The Sedona Conference® Commentary on Achieving Quality in the E-Discovery Process is a must read for anyone seeking to improve their skills in project management, especially in the core functions of search and review. One of its most important insights is that metrics and statistics are now indispensable tools of discovery. The importance of statistics to the law is actually an old insight that has taken a long time to materialize. The Sedona Commentary quotes the great jurist Oliver Wendell Holmes, Jr., shown right, who said in 1897:
For the rational study of the law the black letter man may be the man of the present, but the man of the future is the man of statistics and the master of economics.
The future has taken a lot longer to reach most lawyers than Justice Holmes expected. Statistics is still a stranger to most litigators, especially as a tool of discovery. But by the time you finish studying this new Commentary on Quality, you will see that statistics is a powerful tool of the here and now.
We can thank the hardworking Sedona Editors-in-Chief for this excellent new publication, Jason R. Baron and Macyl A. Burke. They were assisted by Senior Contributing Editor, Thomas Y. Allman, and Executive Editors, Richard G. Braman and Kenneth J. Withers, with input from Members of Working Group 1. This commentary is a project of The Sedona Conference® Working Group on Best Practices for Document Retention and Production.
Here is the opening paragraph of the Executive Summary, which should entice you, like it did me, to read more:
The legal profession is at a crossroads: the choice is between continuing to conduct discovery as it has “always been practiced” in a paper world — before the advent of computers, the Internet, and the exponential growth of electronically stored information (ESI) — or, alternatively, embracing new ways of thinking in today’s digital world. Cost-conscious clients and over-burdened judges are demanding that parties now undertake new approaches to solving litigation problems. The central aim of the present Commentary is to introduce and raise awareness about a variety of processes, tools, techniques, methods, and metrics that fall broadly under the umbrella term “quality measures,” and that may be of assistance in taming the ESI beast during the various phases of the discovery workflow process. These include greater use of project management, sampling, and other means to verify the accuracy of what constitutes the “output” of e-discovery. Such collective measures, drawn from a wide variety of scientific and management disciplines, are intended only as an entry-point for further discussion, rather than any type of all-inclusive checklist or cookie-cutter solution to all e-discovery issues.
The truth is, there can be no cookie-cutter solution or all-inclusive checklist for a subject as complex and dynamic as e-discovery. Yet, those of us who specialize in this area get asked for such easy-buttons all of the time. The best that can be hoped for is competency training and the gift of this writing: quality control procedures.
Critique of the Five Reasons Stated for Quality Control
Why is quality so important to the e-discovery process? The obvious answer is to avoid mistakes and the sanctions that can come with mistakes. The Sedona Commentary agrees that such risk management is the primary factor, but then, at page one lists four other reasons quality control is important:
- “Failure to employ a quality e-discovery process can result in failure to uncover or disclose relevant evidence which can affect the outcome of litigation.” I call this the “adequate recall factor” where you find enough of the truth for justice to be done in a case.
- “An inadequate e-discovery process may allow privileged or confidential information to be inadvertently produced.” I call this the “adequate confidentiality factor” where you protect enough of the confidential information to meet your client’s specifications. Some may be quite paranoid regarding disclosure of any privileged or confidential information, and willing to spend vast sums of money to avoid it. Others may not care as much if a few attorney-client emails, out of thousands, slip through the cracks and are more willing to rely upon the claw-back protections of Evidence Rule 502 to save money.
- “Procedures that measure the quality of an e-discovery process allow timely course corrections and provide greater assurance of accuracy, especially of innovative processes.” I consider this to be a secondary “adequate recall factor,” which allows for quality adjustment of search protocols and other processes as results are measured, new facts uncovered, issues evolve, and insights gained.
- “A poorly planned effort can also cost more money in the long run if the deficiencies ultimately require that e-discovery must be redone.” I call this the “do-over avoidance factor,” where if you do something right the first time, you do not have to pay to do it again. This, for me, is just one aspect of a larger “money-savings factor” that can result from quality processes.
I agree that all five factors are important, but I am inclined to think that the economic savings that can result from quality control are equally important to the e-discovery process as risk management, although perhaps not as obvious. As Justice Holmes said over 100 years ago, we lawyers of his future must not only be men and women of statistics, but also “masters of economics.” The savings can not only be realized by avoiding costly do-overs, as the Commentary points out, but also by increasing culling quantity and review speed. Quality controls make it possible to significantly reduce the amount of ESI to review, and to reduce the ESI volume in a manner that is legally defensible. This culling process, as shown generally in the diagram below, is key to dramatic reduction of the costs of e-discovery.
The process of intelligent reduction of data size prior to review is an essential component of what is now being called “early case assessment.” One of the Senior Editors of the Commentary on Quality, Jason R. Baron, wrote about this type of case assessment culling in another short paper he wrote with Ronni D. Solomon for use at a Sedona Conference Institute CLE entitled Bake Offs, Demos & Kicking the Tires. In spite of this title, it is well worth reading. See especially their Tip 5 where they talk about utilizing keyword “black lists” to reduce ESI size before full review. They also mention the critical need to see the results of keywords, not just hit totals, in order to make an intelligent choice of effective keywords. It is no longer an acceptable practice to choose keywords in the blind. It results in weak culling and thus excessive review. We need to see the results of keyword filtering to be able to aggressively reduce the volume of ESI and still maintain quality.
This type of early case assessment culling is critical because the review stage is by far the most expensive step in e-discovery. Anything that cuts the amount of ESI to be reviewed has a direct, substantial impact on the bottom line. Of course, better, faster, and more efficient review of the ESI can also reduce costs. See Bake Offs, Tip 3, and the discussion of clustering tools. With proper quality controls, the costs savings from culling can be realized without sacrificing the other four goals stated in the Commentary on Quality.
In today’s economy, the money-saving aspects of quality control are just as important as risk management, and, in my view, more important than the secondary benefits of “adequate recall factors” and “confidentiality factors.” E-discovery costs must be significantly reigned in for the civil justice system to avoid the danger of replacement by private arbitration, or worse, by self-help. For that reason, we must leave the old paradigm of total-recall in favor of a more realistic, cost-controlled view. As I often say these days, how much of the truth can a particular dispute afford in view of the constraints of proportionality and Rule 26(b)(2)(C)?
This is the public comment version of the Sedona Commentary on Quality and it is my hope that the final version will place a greater emphasis on the economic savings possible with quality controls in an early case assessment environment. As Justice Holmes said: “It is the province of knowledge to speak and it is the privilege of wisdom to listen.”
Four Guiding Principles Behind the Commentary
The Executive Summary also contains an explanation of the four guiding principles behind the commentary:
Principle 1. In cases involving ESI of increasing scope and complexity, the attorney in charge should utilize project management and exercise leadership to ensure that a reasonable process has been followed by his or her legal team to locate responsive material.
Principle 2. Parties should employ reasonable forms or measures of quality at appropriate points in the ediscovery process, consistent with the needs of the case and their legal and ethical responsibilities.
Principle 3. Implementing a well thought out e-discovery “process” should seek to enhance the overall quality of the production in the form of: (a) reducing the time from request to response; (b) reducing cost; and (c) improving the accuracy and completeness of responses to requests.
Principle 4. Practicing cooperation and striving for greater transparency within the adversary paradigm are key ingredients to obtaining a better quality outcome in e-discovery. Parties should confer early in discovery, including, where appropriate, exchanging information on any quality measures which may be used.
I was pleased to see that cooperation was included as the fourth principle behind quality control. As I explained in my blog, There Can Be No Justice Unless Lawyers Maintain High Ethical Standards, cooperation is an ethical imperative that will necessarily result in substantial costs savings by the avoidance of litigation churning. The Commentary correctly notes that cooperation can also cause a significant increase in quality.
All four of these Principles behind Quality are, in turn, based on the fourteen Sedona Principles, but especially on Principle 11, which states:
A responding party may satisfy its good faith obligation to preserve and produce relevant electronically stored information by using electronic tools and processes, such as data sampling, searching, or the use of selection criteria, to identify data reasonably likely to contain relevant information.
For that reason, much of the Commentary on Achieving Quality in the E-Discovery Process is focused on electronic tools and processes, including sampling and other search methods.
The Commentary Is Not A Recipe Book
This Commentary does not purport to give specific advice on how to maintain quality. So, if you are looking for forms and recipes, you will be disappointed. Sedona was quite correct to so limit this project. Here is their explanation, with which I fully agree:
This Commentary is not intended to serve as a comprehensive roadmap covering all possible uses of quality measures and metrics throughout the e-discovery process. The creativity of the vendors and the bar will ensure that concrete applications of quality techniques will be advanced. Nor have we any bias towards particular methods, tools or technologies or a point of view that asserts that sampling or other types of quality measures are invariably required in every type of litigation. Indeed, the drafters believe that the solution to problems created by scale are not solved by technology per se, which is merely a tool, but by better use of team leader skills, project management, and quality measures.
Although the Commentary discusses many tools including my favorite — sampling — it does not set out a step-by-step process for handling e-discovery projects. There is too much variability in cases and facts for that, and frankly, it is still too early for any such best practices specifications. Electronic discovery project management is still in its early stages and many of us who have developed methods tend to keep them close to the vest for competitive purposes. Plus, it is not Sedona’s role to endorse one particular technique over another, but rather to address issues on a high level and facilitate further study and dialogue. If you read this Commentary with this expectation, you will not be disappointed. Indeed, you will find that it contains many valuable treasures and insights. As Oliver Wendell Holmes, Jr., said: “A moment’s insight is sometimes worth a lifetime’s experience.”
Key Elements of Successful Project Management
The Commentary lists seven elements that Sedona considers critical for successful project management. They do not claim that the list is exhaustive, nor do they purport to rank them in importance. These elements are:
Although Sedona will not rank these seven, I will. I think that Leadership and Measurement are the two most important factors and the two most difficult. A close third is Documentation, which is key to protection, especially if any of the million decisions made in the course of a project prove to be a mistake. The courts do not expect perfection from attorneys, only expertise, reasonability, and good faith. Good documentation of the process can make it easier to recreate what happened. It will help you to convince the court that you thought you were doing the right thing at the time, even though later events may suggest otherwise. Personally, I do not like the documentation process at all and find it to be as odious as filing. But I know from experience just how valuable it can be, not only to later show what you did, but also in real time to help you to keep track of what you are doing.
Leadership is obviously important. Someone has to be in charge of a project and it needs to be an attorney with special expertise. It is a mistake for the trial attorney to try to fill this role because they probably do not have the necessary skills nor time. But it is also a mistake to delegate the job to a non-lawyer. I continue to believe that lawyers must remain masters of the discovery process and they abdicate their responsibilities when they over-delegate to vendors. The Sedona Commentary (at page 7) agrees that attorneys should retain the “Team Leader” position and should only look to outside vendors for competent assistance, which, it correctly states, is “often essential.”
The importance of “Measurement” may not be as obvious as Leadership, but the Commentary does a good job of explaining just how indispensible it is to quality control. It lists five types of quality measurements that are especially useful in e-discovery:
1. Judgmental Sampling
2. Independent Testing
3. Reconciliation Techniques
4. Inspection to verify and report discrepancies
5. Statistical Sampling
Of these five measures of quality, the two types of sampling are, in my opinion, the most important. The two types of sampling are defined and explained in detail in Appendix A: Sampling 101 for the E-Discovery Lawyer:
Judgmental Sampling: Sampling performed on a sample set that was selected based on the judgment of the person doing the sampling. … A common example in the e-discovery context would be keyword searching itself, which is a more-or less informed technique universally used by lawyers and legal professionals to produce a sample slice of a given ESI universe of data, based on the a priori judgment of those selecting the keyword terms.
Statistical Sampling: Probability sampling, or random sampling, is a sampling technique in which the probability of getting any particular sample may be calculated. … A random sample is one chosen by a method involving an unpredictable component.
In the body of the text at page eleven, the Commentary explains that “statistical sampling can serve as a check on the effectiveness of search terms and other automated tools in identifying responsive information.” You can, for instance, use random sampling to test small subsets of the data selected by judgmental sampling, such as keyword culling. Then, by quick reviews of random sample subsets, you can determine the effectiveness of the keywords to identify responsive information. You can then adjust your keywords accordingly and try the new search cull terms again on a new sample. Here is how the Commentary describes this process at page thirteen: “Trial or pilot runs of combinations of words may be tested in an iterative fashion to extrapolate the effectiveness of the chosen set.”
You can use this iterative method to increase the culling rate of ESI to a size where cost projections of final review expenses finally come within the project budget. This use of both judgmental and random sampling methods, coupled with cost estimations, is close to the kind of early case assessment quality control procedure that I have developed to control e-discovery expenses in a legally defensible manner. I would tell you more, but it gets extremely complicated, is case sensitive, and frankly, leads into proprietary territory.
Applying Quality Measures
Part Three of the Commentary, entitled “Applying Quality Measures in E-Discovery,” is probably the section that will be of most interest to practitioners. It divides the analysis into two segments: the Data Collection Phase; and, the Review and Production Phase. The Data Collection Phase is examined in three segments: “Building on Traditional Approaches to Document Collection; Applying Measures of Quality to the Data Collection Process; and, Best Practice Guidelines.”
The Best Practices Guidelines at page fifteen begins with the following good advice:
The selection, organization and filtering of ESI through the use of a search protocol is a critical element in reducing the volume of information to be collected and thus the time and cost of collection. In addition, keyword search techniques are well known and may be used for this purpose. More advanced technologies have emerged that employ complex algorithms for ESI filtering and organization and may, in some cases, be useful at the collection stage. Regardless of the technology chosen, all filtering methods require a well-defined process. Without these basic steps, the use of any filtering technology will likely result in gross over- or under-inclusion of responsive ESI. The process includes several steps:
• Understanding the composition of source ESI;
• Defining the goals of the filtering;
• Applying the filter and testing the results.
On this last filtration/testing step the Commentary makes the basic, yet important point missed by most practitioners still using negotiated keyword searching, that:
The filtering process should be iterative and needs to be repeated until the desired goals are met. It is not sufficient to blindly run a filtering tool and trust that it is achieving the desired results. One must evaluate the outcome of the search, looking to identify errors in how the filter rules were set up or applied. Key metrics, such as the number of included or excluded documents by keyword or filtering criteria, can be used to evaluate the outcome. Examining the high and low number of search hits can uncover issues with how the search was constructed, the choice of terms, or even issues with the data.
No one is good enough to pick good keywords off the top of their head, much less negotiate a good set of keywords. Words are, after all, so maliable and differ tremendously between person to person. As Oliver Wendell Holmes said:
A word is not a crystal, transparent and unchanging, it is the skin of a living thought and may vary greatly in colour and content according to the circumstances and time in which it is used.
Yet keyword search is still the practice used by most lawyers today and is often ordered by the court. American Family Mutual Ins. Co. v. Gustafson, 2009 WL 641297 at *3 (D.Co. March 10, 2009) (“the parties shall forthwith meet, confer, and agree upon the search terms”). This is a mistake. The “skin of living thoughts” is not so easily snared. Testing of proposed search terms should always be required. Otherwise, your review will be haphazard at best, and a complete waste of time and money at worst. As Oliver Wendall Holmes also said: “Lawyers spend a great deal of their time shoveling smoke.”
Quality controls in the review and production phase are also examined at length in the Commentary. The discussion includes: automated methods to reduce the initial burden of review; “clawback” agreements, Rule 502, and reliance on automated methods; quality control guidelines for responsiveness and privilege; and, final quality checking at production.
The Commentary conclusion begins with a quote I like a lot by William Gibson: “The future is already here – it’s just not evenly distributed yet.” I know many people in this field feel like that is the story of their life. Certainly there is a wide variation in the U.S. and around the world in how the discovery of written evidence is conducted. The Commentary ends with these fine words of wisdom:
In the end, cost-conscious firms, corporations, and institutions of all kinds intent on best practices, as well as over-burdened judges, will demand that parties undertake new ways of thinking about how to solve discovery problems — including employing better project management and better measures of quality to achieve optimum results, as outlined here. The technical and management-orientated quality processes discussed above need to be incorporated into every trial lawyer’s continuing education and daily practice. These processes also dovetail with, and support The Sedona Conference® Cooperation Proclamation — which calls for incorporation of the best thinking of “disciplines outside the law” to achieve the goal of the “just, speedy, and inexpensive” determination of every action. In the end, striving to attain a quality outcome in the conduct of litigation is consistent with the highest ethical calling of the legal profession.
I agree with these noble aspirations, but think it is unrealistic to think that these processes will, or even should, “be incorporated into every trial lawyer’s continuing education and daily practice.” Not every trial lawyer will be interested in random sampling, iteration, linguistic analysis, the latest concept-search engines, ESI architecture, leadership of complex e-discovery projects, and ESI architecture, not to mention the ever changing technologies that create and store electronic information. I agree that all trial lawyers should have some exposure to this and to the idea of quality control, in the same way that all lawyers should have some exposure to antitrust law. But I doubt very much that the subjects in the Quality Commentary are going to be part of “every trial lawyer’s daily practice” anytime soon (if ever). Instead, they will be front and center in the practice of attorneys who specialize in e-discovery.
In the inconsistent future here-now that I see, e-discovery specialists will work closely with trial lawyer specialists. Team-work will be common, even on small cases. Some trial lawyers may have the time and inclination to handle e-discovery themselves, especially in less complicated situations. But, for the foreseeable future, they are likely to be few and far between. As William Gibson says: “Time moves in one direction, memory in another.” Instead, most trial lawyers will work with, not replace, the e-discovery lawyers. This could be a very small team of just two persons, like Perry Mason and Paul Drake, with a clear division of labor and skills, or, in the largest cases, it could be a team of many lawyers, paralegals, technicians, engineers, and information scientists.
The new age of information is too complicated to continue the old practices and traditions where both trial and discovery skills were combined and held by all trial lawyers. It worked when the documents were paper and few in number. But those days are nearly gone. Now we have ephemeral electronic paper that throws itself away when you are not looking. We have needles of relevant evidence hidden in vast electronic haystacks that are larger and more complicated that you can imagine; haystacks that daily change and grow. As our best experts tell us, search is hard. To use Gibson’s words:
I don’t have to write about the future. For most people, the present is enough like the future to be pretty scary.
As a consequence, document discovery is far more complicated than it was before and requires special skills to be done right. It is time for the profession to change. As Justice Holmes said:
I find the great thing in this world is not so much where we stand, as in what direction we are moving — we must sail sometimes with the wind and sometimes against it — but we must sail, and not drift, nor lie at anchor.
The basics of e-discovery can and should be taught to all trial lawyers. Since they are generally a very smart group, they can learn the basics, if they will take the time and effort needed to do so. (So far, not many have been inclined to make this effort. Most seem to hate e-discovery, but this will change soon.) These basic skills, once learned, can suffice for many small cases, with just an occasional assist from a 21st Century version of Paul Drake. But the larger, more important cases will need the skills of a specialist; skills such as those outlined in The Sedona Conference® Commentary on Achieving Quality in the E-Discovery Process.
In today’s world of dispute resolution, the client with a sophisticated matter is better served by specialized services with a division of labor. The e-discovery lawyers will possess the skills and quality control techniques discussed in this Commentary, as well as the many other skills discussed in the many other Sedona Commentaries and other books and articles on the subject. These skills take time to learn and time to practice and maintain. There is not enough time to also learn the many, very different skills needed by a trial lawyer. As a result, in complex cases the discovery lawyers will go through the electronic maze to find the facts and their trial lawyer partners will present them to the court and argue their significance.
Litigation is already a team effort in most law firms. This trend will continue to grow and the clients will be better served because of it. Far from being more expensive, as you might think because more people are involved, the discovery and trial lawyer team will save money. The Fannie Mae type cases of outrageous e-discovery expenses only happen when trial lawyers dabble in e-discovery and make huge mistakes. A true specialist will not only do things right the first time, with quality, but do them quicker and less expensively. In this way, the e-discovery teams of the future will help preserve our system of justice by making discovery affordable again.