they were shown as CCs on emails transmitting relevant documents that were already produced as part of the 230. No other reason provided. Let’s listen to how the Sage from D.C., who loves the mathematical proportionality of Johann Sebastian Bach (not to mention Bruce Springsteen), deals with such a dissonant request:

In excruciating, but highly educational and useful, detail, she (one of Halliburton’s e-discovery compliance managers) explains how the process and collection of data from a current or former employee takes place, from what may be many different sources … Once the data is found, it must be copied in a forensically appropriate manner to preserve its metadata and prevent its alteration. … The data must then be processed to be rendered searchable by the review tool being used, a process that can overwhelm the computer’s capacity and require that the data be processed by batch, as opposed to all at once. …

All discovery, even if otherwise permitted by the Federal Rules of Civil Procedure because it is likely to yield relevant evidence, is subject to the court’s obligation to balance its utility against its cost. Fed.R.Civ.P. 26(b)(2)(C). More specifically, the court is obliged to consider whether (1) the discovery sought is unreasonably cumulative or duplicative, or obtainable from a cheaper and more convenient source; (2) the party seeking the discovery has had ample opportunity to obtain the sought information by earlier discovery; or (3) the burden of the discovery *241 outweighs its utility. Id. The latter requires the court to consider (1) the needs of the case; (2) the amount in controversy; (3) the parties’ resources; (4) the importance of the issues at stake in the action; and (5) the importance of the discovery in resolving the issues. Fed.R.Civ.P. 26(b)(2)(C)(iii).

While the present record does not permit a precise conclusion, I can presume, given the numbers of hours for which the defendants billed and the period of time at issue, that the amount in controversy is great and that the defendants’ resources are greater than the relator’s. Claims of fraud in providing services to military personnel raise important, vital issues of governmental supervision and public trust. Thus, these factors might weigh in favor of the discovery sought.

On the other hand, the defendants protest, and relator does not deny, that they have already spent a king’s ransom on discovery in this case–$650,000–without the addition of attorneys’ fees.  They have produced more than two million paper documents, thousands of spreadsheets, and over a half a million e-mails.

Id. at pgs. 240-241 (record citations omitted).

Judge Facciola goes on, and note once again that he never says proportion* or any variation thereof:

Given the discovery that relator has had, what defendants have already spent, and the detailed showing made of how much more time and money will likely have to spent to search an additional thirty-five custodians, surely relator has to make a showing that the e-mails not produced are crucial to her proof. She has not made such a showing, and they are not. First, she has the LOGREP reports, and no one is pretending that defendants are (or could) be asserting that, without the transmitting e-mails, she cannot establish the submission of a false claim. To the contrary, the defendants conceded that point. Thus, the transmitting e-mails seem to be hopelessly insignificant.

In this context, it is telling that relator does not show from the e-mails she has received that there is good reason to believe that the ones she claims are missing are highly probative of some fact. Indeed, there is no showing whatsoever from what has been produced that those e-mails not produced will make the existence of some crucial fact more likely than not. It is, after all, unlikely that a transmitting e-mail will do any more than transmit attached information and, by copy, alert others of that transmittal.

Without any showing of the significance of the non-produced e-mails, let alone the likelihood of finding the “smoking gun,” the search relator demands cannot possibly be justified when one balances its cost against its utility. The motion will be denied.

Id. at 241.

Well put Judge Facciolla. This is music to the ears of any e-discovery devotee. I am reminded of my favorite, the Goldberg Variations. There are additional melodic words in this opinion about a party not losing their right to proportionate discovery because of preservation failures, but I suggest you read the opinion yourself for that last variation. Id. at 241-242.

The Best Case: DCG Systems

End of Part One. Stay tuned for the exciting conclusion, where you will hear the happy ending to this tale of three cases with the story of the best case, DCG Systems, more case-law on proportionality, and an opinion-riddled conclusion that will leave your jaws gaping.

As a sneak preview on the ending I will say this much. Although I agree with Siri’s explanation of DCG Systems, err, I mean, Ken Withers’ explanation: DCG Systems is a case that “proactively cuts the Gordian Knot by adopting Judge Rader’s Model Patent Discovery Order,” (yes, he actually talks like that, and with links too), I nevertheless have some criticisms of this case and the Model Order on which it is based. Its reliance on five keywords is flawed. Still, given the cost of most vendor’s predictive coding software these days, and the weak understanding most lawyers have of legal search, this reliance on outdated technology and search methods is to be expected.

Now for something completely different, I leave you with a few lyrics from Judge Facciola’s favorite musician-poet, Bruce Springsteen:

That you know flag flying over the courthouse
Means certain things are set in stone
Who we are, what we’ll do and what we won’t.

“Long Walk Home”

God have mercy on the man who doubts what he’s sure of.

“Brilliant Disguise”

But it’s a sad man my friend who’s livin’ in his own skin
And can’t stand the company.
Every fool’s got a reason to feelin’ sorry for himself
And turn his heart to stone.
Tonight this fool’s halfway to heaven and just a mile outta hell
And I feel like I’m comin’ home.

“Better Days”

So let’s take the good times as they go and I’ll meet you further on up the road…

“Further On (Up the Road)”

5 Comments | Lawyers Duties, New Rules, Review, Search, Technology | Permalink
Posted by Ralph Losey

Predictive Coding Based Legal Methods for Search and Review

March 25, 2012

Predictive coding type algorithms are designed to leverage the expertise of human input, preferably attorneys who are subject matter experts of the case at hand. A classification of one document by an attorney results in a recommended classification of hundreds, if not thousands of other documents that the computer identifies as similar. The computer examines the entire data set, the corpus, and predicts the probability of each document therein fitting within the same classification. The expert then tests and correct the predictions in an iterative process. The cost savings in this approach are obvious, particularly considering the high expense of attorney review time. Reviewers can break through the current linear review speed barrier of approximately 100 files per hour, to 1,000, or even 10,000 files per hour. These supersonic review speeds are what it takes to handle e-discovery today in an effective and economical manner.

Predictive coding software tracks and learns from human expert decisions in a feedback loop. This increases accuracy, allowing greater reliance and faster culling of irrelevant (or unresponsive) ESI. Only the ESI predicted to be relevant then needs to be reviewed by attorneys for final quality control verification before production. The extent and degree of the final human review for quality control  varies according to two considerations: proportionality constraints related to the value and importance of the case; and, the parties risk tolerance for disclosure of confidential or privileged information.

This kind of software features that I here refer to as predictive coding algorithms use what is known in information science as classification systems based on probability theories. The Sedona Conference Best Practices Commentary on the Use of Search & Information Retrieval Methods in E-Discovery, 8 Sedona Conf. J. 189, 218-221. (2007):

Probability theories are used to make decisions regarding relevant documents. The most prominent of these are so-called “Bayesian” systems or methods, based on Bayes’ Theorem, and the related  “dimension reduction systems.”

With predictive coding human review of all files is eliminated. The operative word here is all. As mentioned, review of some files by attorneys is still required for a variety of reasons. Primary among them are training of the artificial intelligence agents, and quality control verification before information is transferred to the requesting party. But all files will not be reviewed by human eyes. In fact, in some cases only a small minority may be subjected to human review. The percent of files not reviewed by experts, but instead only reviewed by artificial intelligence algorithms, will depend on the costs and time factors in the case — the proportionality analysis under Rule 26(b)(2)(B) — and the producing party’s risk tolerance.

Experienced, Skilled Attorneys Are Key to Successful Use of Predictive Coding Software

The input of human experts to train the artificial intelligence algorithms is essential and has a direct impact upon the quality of performance of the predictive coding. If, for instance, the expert is inconsistent in making initial calls on the relevance of documents, then the computer extrapolations based on these calls will also be inconsistent. It is the perennial computer situation of garbage in, garbage out. The quality of the predictions made by the computer is, in large part, based on the quality of the training, the input, provided by the human experts. Quality is, of course, also impacted by other factors, including the nature of the ESI under review, the number of iterative human expert input cycles, and the quality of the software itself. But any deficiencies in these quality factors can be detected and corrected by statistical random sampling of the final predictions. This is ultimately why the black box itself, i.w. exactly how the software works, is not really that important. The final quality controls of random sampling protect against any errors, including software errors.

Although random sampling protects against errors, and greater reliance is placed on these new technologies, the skills, or lack of skills, of attorneys managing a discovery project are still critical. Unskilled attorneys may, for instance, be inconsistent in coding of relevance over multiple documents. They may also make poor choices as to what documents should be considered as representative of highly relevant documents. Even errors in merely relevant coding can have cascading effects.

Predictive coding software models future choices after the documents submitted to it as representative. If the documents have not been properly evaluated or selected by attorneys to begin with, then the results of the automated analysis will also be flawed. If you show the computer what a smoking gun document actually looks like, then the computer will have a good chance to find it. But if you show it a poor imitation, not a real smoking gun document, then you should not be surprised when the computer only finds duds.

When I speak of skilled attorneys, I am not simply talking about attorneys who know how to operate a particular vendor’s software. I am talking about attorneys who really understand the particular issues of a case and the relative probative value of various documents. I am also talking about attorneys who understand the new iterative processes and legal methods to properly use the software.  The diagram below provides a high-level overview of these new processes that I recently prepared. I will go into greater detail about these new iterative predictive coding methods for ESI productions in a future blog.

Do not simply use in an unthinking manner the standard work flow methods that vendors set up as defaults for software. Also, if a vendor offers advice on legal methods and best practices (and good vendors know better than to do that), take it all with a big grain of salt. Vendors are not lawyers. E-discovery vendors are not allowed to provide legal services or legal advice. They are prohibited from doing so, even if the expert has a law degree and is licensed to practice law. It is even worse when a tech expert, one who has never even gone to law school, provides opinions on legal methods and what is reasonable and what is not. Vendors and techs are permitted to be advisers on technology and software functions, not the law. Legal methods must be determined by lawyers, not techs. Legal advice may only be provided by practicing attorneys in firms, or acting as solo practitioners. Still, unless you are already an expert in a particular kind of software or technology, it is a very good idea to consult with a software expert or bona fide information scientist in order to formulate your legal opinions. The multi-disciplinary team approach is required in e-discovery work.

Bottom line: you must never let the tail (the software) wag the dog (legal search). Attorneys must at all times remain in control and not abdicate their responsibility as legal advisers to non-lawyer tech experts, much less to software. In other words, even the best software is only as good as the lawyers who use it.

Proving a Negative

The inability of attorneys to uncover documents requested does not necessarily mean that the attorneys are unskilled or the software is defective (although it might). The template for the smoking gun document might be accurate, but nothing like it is found for the simple reason that none exist. This is a key point that should never be forgotten. It is also one reason that skill alone is not enough to know whether you have proven a negative, or have made a mistake. Art and experience are important in this area of the law, just like in any other area of the law. If there is no needle in the haystack to begin with, then no amount of skills, quality controls, or repetition in search will find one.

An experienced attorney gets a feel for the data under review and the people involved. After a while of studying a large data set like a multi-custodian collection of email they know when proceeding further is likely futile. The data examined allows for an accurate assessment of the probabilities of the data not yet examined. It looks like this may be a kind of intuitive application of Bayes’ theorem of conditional probabilities.

It is not uncommon for a requesting party to imagine or hope for the existence of emails or other documents that, like a Unicorn, simply do not exist. The requester is on a mere fishing expedition, and often a delusional one at that. If only a few relevant documents are uncovered by a search, that does not necessarily mean garbage in, garbage out. It could be treasure in, treasure out, as far as the responding party in concerned. The search was successful in proving that what was alleged to have happened, in fact did not. It is proof of a negative by failing to discover a positive. This does not require that every stone be turned as some still think. Probability of data is quite sufficient in the law in these circumstances for a number of reasons, including cost proportionality restraints and burden of proof requirements. 

The old adage “where there’s smoke there’s fire” is usually true, but the converse is also true. Moreover, sometimes there is just smoke and mirrors without any fire at all. It is just a conjured wrong that doesn’t really exist. Knowing the difference is where experience and art in law come into play. There is no artificial intelligence for that, although Bayesian probability analysis comes close. This is something that most software designers, vendors, and inexperienced attorneys do not fully appreciate.

Positive Learning Feedback Loop

Even the most highly skilled attorneys will make mistakes from time to time. All experienced attorneys know that. That is why the iterative processes of predictive coding are so valuable. When mistakes in input are made by the subject matter expert attorneys who create the first seed set in the predictive coding process, it is not a one-time, all or nothing procedure.

In predictive coding the linear processes are replaced by cyclic ones. A review by attorneys of the results returned by their initial input creates a positive learning feedback loop for both the software and attorney reviewers. They will quickly detect any errors in their analysis and make corrections accordingly. It is a way of repeatedly checking their work. If they get a return of duds, they know to improve the picture of the smoking guns documents they are looking for, or to look for something else. Proper use of predictive coding tools will in effect make the attorney reviewers smarter and more effective. This in turn allows better input from the attorneys and improves the computer’s efficiency in the next round. The iterative nature of the process allows the creation of powerful positive feedback loops.

The more advanced trans-keyword search methods include the following as described by one of the leading jurists in this area, Judge Paul Grimm:

In addition to keyword searches, other search and information retrieval methodologies include: probabilistic search models, including “Bayesian classifiers” (which searches by creating a formula based on values assigned to particular words based on their interrelationships, proximity, and frequency to establish a relevancy ranking that is applied to each document searched); “Fuzzy Search Models” (which attempt to refine a search beyond specific words, recognizing that words can have multiple forms. By identifying the “core” for a word the fuzzy search can retrieve documents containing all forms of the target word); “Clustering” searches (searches of documents by grouping them by similarity of content, for example, the presence of a series of same or similar words that are found in multiple documents); and “Concept and Categorization Tools” (search systems that rely on a thesaurus to capture documents which use alternative ways to express the same thought).

Victor Stanley, Inc. v. Creative Pipe, Inc., supra at FN 9 of the opinion. Also see videos of my short-hand explanations of predictive coding at LegalTech interviews: My Impromptu Video Interview at NY LegalTech on Predictive Coding and Some Hopeful Thoughts for the Future; and, Interview at Legaltech 2012 on predictive coding and transparency.

The current linear review and culling process derived from paper discovery, where one step follows after the last one is completed, should be replaced by a non-sequential, four-dimensional approach. Here all steps are pursued at once, to some degree, and are repeated over time (the fourth dimension) within predetermined budgetary constraints. This is a use of iterative cycles, where multiple decisions are made at the same time with computer assistance, but later verified and quality controlled by sampling. This is all done within case specific proportionality limits under Rule 26(b)(2)(C), Federal Rules of Civil Procedure and related case law. This  type of iterative project management may be new to the law, but is well established in technology development where diagrams like this are common.

Under the new procedures, the old style ad hoc quality control methods are replaced by judgmental and statistical sampling and other metric based quality control systems. The old school muddling management approach is replaced by proven project management techniques. See The Sedona Conference® Commentary on Achieving Quality in the E-Discovery Process; (identifies seven elements critical for successful project management: 1. Leadership.
 2. Tailoring. 
3. Expertise. 
4. Adaptability. 
5. Measurement.
 6. Documentation. 
7. Transparency.) The new procedures also require and depend on cooperation and disclosure. See eg. Animation Showing How Not To Cooperate In An E-Discovery Conference.

Six Search Ideas

In Secrets of Search Parts One, Two and Three, I outlined the five key characteristics of effective search today, using the rubric of secrets. In Part Three I summarized my ideas on search and review using the symbol of the Pythagoreans, the five-sided polygon, or pentagon:

With my blog, Bottom Line Driven Proportional Review  I added the sixth idea, where the process gets real and takes money into consideration. In that blog I shared my method to use estimation, projections, budget, cooperation, transparency, and the legal doctrine of proportionality to control the costs of search and review. This final piece competed the outline of a new gold standard of search and review:

Predictive coding is just one tool, one part of the six search ideas. It is part of the “hybrid multimodal,” one of many search techniques, including keyword, that make up multimodal, meaning multiple types of search techniques. Predictive coding is currently the best search tool to be sure, but it is just another high-priced hammer without a well-trained carpenter to use it, one who knows the new multidimensional, iterative, largely automated, cooperative methods. (By the way, have I mentioned yet that most software is priced too high? Come on vendors, go for quantity of sales by lowering the prices.)

Conclusion

The number of documents we have to review seems to double every two to three years, so this new six-fold legal method is imperative. New and better software, especially predictive coding type, is also important. The ranking of relevancy and other categories built into the latest algorithms is, under Bottom Line Driven Proportional Review, an especially helpful new capability. You rank the documents that the computer predicts will be the most relevant to your case and still fit within your budget. You plan an open proportional review, not an over-burdensome one.

Predictive coding is a powerful new tool, but technology alone is not enough. We must also have new legal methods. Technology and law have to work together, grounded in science, to create a new gold standard. The law, driven as it is to stop the run-away costs of e-discovery, is now ready to adopt these new ideas and methods. It is ready to change from a linear, confrontative, one-dimensional, mostly-manual, Bates stamp approach to discovery, to a multidimensional, cooperative, iterative, largely-automated, hash value approach.

Change is possible. Just keep working on it, a little bit every day, and then suddenly, a breakthrough. Once you break the ice, and get things started, the pace of change can quicken. New technologies like predictive coding and new legal methods like bottom line driven proportional review can be quickly accepted.

But we need your help to establish and explain these new ideas and methods to the Bench and Bar. We need more lawyers and other experts to implement the new approach. Remember the lessons of Zuckerberg’s success: Impactful, Fast, Bold, Open, Values: Guidance of the “Hacker Way”. Don’t wait any longer. Affirmance is on its way. Be bold, try the new methods, the new software, and don’t be afraid to be open about what you are doing. We need the impact of these values to quickly stem the tide of ESI overload and excessive costs.

As Margaret Mead said:

Never doubt that a small group

of thoughtful, committed, citizens

can change the world.

Indeed,

it is the only thing that ever has.

6 Comments | Evidence, Metadata, New Rules, Review, Search, Technology | Permalink
Posted by Ralph Losey

Perspective on Legal Search and Document Review

March 11, 2012

For millennium writings were on paper. For centuries the legal profession depended upon writings, referred to in the law as documents, as the key evidence to resolve disputes in a fair and just manner. Losey, Mathematical Formula for Justice Proves the Importance of ESI in Civil Litigation, Chapter 4 of Electronic Discovery (West 2010). Paper documents were well-known and mastered by every lawyer and judge who swore an oath to uphold the law.

This all changed in a historical blink of the eye. In just one generation documents have dematerialized and transformed into a dizzying array of digital media.

New Age of Technology

Many see this as a much more profound cultural revolution than that precipitated by Guttenberg, which took centuries to play out, not decades. George L. Paul and Jason R. Baron, Information Inflation: Can the Legal System Adapt? 13 RICH. J.L. & TECH. 10 (2007). Paul and Baron’s well known article explains how writing co-evolved with civilization over the past 50 centuries or longer, with a slow but steady increase in information as our writing technologies slowly improved. They point out that this all changed about 25 years ago when Mankind invented a totally different form of electronic writing, free from physical confines, that triggered a Big Bang like explosion of a new Universe of virtually unlimited information. Paul and Baron predicted the legal profession will necessarily have to significantly change and adapt new strategies of practice to cope with this information revolution.

Documents originally created on paper still exist in our society, but are rare. Zubulake v. UBS Warburg LLC, 217 F.R.D. 309, (S.D.N.Y. 2003) (FN 5 cites Wendy R. Liebowitz, Digital Discovery Starts to Work, Nat’l L.J., Nov. 4, 2002, at 4 reporting that in 1999, ninety-three percent of all information generated was in digital form).

Most of the paper documents we see are merely printouts of one dimension (the text) of the original electronic information. The law recognizes this transformation and the Federal Rules of Civil Procedure were amended in 2006 to include electronically stored information (ESI) as information that can be discovered and used as evidence in lawsuits. Rule 34(a)(1) Federal Rules of Civil Procedure (2006). ESI is not specifically defined in the rules. The Rules Committee Commentary explained why:

The wide variety of computer systems currently in use, and the rapidity of technological change, counsel against a limiting or precise definition of electronically stored information. Rule 34(a)(1) is expansive and includes any type of information that is stored electronically.

Even without specific amendments to rules, all courts today, state and federal, treat ESI as potentially admissible evidence subject to discovery. The first Sedona Principle is now commonplace:

Electronically stored information is potentially discoverable under Fed. R. Civ. P. 34 or its state equivalents. Organizations must properly preserve electronically stored information that can reasonably be anticipated to be relevant to litigation.

Stressed Out Lawyers

The legal profession has been severely stressed by the rapid, ever-accelerating advances in technology. The changes in writing and resulting information explosion have been the key stressors. ESI is not only changing and evolving new into forms every year, but is now multiplying at an exponential rate that is almost beyond comprehension. See Eg., Rowe Entm’t, Inc. v. William Morris Agency, Inc., 205 F.R.D. 421, 429 (S.D.N.Y. 2002) (explaining that electronic data is so voluminous because, unlike paper documents, “the costs of storage are virtually nil. Information is retained not because it is expected to be used, but because there is no compelling reason to discard it”), aff’d, 2002 WL 975713 (S.D.N.Y. May 9, 2002); Data, Data Everywhere (The Economist, March 2010); Baron and Losey E-Discovery: Did You Know video.

Most lawyers are unfamiliar with ESI and the complex systems that store it. They prefer the familiar paper and alphabetical filing cabinets. They are paper lawyers living in a digital world.

The astronomical volumes and complexity of ESI has made the traditional process of legal discovery very expensive and burdensome. Many are calling this a crisis in our legal system that threatens our system of justice. Final Report on the Joint Project of the American College of Trial Lawyers and the Institute for the Advancement of the American Legal System (2009) at pg. 15:

Although electronic discovery is becoming extraordinarily important in civil litigation, it is proving to be enormously expensive and burdensome.

The old methods to review digital writings that lawyers used for paper are too expensive. Few can afford the time and effort required to locate, review, and produce all relevant evidence using those old methods. The costs and burdens that way often exceeds the value of entire case. There is a real danger that the resolution of disputes in a court of law based on both testimony and writings will be a luxury available only to the wealthiest parties. See eg. Justice Stephen Breyer’s Preface to Sedona Conference Journal, Vol. 10 Supplement, Fall 2009:

[The articles in this Supplement] suggest that if participants in the legal system act cooperatively in the fact-finding process, more cases will be able to be resolved on their merits more efficiently, and this will help ensure that the courts are not open only to the wealthy. I believe this to be a laudable goal, and hope that readers of this Journal will consider the articles carefully in connection with their efforts to try cases.

Even though many scholars, jurists, and practitioners recognize the problems created by the inability of lawyers to keep pace with technology, most law schools still only train students in paper evidence and discovery. Students graduate unprepared to handle the ESI where the truth of past events is now stored. William Hamilton, The E-Discovery Crisis: An Immediate Challenge to our Nation’s Law Schools, Chapter 39 of Electronic Discovery (West 2010); Shannon Capone Kirk, Kristin G. Ali, Teach Your Children Well: A Case for Teaching E-Discovery in Law Schools, Chapter 38 of Electronic Discovery (West 2010); Shira Scheindlin, Ralph Losey, E-Discovery and Education, Chapter 33 of Electronic Discovery (West 2010); Ralph Losey, Plato’s Cave: Why most lawyers love paper and hate e-discovery and what this means to the future of legal education, Chapter 32 of Electronic Discovery (West 2010).

The law remains as dependent as ever upon documents to prove the truth, but the vast majority of lawyers are untrained and unprepared to handle the electronic documents upon which the world is now built. E-Discovery Competence is a Fundamental Ethical Challenge Now Faced by the Legal Profession, Chapter 34 of my book Electronic Discovery (West 2010). In fact, most lawyers, even those who specialize in litigation, dislike e-discovery and try their best to avoid it. See my blogs: Tell Me Why?, and Spilling the Beans on a Dirty Little Secret of Most Trial Lawyers. Lawyers are trained and prepared instead to handle paper documents following systems developed in the 20th Century.

Old 20th Century Methods

These old paper-based legal search and review methods are one dimensional and linear in nature. They typically follow a sequential Bates stamp organizational model created in the 1890s. The simple paper evidence discovery processes worked pretty well for decades before computers. It should be noted, however, that even before technology moved away from  paper typing machines to computers in the 1980s, the discovery processes were already severely taxed by the growing volumes of paper documents generated from the 1960s forward. The increase in paper volume was caused by another technological innovation, the photocopy machine and by ever more complex transactions. Still, the legal profession coped somehow for the rest of the Twentieth Century. Lawyers added more numbers to the Bate stamps and used larger teams of lawyers and paralegals to manage the additional papers. They were still on familiar ground.

The linear systems developed in the 19th and 20th Centuries for the discovery and production of documents continue to be used today by most attorneys for both ESI and paper discovery. See Eg. the D’Onofrio saga of four opinions by Judge Facciola which describes the processes used in this case and his many orders resolving disputes, including an order requiring production of a sample of the 9,413 documents listed on the privilege log. D’onofrio v. SFX Sports Group, Inc. 247 F.R.D. 43 (D.D.C. 2008); D’onofrio v. SFX Sports Group, Inc. 254 F.R.D. 129 (D.D.C. 2008). The cases are described in Chapter 27 of my book Electronic Discovery (West 2010). This use of old paper based systems with today’s ESI is a big mistake and the primary reason e-discovery is so expensive.

The Paper Derived Processes and Methods for Search and Review Did Not Work When Applied to High Volumes of ESI

The old linear review methods involved serial culling of documents down to a final production set. The process generally required multiple reviews of the same document for different purposes. It was inefficient. It was expensive. Moreover, the quality control of human eyes on paper did not work with high volumes of documents. This is shown by the latest scientific experiments where the agreement rate among professional legal reviewers was found to be just less than 50%. Cormack, Grossman, Hedin, Oard, Overview of TREC 2010 Legal Track (February 21, 2012).

This tradition of multiple manual reviews, with only limited computer assistance, typically on a linear based review platform, still continues today. But it is too inexpensive and inefficient with high volumes of ESI. This will only get worse as the amount of information continues to grow exponentially. Jason Baron, the Head of Litigation for the National Archives and Records Administration, which is in charge of all federal records, including white house email, explains this as a problem of scale. He projects the number of White House emails will soon exceed One Billion. He estimates it would cost over $2 Billion to search that many emails. That assumes a team of one hundred full-time lawyers working over fifty-four years and a very low billing rate of $100 per hour. Information Inflation: Can the Legal System Adapt? Supra at pgs. 13-15. It also assumes computer-assisted review tools, but following the old paper-based linear review models. Also see: Baron, Jason, E-discovery and the Problem of Asymmetric Knowledge (Presentation at the Mercer Law School Ethics in the Digital Age Symposium, Nov. 2008).

Moreover, too many mistakes are being made when these traditional linear review methods are applied to the astronomical volumes and new medias of ESI. See Eg.: Diabetes Centers of America, Inc. v. Healthpia America, Inc., 2008 U.S. Dist. LEXIS 8362, 2008 WL 336382 (S.D. Tex. Feb. 5, 2008):

Plaintiff’s counsel conceded at the hearing that the task of searching Plaintiff’s records for relevant emails in response to Defendants’ discovery request was entrusted to a junior associate. It is apparent that the associate worked with little or no direction or supervision. The search terms used by the associate were inadequate — they did not even include the term “phone” — and, as a result, she failed to locate or perceive the significance of the emails about which Defendants now complain.

Also see Danis v. USN Communications, Inc., 2000 WL 1694325 (N.D. Ill. 2000) ($10,000 fine imposed against CEO personally when the young general counsel he hired to supervise ESI preservation was grossly negligent); Mt. Hawley Ins. Co. v. Felman Production, Inc., 2010 WL 1990555 (S.D. W. Va. May 18, 2010) (serious mistake made resulting in waiver of privilege in spite of sophisticated counsel with very elaborate processes and safeguards).

Outsourcing Alone is Not the Answer

Some are looking for an answer to these expense issues by keeping the old processes, but outsourcing the work of manual review to less expensive contract lawyers. They are called “contract lawyers” because the law firm that represents the client does not employ them. Instead, they work for the firm or some other company under a contract to do only review work. These contract lawyers may be located in India or other countries, or may be down the street from your office, or down the hall.

Even if you accept the abilities of contract lawyers to adequately perform the task of the first level relevance review (which I do, with qualifications), this is still just a stopgap measure. The volumes of ESI are increasing exponentially. Outsourcing may help with expenses to a point, but is still futile as a stand-alone long-term strategy, still just treading water in the midst of a flood. Indeed, when this outsourcing strategy was employed by the Department of Justice to try to reduce the costs of a privilege review, it still cost $9.09 per file to do a review at a total cost of $6,000.000. In re Fannie Mae Securities Litigation, 552 F.3d 814, (D.C. App. Jan. 6, 2009). The government reviewed 660,000 files at a cost of $6,000,000 using contract lawyers to respond to a non-party subpoena. The Order denying the government’s motion for cost-shifting was upheld by the appellate court.

The answer does not lie in modifying the system somewhat to employ cheap labor to do manual review. Outsourcing may still be part of the answer, and is likely to remain as a viable long-term strategy. But outsourcing only works when combined with the new systems and technologies like predictive coding, and only when properly supervised and performed by highly skilled professionals. The answer is not outsourcing. The answer is a whole new system for e-discovery.

Conclusion

Our current linear, confrontative, one-dimensional, largely manual, costly, Bates stamp approach to discovery must be replaced with a cooperative, iterative, largely automated, predictive coding based, proportionality cost-controlled, hash value approach.

For more on hash values and e-discovery, see: Losey, R., HASH: The New Bates Stamp, 12 Journal of Technology Law & Policy 1 (June 2007). For more on cost control see my blog Bottom Line Driven Proportional Review. For more on the new approaches in general and predictive coding in particular, see eg. Chris Dale’s Judge Peck’s Predictive Coding Opinion – reporting the reaction and the case and articles cited therein.

4 Comments | Evidence, Lawyers Duties, Metadata, New Rules, Review, Search, Technology | Permalink
Posted by Ralph Losey

LegalTech Interview of Dean Gonsowski on Predictive Coding and My Mission to Make Predictive Coding Software More Affordable

March 9, 2012

I highly recommend you watch this interview of an expert on e-discovery search whom I respect, Dean Gonsowski, Symantec eDiscovery Counsel. Dean is a former practicing attorney and general counsel who is now an expert on technology for Clearwell-Symantec. He has a good handle on the big picture of predictive coding type technologies, not only in litigation, but also in corporate information governance. The interview is well-done by Don McLaughlin, CEO of Falcon Discovery.

Stay to the end of the interview and learn about the possible use of predictive coding as an email management tool to file your email. I can just see the commercials of the near future! Let your computer do your filing for you. That way the computer can more easily retrieve your email when you want it. Manual self-filing will become a thing of the past, freeing up hours of time for the average information worker. I am all for that vision of the future. This is outlined by Dean, and so is the use of predictive coding in litigation. Dean talks at length, as did I at the time, about there being no real need to wait for a big decision approving predictive coding. We all used keyword search without an opinion approving its reasonability, so why should these new search technologies be different? Dean, like everyone at the time, pointed out that judicial approval of predictive coding might never come, or its coming could take years.

Dean is one of the key e-discovery people from Clearwell who are now helping to make Symantec a more dynamic, information-driven company. Their blog, e-discovery 2.0: thoughts about the evolution of e-discovery, is one of my favorites. They have new software called Clearwell Transparent Predictive Coding that will come out in a few months (price to be announced). I saw a private demo of it at LegalTech from the top people of Clearwell-Symantec. I told them then what I tell all of the vendors, bring the prices down! I will keep beating that drum so that predictive coding culling is affordable in all sizes of cases. It is the ideal tool to implement my Bottom Line Driven Proportional Review method. 

In the video interview above Dean Gonsowski gives a new perspective on technologies and legal methods that goes back to Dean’s early days in e-discovery in 2003. Poor Dean never gets much of a chance to talk when I appear on panels with him, which I have done now for Clearwell for the last two years at LegalTech. He is usually graciously allowing his panelists to speak while he asks questions. When the tables are turned he has many interesting things to say, as this You Tube video shows. Frankly, I have never heard Dean speak for so long as in this video.

As a final note you might want to calendar a free webinar I’m doing for Symantec and Inside Counsel on March 22, 2012 at 2:00 ET. It is called Is Keyword Search in eDiscovery Dead? I’ll answer that question, talk about the old days of endless games of Go Fish, some of my new search methods, and the Kleen Products case, but not the other case. Presenting with me and trying to get a word in edgewise is Matt Nelson, eDiscovery Counsel for the Symantec Intelligent Information Group.

I wonder who counsel is for their Dumb Information Group? Glad I’m not presenting with that guy! Anyway, did I mention it’s free? So give it a listen and ask a question or two, such as, what will your predictive coding software cost when it comes out, and can you make it less expensive? I will cheer you on, and even if you don’t ask (or, more like it they don’t read your question), I will keep on trying to influence all of the software leaders.

I want to make predictive coding software an affordable, everyday item. Now that I’ve helped to open the door, I want as many people as possible to be able to walk through. With the right methods to use this tool, a new world of affordable e-discovery awaits.

3 Comments | Evidence, Forensic Exam, Lawyers Duties, Metadata, New Rules, Review, Search, Technology | Permalink
Posted by Ralph Losey

Ethics of Electronic Discovery – Part Two

March 8, 2012

For Part One, see here.

Ethical Analysis of the Hypothetical

What ethical considerations and rules or professional conduct come into play in this scenario? Let us analyze the facts of the hypothetical one rule at a time and consider the impact of all six of the key rules: Rule 1.1 – Competence; Rule 1.3 – Diligence; Rule 1.6 – Confidentiality; Rule 3.2 – Expediting Litigation; Rule 3.3 – Candor Toward the Tribunal; and, Rule 3.4 – Fairness to Opposing Party and Counsel.

Competence. The competency issue here is critical, as it is in most ethical considerations. One party here, the plaintiffs’ counsel, does not really understood e-discovery. For instance, they did not really understand many of the technical reasons behind defendant’s position, such as deduplication and email copies throughout the system, nor the sampling disclosures. They did not understand the benefits of phased discovery and going first for the low hanging fruit, the emails of the best key custodians.

They did not cooperate in e-discovery because they did not know how. They probably had never even seen the cooperative model in action before, and so when they saw it here, they did not recognize it. They may instead have mistaken it for weak or timid opposing counsel. It just provoked them to be more caveman like. They got more aggressive in the face of the unknown. They also ran away from any real discussion on technical issues. A classic case of the fight or flight instincts of a pre-computer-literate-lawyer faced with e-discovery.

Plaintiffs’ counsel in this hypothetical did not trust the other side because they assumed they were like them. They were not competent enough to recognize or understand the new approach. They only understood “cooperation” from the outside, as just another tool, or keyword, in the game of hide-the-ball. They certainly had not read Professor Gensler’s article and had no understanding at all on how the new cooperative, trans-negotiation model might help their clients. In Professor Genzler’s words:

The Cooperation Proclamation is exactly right when it urges lawyers to see cooperation as a means for advancing their clients’ interests and not as a retreat from their duties as loyal advocates. As I have written elsewhere, the lawyers who default to battle mode in discovery – who fail even to consider whether cooperation might yield better results – are the ones who truly fail to serve their clients’ interests.

The Bull’s-Eye View of Cooperation in Discovery, 10 Sedona Conf. J. 363, at 363 (2009 Supp.)

They had also not read the article by the Sedona Conference making the legal case for cooperation, The Case for Cooperation, supra. The editor-in-chief of this article was well-known plaintiff’s counsel, William P. Butterfield. The concluding paragraph of The Case for Cooperation succulently warns of what may happen to our system of justice if the new cooperative model to discovery is not adopted:

If parties are expected to continue to manage discovery in the manner envisioned by the Federal Rules of Civil Procedure, cooperation will be necessary. Without such cooperation, discovery will become too expensive and time consuming for parties to effectively litigate their disputes.

The Case for Cooperation, Id. at 362.

Plaintiffs’ counsel here did not even seem to understand what most truly competent plaintiff’s counsel do, that every dollar spent on useless discovery is another dollar not available for settlement. The best, most competent counsel have always saved their powder for the real battles that count, on the law and application of the law to the facts. They have always understood that the true, and only valid purpose of discovery is to get at the key facts to allow reasoned evaluation of the case, not to prepare mountains of data, or extort the responding party, or bury the requesting party in a document dump. Competent legal counsel do not engage in discovery as abuse.  Those that do soon develop a reputation that follows them into a court room, even if it is on the other side of the country of where they usually practice. See Discovery As Abuse, 69 B.U. L. REV. 635 (1989). As the Sedona Conference noted, the “risk of gaining a reputation among the judiciary as unduly combative during discovery, encourages cooperative behavior.” The Case for Cooperation, supra at pg. 362.

The only possible conclusion here is that Plaintiffs’ were not competent to handle the e-discovery issues in this class-action employment case. For that reason alone, under this scenario they behaved unethically. They violated Rule 1.1.

Many attorneys in this situation attempt to meet their ethical obligation of competence by hiring an e-discovery vendor to advise them. Unfortunately, this usually does not work, for such experts frequently only tell the attorneys who hired them what they want to hear. The hired guns simply supply arguments jazzed up with tech-speak to support the legal argument of the attorneys who hired them.

In any event, e-discovery vendors are technologists, not lawyers, and even when a rare e-discovery expert at a vendor does also have a law degree, they do not provide legal advice. They are not allowed to provide legal advice. The only way an attorney who is not competent in the law and practice of e-discovery can fulfill their ethical duty is by taking the time and considerable efforts needed to become competent, or by bringing in legal counsel who is competent to assist him or her. Vendors cannot do that. They cannot render legal advice, but I appear to be the only one pointing out that obvious fact. See EDRM Model Code of Conduct. In fact, the EDRM proposed vendor ethical code of conduct, Principle 4 – Sound Process, seems like it encourages vendors to provide legal advice: “Service Providers should define, implement and audit documented sound processes that are designed to preserve legal defensibility.” This is a gray I know, but I strongly encourage the experts who are involved in this EDRM project to carefully consider the whole UPL issue and clarify draft Principle 4 and its comments on legal defensibility. By the way, except for my issue with Principle 4 and UPL, I am favorably impressed with the proposed model e-discovery vendor ethics guidelines that EDRM has come up with and encourage all vendors to study it carefully.

Diligence. Plaintiff’s counsel here thought they were being diligent because they were engaging in what they saw as vigorous advocacy. Too bad they were playing the wrong game. True vigorous advocacy here would have entailed detailed examination of the reasons provided by defense counsel for the five custodians and seven issues. It would have required reciprocal disclosures on their part on their thinking and analysis of the importance of the various custodians to the case. Diligence would have required immediate study of the emails produced. It would have taken the dispute out of rhetoric and knee-jerk opposition to everything the other side proposes, and into the facts and legal analysis.

Plaintiffs’ counsel in this scenario was not diligent at all. Their incompetence made that impossible. They thought they were being very diligent, and no doubt so did their clients, who are usually very easily impressed by the kind of saber-rattling in which they engaged. But in fact they did all of the wrong things.  They engaged in knee-jerk opposition to everything the other side proposed, and this pretend diligence was really not diligent at all. Discovery as abuse is not an exercise in competent diligence. It is abuse, pure and simple.

Confidentiality. Plaintiffs’ counsel here thought they were being very ethical by refusing to disclose their work product. They would not give the defendant an idea on their thinking of the case, on what information they thought would be highly relevant. They would not disclose why they thought some custodians and issues were more important. That was their confidential thinking, and they thought they should keep it secret. They hoped to keep their analysis of the case (assume they had one, and this was not just a superficial form-driven lawsuit) to themselves. They wanted to surprise defendants as much as possible. Indeed, they were initially surprised by how much confidential information the defense counsel here provided, which, again, they mistakenly mistook as a sign of weakness and egged them on to keep demanding more and more. Then they were surprised again when defense counsel said no, and never budged from the initial 5/7. And finally were surprised again by the court ruling against them and went with the defense plan.

So which attorneys in this scenario met their ethical duty under Rule 1.6 to “not reveal information relating to the representation of a client” and which did not? Rule 1.6 has numerous exceptions to the duty of client confidentiality, including where the client gives informed consent or the “disclosure is impliedly authorized in order to carry out the representation.”

Here the disclosure by defense counsel was needed to carry out the representation and so impliedly authorized, even if not specifically authorized. Note that under the scenario no attorney-client privileged documents are disclosed, nor any privileged communications. Even client confidential documents are protected under a strong confidentiality order. Any non-relevant documents produced would also be covered by the confidentiality order and would be promptly returned. Further protection was provided by a strong clawback order.

The only confidential information disclosed here is some of the work-product privileged thinking and analysis of defense counsel. Unlike the attorney-client privilege, which is held by the client and can only be released by the client, the work-product privilege is held by the attorney, and released by the attorney. The disclosure of the attorney’s work product was made here to advance their client’s interests. It was made to facilitate efficient, cost-effective search and production, and later to obtain a protective order preventing unduly burdensome, disproportionate e-discovery. This kind of disclosure does not violate defense counsel’s duties under Rule 1.6.

My conclusion is based on an understanding of how work product privilege waiver is different from attorney-client privileged waiver. It does not automatically open up the door for further inquiries as attorney-client waiver might do. My thanks to Professor Mark S. Sidoti, who is part of Professor Hamilton’s excellent email discussion group for e-discovery professors, full-time and adjunct. Professor Sidoti, whose day job is at Gibbons, pointed out a key case in this area, Williams & Connolly LLP v. U.S. Securities and Exchange Commission, 2010 U.S. Dist. LEXIS 78570 (D.D.C. Aug. 4, 2010):

The work product doctrine protects both deliberative materials such as mental impressions, conclusions, opinions, and legal theories and factual materials prepared in anticipation of litigation. *** In his declaration in support of summary judgment, David Frohlich, an Assistant Director in the SEC’s Division of Enforcement, explains that the handwritten notes sought by Plaintiff were generated during the Cendant investigation in anticipation of litigation with Mr. Corigliano, Mr. Kearney, and others. Plaintiff does not dispute this statement and concedes that these documents qualify as attorney work product. However, Plaintiff contends that the SEC has waived any work product privilege with respect to these documents because similar handwritten notes were disclosed to Plaintiff during the criminal prosecution of Mr. Forbes. According to the declaration of Christopher R. Hart, an attorney at Williams & Connolly LLP, at least eleven of the documents of handwritten notes identified in the Vaughn index were produced by the government during the Forbes matter. Plaintiff thus contends that the government has “waived the work product privilege as to entire subject matter of handwritten notes between SEC staff and Corigliano, Kearney, or their respective counsel.”

As with the attorney-client privilege, a party may waive the work product privilege through disclosure. See In re United Mine Workers of Am. Employee Benefit Plans Litig., 159 F.R.D. 307, 310 (D.D.C. 1994) (“It seems . . . clear in this Circuit that the disclosure of documents protected by the attorney work product privilege waives the protections of the attorney work product privilege as to the documents disclosed.”). However, “the test for waiving attorney work product protection is more stringent than the test for waiving attorney-client privilege.” Goodrich Corp. v. U.S. Envtl. Prot. Agency, 593 F. Supp. 2d 184, 191 (D.D.C. 2009). Although disclosure of documents waives attorney-client privilege with respect to all other communications related to the same subject matter, the scope of “subject matter waiver” with respect to work product materials is more limited. “[A] subject-matter waiver of the attorney work product privilege should only be found when it would be inconsistent with the purposes of the work product privilege to limit the waiver to the actual documents disclosed.” “Several factors figure into the analysis: whether disclosure was intentional or inadvertent, the breadth of the waiver sought, and the extent to which the requested documents would reveal litigation strategies or trial preparations.” (emphasis added)

The law provides greater protection to work-product waiver and so the disclosures made in the hypothetical should leave defense counsel protected from any further unwanted intrusions. Also see Rule 502 Federal Rules of Evidence, and note that our hypothetical assumes a strong confidentiality and clawback order under Rule 502(d).

As for the conduct of plaintiff’s counsel, it may surprise you to know that I do not conclude they have violated Rule 1.6. I must conclude that plaintiffs stonewalling actions in this case are also consistent with Rule 1.6. Although their behavior in this hypothetical was unethical under the other rules here considered, it was not under Rule 1.6. They have the right to keep all of their work product to themselves, even though it was stupid to do so, and thus a probable violation of Rule 1.1 on competence, but it is was their work-product. Of course, no rule should be considered in isolation, and work-product thinking is required to be disclosed under many rules of civil procedures, including especially Rule 26(f), and any time you are seeking relief from the court, or making final preparation for trial, not to mention trial itself.

In the same professorial discussion group my esteemed LegalTech debating adversary, Professor Craig Ball, had this to say concerning the policy behind limited work product disclosures, starting with a comment on pre-digital paper productions:

You may argue that we never got to look behind a lawyer’s decisions to produce or withhold even if the lawyer made the selections by throwing documents down the stairs and producing only what hit the next landing. But perhaps that’s not an approach we want to replicate in post-digital practice. We indulged ourselves in the belief that, right or wrong, relevance and privilege determinations were a lawyer’s to make and largely immune from being second-guessed absent evidence of gross dereliction or misconduct. But, as we move into the realm of search technology–and especially those like keyword search only lately appreciated to be deeply flawed–the tools and methods employed must be closely examined and tested, including by exposing them to adversarial challenge.

Perhaps we need to step away from our reflexive “we are lawyers and we are special,” in order to consider what approaches are calculated to best serve the ends of justice. Cooperation demands communication. If you believe the first is more than a pipe dream, you need to embrace the latter.

I have to agree with Craig on this one. Opposing counsel need to talk with each other and explain what they are doing in the area of search and production. You do not have to give up all your trade secrets, nor reveal all of your strategies, but you have to be prepared to disclose enough to show your reasonable, good faith efforts. What does cooperation look like? It looks like lawyers talking and making mutual disclosures needed to plan discovery in a case.

Expediting Litigation. The conduct of plaintiffs’ counsel here violated their duty to expedite litigation. They did so in spite of the fact that expedited litigation is almost always consistent with a plaintiff’s interest in a civil proceeding for a speedy trial. Plaintiffs’ counsel here unethically violated this duty to expedite by forcing unnecessary motion practice. Further, the closed approach of plaintiffs’ counsel to try to conceal the facts they really wanted in discovery probably also violated Rule 3.2, even of it did not violate Rule 1.6. See eg. DeGeer v. Gillis, 755 F. Supp. 2d 909, 930 (N.D. Ill. 2010) (if parties had participated in “candid, meaningful discussion of ESI at the outset of the case,” expensive and time consuming discovery and motions practice could have been avoided).

Candor Toward the Tribunal. I think the conduct of plaintiffs’ counsel in this hypothetical violated this fundamental Rule of Professional Conduct too, but admit their violation might be seen as technical, and is certainly the kind of conduct that goes on every day and is tolerated by both Bench and Bar. Plaintiffs’ counsel here did not “make a false statement of fact or law to a tribunal … or, offer evidence that the lawyer knows to be false.” But they did not provide the whole truth either. They made a demand in their motion to compel for 30 custodians and 30 issues, and they continued in this demand at the hearing before the tribunal. Let us assume that they also made oral representations to the judge at a hearing that they thought this 30/30 demand was necessary and appropriate under the facts of the case and governing law. But that is at best a half-truth because they have been willing all along to accept 20/20.

For that reason they did not display the kind of candor to the tribunal that I personally think is appropriate. Instead, they continued to play the only game they knew how, the negotiation game, and they treated the judge as just another player in the game. In a negotiation game, you are never completely candid. That would defeat the whole point of the game. If plaintiff’s counsel here were in fact candid to the tribunal, they would admit that 20/20 is their goal, but then again, that could be a slippery slope for them. It could also lead to admission of their basic incompetence to do e-discovery to begin with, to engage in meaningful discussion and analysis, much less to enter into a true cooperative dialogue with the other side.

Fairness to Opposing Party and Counsel. Some attorneys are surprised when they see the terms of this important Rule of Professional Conduct.  Candor to the tribunal is one thing, but fairness to the opposing party and their attorneys? That is simply not part of the culture of many lawyers and law firms! They seem surprised when they are reminded that the requirement is built into our rules of ethics. It is not a mere professional courtesy, as some think, it is an ethical imperative. Under Rule 3.4 a lawyer shall not “unlawfully obstruct another party’s access to evidence.” A lawyer shall not “conceal a document or other material that the lawyer knows or reasonably should know is relevant to a pending or a reasonably foreseeable proceeding; nor counsel or assist another person to do any such act.” A lawyer shall not in “make a frivolous discovery request or fail to make reasonably diligent effort to comply with a legally proper discovery request by an opposing party.”

The attorneys for the party responding to the discovery request in this hypothetical, defense counsel, complied with their ethical duties. My conclusion assumes that they sought to limit discovery for the grounds stated, burdensomeness and likely relevance. I assume they did not have any bad faith ulterior motives, such as an attempt to conceal evidence that they knew to be unfavorable to their client. I assume they were not just trying to obstruct access to evidence or conceal relevant ESI. Finally, I assume that their search and production efforts were reasonably diligent.

With these assumptions I conclude defense counsel acted ethically because their refusal to review and produce any more that five custodians and seven issues in the first round of phased discovery did not prevent disclosure of any evidence they knew to be relevant, and did not prevent plaintiffs’ later discovery from other custodians or issues that might prove necessary. It might delay it to a second phase, but not prevent it. If defense counsel should uncover a document outside of this scope, one that is obviously relevant, then under this rule they would be required to disclose such a document and should not conceal it.

I conclude to the contrary that under the given facts plaintiffs’ counsel violated the duty of fairness to opposing counsel because their initial request for discovery was frivolous. They signed and served a discovery request for 50 custodians and 50 issues, knowing that was excessive, and for that reason quickly dropping to 30/30, and ready to accept 20/20, which is what they really thought was reasonable all along. They used a request for production as a mere negotiation tool, and in so doing made an unethically frivolous discovery request. Note they also violated Rule 26(g) FRCP, which Judge Grimm in Mancia said was the most misunderstood and under utilized of all rules of procedure. Rule 26(g) is similar to Rule 11, but applies only to the discovery pleadings.  Rule 26(g)(1)(B) states that a signature on a discovery request constitutes an attorney’s certification that the request is:

(ii) not interposed for any improper purpose, such as to harass, cause unnecessary delay, or needlessly increase the cost of litigation; and…

(iii) neither unreasonable nor unduly burdensome or expensive, considering the needs of the case, prior discovery in the case, the amount in controversy, and the importance of the issues at stake in the action.

They thought 20 custodians and 20 issues was proportional, yet they signed a request for production seeking 50. They did so for purposes of setting up a negotiation. This is, I suggest, an improper purpose under 26(g)(1)(B)(ii). They knew 50 was unreasonable and unduly burdensome, yet they still signed the discovery request. Plaintiffs counsel in this hypothetical thereby intentionally violated 26(g)(1)(B)(iii) triggering a mandatory obligation under the rule for the court to impose sanctions.

This conduct by the requesting party was not only a civil rule violation, it was unethical. It violated Professional Rule of Conduct 34(c) by “knowingly disobeying an obligation under the rules of a tribunal,” namely Rule 26(g)(1)(B) and as explained before, violated ethics Rule  34(d) because it was frivolous.

Conclusion

Some may say my hypothetical is far-fetched, that attorneys do not engage in this type of behavior. I say, get real. It is an everyday occurrence. The only thing far-fetched about it is the simplicity of the facts, which I necessary injected into the hypothetical, and, the relatively mild nature of the violations. In my position as national e-discovery counsel for a 700 attorney law firm with 48 offices around the country, I see equivalent or worse behavior by opposing counsel almost every week. It is not exactly a crazy hell-zone as I used to suspect, and many still believe. Krueger v. Pelican Products Corp., C/A No. 87-2385-A (W.D. Okla. 1989) (J. Alley) (“If there is a hell to which disputatious, uncivil, vituperative lawyers go, let it be one in which the damned are eternally locked in discovery disputes with other lawyers of equally repugnant attributes.”) But it is bad, and we need to work together as a profession to break out of this prison.

Even though this hypothetical is an all-too-common scenario, as far as I know, no attorney has ever been reprimanded for an ethical violation of Rule 34 upon these circumstances. This ethics rule, like the 26(g) procedure rule, is a paper tiger. Indeed, you could say that about all of the ethics rules here discussed in the context of e-discovery. It is a new area of the law and State Bar Associations are naturally reluctant to enforce its rules in this virgin territory.

But just because attorneys are no yet being reprimanded or losing their licenses for these kinds of rule violations, does not mean we should not care about compliance. Following these well-established rules is the best way to stay on the straight and narrow when it comes to e-discovery. We cannot let old hide-the-ball practices morph into hide-the-byte operating systems.

Compliance with the rules of professional conduct, and the new doctrine of Cooperation that implements these rules (and the rules of civil procedure as Judge Grimm and Professor Genzler have pointed out), is the best way to avoid the threat warned of by Sedona and many others. It is the best way to avoid a future world of litigation where standard hide-the-byte operating systems make “discovery too expensive and time consuming for parties to effectively litigate their disputes.” The Case for Cooperation, Supra at 362.

Although a good argument can be made for the enactment of more new rules of civil procedure to adapt to the challenges of e-discovery, new rules of professional conduct are not needed. The ones we have are sufficient to guide us, but we need to take the time and effort to study and understand these rules. We need to discuss these rules and how they apply to the new situations presented in e-discovery practice. It is my hope that this early effort in that direction will stimulate more discussion and analysis of the subject.

4 Comments | Lawyers Duties, New Rules, Technology | Permalink
Posted by Ralph Losey

« Previous Entries
Next Entries »