Practical Considerations in Dealing with Electronic Evidence

By Richard T. Seymour*

Click here to download a PDF version of this paper

* Law Office of Richard T. Seymour, P.L.L.C., 1150 Connecticut Avenue N.W., Suite900, Washington, D.C. 20036-4129. Telephone: 202-862-4320. Cell: 202-549-1454. Facsimile: 800-805-1065. E-mail: [email protected]. Some of the information in thispaper is used with permission from an upcoming edition of Richard T. Seymour and John F. Aslin, Equal Employment Law Update (Bureau of National Affairs, Washington, D.C., 2007), copyright © American Bar Association, 2007. For copies, contact BNA at 1-800-960-1220; seminar attendees are entitled to a 25% discount. Mention priority code EQL in order to receive the discount.

My web site,, has downloadable CLE papers.

Table of Contents

A. Case Selection Information

In deciding whether to take a class action or FLSA collective action, the potential cost of litigation is an important factor. If the defendant and its vendors have much of the critical information in computer-readable form, it will enormously decrease transaction costs for both sides.

Analysis of the electronic data may also enable each side to form relatively quick estimates of the value of the case, enabling the parties to engage in early settlement discussions.

No matter how many employees potential plaintiffs’ counsel may interview in the pre-filing or even pre-retainer investigation, it often occurs that their perceptions of the causes of their problems differ markedly from the real policies or practices causing their problems. Many employers see no need to explain themselves to their employees, or provide such partial information that they sow confusion. They pay a price in more frequent litigation, but are happy because they get summary judgments. It would be better to be clearer and not be sued in the first place, but until they absorb that lesson plaintiffs’ counsel will have to continue to be very careful about relying on their clients’ perceptions. Access to the electronic information leads with some frequency to an understanding of the real nature of the problem, requiring a fresh evaluation of the case.

Some companies or agencies with extensive computer-readable information can make use of it to persuade plaintiffs’ counsel, even prior to the filing of litigation, that there is no good case to be made. This obviously requires attorneys of high integrity and professionalism on both sides of the case.

B. Whose Electronic Evidence?

Practitioners should never fall into the trap of assuming that only the company may have relevant electronic evidence.

State Employment Services and private employment agencies may have computer files showing referrals to the employer, with information on qualifications, race, gender, age, and disability.

Numerous employers contract out the initial screening to third-party companies, which may administer a test on-line or over the phone, and which may administer detailed questionnaires on education, prior experience, and minute details of qualifications. The third party may have records of applicants the defendant never obtained, or obtained but no longer has accessible.

For higher-level positions, a corporate parent may receive information and issue instructions. For lower-level positions, it may issue changes to local qualification standards. For example, there are still companies that insist on higher educational and other qualifications whenever nonwhites are a large proportion of the available pool, and lower qualifications elsewhere. It is good to know this.

Plaintiffs’ counsel or their expert may generate the most important computer-readable files. In one case, we computerized the entire job and department initial assignment, promotion, and disciplinary information, with details of seniority, experience and education, for all employees in a textile mill over the 15 years since the mill opened. We needed it.

Hybrid files are also common. These occur when employer or third-party files are supplemented with other information counsel deem relevant, either of their own creation or by merging data from multiple sources. For example, an employer’s data file on hires may be merged with a third-party screener’s files to show information on the relative qualifications of hired male applicants and unhired female applicants.

C. Privilege Considerations

When counsel directly or through an expert or in-house personnel create or merge electronic files for the purpose of litigation, the work-product privilege and sometimes attorney-client privilege have to be considered.

If the new or merged files will eventually give rise to exhibits in connection with a motion or to trial exhibits, most courts will order the parent files to be turned over to the opposing side. The timing of the turn-over may be important in the litigation and in settlement discussions. Counsel may legitimately not reach a decision to use trial exhibits stemming from such files until late in the day, and it is doubtful that any obligation to turn over the privileged parent files could arise before that decision is made.

If opposing counsel have high integrity and professionalism, it is possible to work out a free-flowing exchange of information on merged and created files where counsel believe the files may be used in the future. Both sides may cooperate in the creation of a data base that both sides may then use in the litigation. The creation and checking of such a database can require extensive calendar and professional time, but will still often save a multiple of that professional time in avoiding a myriad of challenges for which the trial court will have little time or patience.

One of the many steep costs defendants justly pay for retaining junkyard dogs as counsel is that these types of case-shortening and cost-cutting techniques are not available.

D. Identifying Potential In-House Electronic Evidence

Before commencing a proceeding, it is advisable to meet with as many employees and former employees as possible, to figure out the company’s recordkeeping system. Paycheck stubs can be useful, as can all of the computer-generated personnel and payroll information employees receive.

For example, if employees receive company newsletters at their home, the employer will be hard-pressed to assert that it does not have a record of employee addresses.

If employees receive printouts of their job histories, qualifications on file, statements of interest in higher-paid positions on file, and hours worked in the current week, potential counsel for the employees knows that these kinds of information will not have to be created from scratch. The case may be litigable with significantly lower transaction costs for both sides, making more relief available to the clients.

It is critical to gather facts in advance of a problem, to support a potential claim for obstruction of justice or spoliation in the event that the company destroys currently available information as soon as it gets wind of a possible legal proceeding. Each current employee in contact with plaintiffs’ counsel should check to see how far back his or her e-mails are currently retained on the company’s system.

E. Ensuring the Preservation of Electronic Evidence

As soon as plaintiffs’ counsel decides to accept the retainer and it is wise to reveal this fact to the prospective defendant, plaintiffs’ counsel should notify the defendant, at a high level, of the need to guarantee the preservation of electronic as well as documentary evidence. If there are no outside counsel yet known to plaintiffs’ counsel, a letter should be sent to the CEO, Chairman of the Board, and General Counsel. If outside counsel are known, the letter should go only to outside counsel.

The letter should refer specifically to electronic evidence; some defense counsel routinely caution their clients to preserve documentary evidence but do not think to include electronic evidence, back-up tapes, etc., in their caution.

Setting the electronic evidence aside does not necessarily preserve it. The files need to be copied onto fresh media every so often, or they will deteriorate.

It is also important to make sure that the employer either preserves the means to read the files, or converts the files into a generally readable format. For example, I was in a case in which intervenors wanted access to electronic files the defendant was under a legal duty to maintain. The files were in fact in the computer room. Unfortunately, their format could be read only by the computer they had sold to a South American company when they upgraded, and they had not bothered to migrate the files onto the new system. On deposition, the IT official said that the law required them to keep the files, but does not require that they be readable.

If there is no response, or if there is any resistance, to the letter, plaintiffs’ counsel should consider filing an action to preserve the evidence or to obtain pre-suit copies of the evidence that can then be preserved properly.

On more than one occasion, it was only in producing the files that the defendant became aware of their deterioration. Sometimes, one expert’s equipment was able to use the underlying data where another expert’s equipment was not. Sometimes, plaintiffs’ copy of the data survives and the defendant’s own original does not.

There are enough problems that it is in the interests of both sides to have a constructive working relationship as to at least the technical factors.

F. Objects of Discovery

1. Data Bases

It is critical to know what is available. Much may be irrelevant, but it is hard to make that decision without knowing what is available.

For example, in one case defense counsel had assured me the company had no computerized record of the race of employees. He had checked personally. When we took a deposition and inquired of a data processor the meaning of each field in the employee database, including the obviously irrelevant ones, it transpired that the company did not use the “Christmas Bonus” field for information on Christmas bonuses, because it did not pay such bonuses. That was the field where race codes were stored.

Practitioners should never assume the honesty of the information in the files. Even defense counsel may be deceived by his or her client’s misdirection in data-processing files. I once had a class case in which defense counsel was convinced his client faced no risk because the vast majority of African American applicants had taken themselves out of the selection process by failing to return required documents. His client submitted the printout as part of sworn discovery responses. When we did a spot-check and called the applicants, however, some of them still had the computer-generated postcards or letters stating that they had failed the hiring test. All they “failed to return” was a passing score.

In a proper case, it may be worthwhile to compare a file that may have been “doctored” with an earlier version available from the back-up tapes.

A file that looks relevant may in fact be unintelligible without other files that do not look relevant. For example, a “Job Assignment” file looks relevant but may contain only two dates and four codes. The dates would be the starting and ending dates of assignment to the job, the first code may relate to a second file containing the employee name and SSN, the second code may relate to a third file containing the facility with the job, the third code may relate to the job name, the fourth code may relate to shift, the facility and job files may in turn relate to a fourth file containing the pay rate, and the employee name and SSN file may relate to an fifth file containing demographic information about the employee, a sixth file containing information about the employee’s qualifications, and a seventh file containing information about the employee’s performance evaluations.

Plaintiffs’ counsel should ask for copies of all manuals, memoranda, checklists, etc., used by any persons in any job categories or departments who enter personnel information into the computer files or get reports out. They will need samples of blank forms and of filled-in forms, and will need to compare all of these sources with what is in the computer files.

These should be studied in detail before a deposition. One cannot “wing” these matters.

Assuming that the files are of manageable size, it is often much cheaper to get the entire cluster of files than it would be to go into the files and redact information. Redaction can add enormously to the costs of production.

2. Getting Documents in their Native Format

The “properties” of a document can contain critical information. For example, an employer will tend to win if plaintiff cannot dent the decision maker’s statement that he had decided to fire the plaintiff and started to prepare the paperwork a week before she filed an EEOC charge. If the plaintiff gets a copy of the termination memorandum from the computer network in “native format”—i.e., with the “properties” information still attached—she will tend to win if the memorandum shows that the termination paperwork was initiated the day after she filed her charge.

One should always keep in mind that the company’s decision to defend a claim may have resulted from its having been deceived by the decision maker. Without the plaintiff’s pointing out the “properties” information, the defendant may honestly believe it is in the right, and refuse to settle on reasonable terms. With that information, it can make a more realistic choice about the risks of litigation.

3. E-Mails

E-mails are an enormously powerful tool for getting to the heart of what happened, and why. A single e-mail may alter the result of the case.

The problem is finding that single e-mail. I get about 200 e-mails a business day, with 30 to 50 more each weekend day. That’s 1,100 a week, or 57,200 e-mails a year. If I am a pack rat and save my e-mails to my hard drive, in the four years I’ve been with my firm I would have accumulated 228,800 e-mails. As with most firms and organizations, some would be in the active e-mail folder, some in an archive folder, some on back-up tapes as active or archived, some only on my hard drive, and all potentially available in senders’ and recipients’ e-mail folders.

To avoid looking at every e-mail—because life is short, resources are limited, and the case has other demands—one must come up with an algorithm to search sources of e-mail records. The needs of the case will produce different search imperatives. We all know how difficult it is to get the precise case we need to come up in a search of computerized data bases based on strings of key words, because there are so many ways courts can phrase things. The problem is compounded with e-mails, because at least courts use formal English.

For example, in the hypothetical we would all agree that the following kinds of e-mails between senior officials and Paula Plaintiff’s supervisor would be of great interest to her:

  • Someone’s been pulling the stats on our picks. I suspect you-know-who. Can you give her enough work so she doesn’t have time for this garbage?
  • Get serious! How many female golf pros can you name?
  • What have you got on our problem situation?
  • Somebody’s getting ideas above their station. A little discouragement is in order, but be careful. We want to keep the employee.

This illustrates why a multi-tiered strategy works best. Key-word searches are important, but it is also important to monitor all the e-mail traffic of key individuals. If substantial traffic consists only of routine reports of some kind—which will be readily apparent in an eyeball inspection—they can be excluded.

It is also useful to test algorithms in advance on small sets of e-mails, with problematic e-mails drafted for the purpose and inserted, to see how well the algorithm does in practice. Computerized legal searches are often fine-tuned when one sees what the search terms included that should have been excluded, and excluded that should have been included.

It is highly desirable that the full search be gotten right, the first time.

4. Document “Retention” Policies

E-mails present special preservation imperatives. The advice of some defense counsel is to ensure that e-mails are purged quickly. Some companies have document-destruction policies—uniformly called “document retention” policies in Orwellian doublespeak—resulting in the automatic erasure of e-mails after three weeks.

Putting aside the legal requirements that records relevant to compliance with the fair employment laws be retained, employers dancing on this sword should remember that it is two-edged, competent plaintiffs’ counsel can often make them slip, and falling on sharp objects—even of your own devising—can be painful.

For example, if plaintiffs can put together a decent case despite the destruction of documents, there are a thousand ways in which capable counsel can bring to the jury’s attention that the documents were not preserved. Juries do not like that. In an extreme case, if defense counsel suggested application of the “document retention” shredder after it became reasonably clear a claim would be filed, defense counsel may be disqualified because counsel and their firm may become co-defendants in an obstruction-of-justice or spoliation count.

Even if the destruction is inadvertent rather than purposeful, it may result in adverse inferences where the defendant violated a duty to preserve the documents, such as a duty imposed by a record preservation order.

Finally, even if a destruction is inadvertent and there was no violation of a court order, the destroyed documents are often the employer’s only record of any nondiscriminatory explanation. In a class action, the absence of supporting documentation means that the employer literally has no idea why any facially-qualified candidate was not selected. Some claims will be defeated on obvious grounds, but there may be no remaining defense as to the vast majority of claimants.

G. Meeting of Counsel on Electronic Discovery

Once the necessary documents are in hand and have been studied, and sometimes after the basic computer information has already been produced, the most effective and cost-efficient step is often to schedule a meeting of counsel, with each side bringing their data-processing people. When data-processing people speak directly, they can often come to a quick understanding of the nature of the data files and of any problems in their structure or use, translate it into English in each other’s hearing for the benefit of counsel, and enable counsel to get to the heart of the matter.

When each side communicates only between counsel, with the computer experts speaking only to counsel, the result is about as effective as trying to connect to the Internet without a modem, by imitating the sounds modems make.

Court reporters can take down the meeting, but it is often more useful to cover the necessary ground informally, and then, if needed, to have a deposition laying everything out in legally useable form. If a relationship of trust can be developed, it can continue throughout the case. My experience is that informal consultations can take place much more frequently with respect to electronic information than with respect to the traditional subjects of discovery, eliminating evidentiary problems and objections.

H. Reducing Transaction Costs

Plaintiffs, with more limited resources, are sometimes the parties most interested in reducing transaction costs. The reduction of transaction costs also provides a saving to defendants, with the ability to pay more to resolve the case because less is being wasted in litigation.

In the course of litigation, a lot of opportunities for cost reduction arise. Plaintiffs’ counsel should raise them repeatedly in writing with defense counsel, even if defendant seems completely uninterested. If nothing else, this will help defeat defendant’s arguments at the end of the case that plaintiffs did more work than was necessary.

Zubulake v. UBS Warburg LLC, 217 F.R.D. 309, 91 FEP Cases 1574 (S.D. N.Y. 2003), a Title VII sex discrimination case, ordered the defendant to produce in discovery e-mails that had been deleted and that now resided only on 94 back-up tapes, at its own expense, estimated to be about $175,000. The court noted that requiring plaintiff to pay the expenses of such discovery could often make the discovery inaccessible regardless of its relevance. It stated that the production of electronic discovery can be cheaper than the production of paper discovery. Finally, the court announced a new seven-factor test to determine who should pay for the discovery of electronic records:

Set forth below is a new seven-factor test based on the modifications to Rowe discussed in the preceding sections. 1. The extent to which the request is specifically tailored to discover relevant information; 2. The availability of such information from other sources; 3. The total cost of production, compared to the amount in controversy; 4. The total cost of production, compared to the resources available to each party;5. The relative ability of each party to control costs and its incentive to do so;6. The importance of the issues at stake in the litigation; and7. The relative benefits to the parties of obtaining the information.2. The Seven Factors Should Not Be Weighted Equally

Whenever a court applies a multi-factor test, there is a temptation to treat the factors as a check-list, resolving the issue in favor of whichever column has the most checks. But “we do not just add up the factors.” When evaluating cost-shifting, the central question must be, does the request impose an “undue burden or expense” on the responding party? Put another way, “how important is the sought-after evidence in comparison to the cost of production?” The seven-factor test articulated above provide some guidance in answering this question, but the test cannot be mechanically applied at the risk of losing sight of its purpose.

Id. at 322–23 (footnotes omitted).

Wiginton v. CB Richard Ellis, Inc., 2004 WL 1895122, 94 FEP Cases 627 (N.D. Ill. Aug. 10, 2004), modified the Zubulake factors by adding a “marginal utility” factor and granted in part plaintiffs’ motion to require defendant to pay the costs of restoring data defendant had rendered inaccessible, by requiring that defendant pay 25% of the costs pendente lite. Toshiba America Electronic Components, Inc. v. Superior Court, 124 Cal.App.4th 762, 21 Cal.Rptr.3d 532

(Cal. App. 6th Dist. 2004), granted mandamus and remanded a discovery order that would have required the producing party to pay all of the potentially $1.9 million to produce requested electronic discovery, without any examination of the reasonable costs of such discovery to be borne by the requesting party.

I. The Role of the Court in Managing Electronic Discovery

As with all other subjects of discovery, judges are essential when one side cannot be reasoned with. The problem occurs when the judge makes a mistake in identifying the unreasonable party.

For example, some hearings on electronic discovery involve defense contentions that the discovery in question will cost millions of dollars, which makes the plaintiff look unreasonable. If the discovery could be provided cheaply without redactions, and if the millions are necessary only to satisfy the defendant’s desire to strike all fields that are not immediately clearly relevant, it’s actually the defendant that is being unreasonable.

For example, some hearings on electronic discovery involve defense contentions that the discovery in question will cost millions of dollars, which makes the plaintiff look unreasonable. If the discovery could be provided cheaply without redactions, and if the millions are necessary only to satisfy the defendant’s desire to strike all fields that are not immediately clearly relevant, it’s actually the defendant that is being unreasonable.

In these situations, the appointment of an independent expert to advise the court may be extremely beneficial.

Where both sides are reasonable, my experience is that judges will happily leave the technical aspects of electronic discovery for counsel to work out together, once the court is satisfied that the parties know what they are doing in this area. Conversely, a court that has to immerse itself in technical details to resolve a myriad of objections and counter-objections is not a happy court.

Sometimes, judges will not have had personal experience with the kinds of problems attendant on this evidence, and will inadvertently make rulings that will prejudice one side or the other. This is one more reason it is useful to work out problems with the other side in advance of the hearing, or to narrow the scope of what is to be decided.

J. The Use of Computer Files at Trial

Exhibits should be simple enough that their meaning should be readily apparent to an ordinary person without technical training.

The problem for both sides is that, even with experts who know how to make themselves understood to persons not technically trained, their exhibits can be hard to understand. When they are easy to understand, they may be misleading.

Ideally, each analytic exhibit should identify its source, any specific subset of employees on which the analysis was done, the time period covered, the geographic scope of the analysis, any exclusions of data, and anything else helping to make apparent exactly what was done. These are not the normal conventions in the expert’s field, but are critical. In addition, the presence of this information on the exhibit alerts the counsel putting on the witness to potential problems not yet fully appreciated.

“Dictionary” exhibits—big bulky printouts that will never be read cover to cover but that are available to provide additional information on employees if desired—can be useful not just at trial, but in ensuing appeals and remedy proceedings. They can answer questions that first occur to someone long after the trial.

Lewis v. Bloomsburg Mills, 773 F.2d 561 (4th Cir. 1985).

But see Lowery v. Circuit City Stores, Inc., 158 F.3d 742, 765 (4th Cir. 1998), vacated and remanded on another point, 527 U.S. 1031 (1999) (denying defendant’s argument that it was deprived of effective defense at trial because lower court refused to order production of data base compiled by non-testifying expert consultants, which testifying expert used to produce trial exhibits).

The destructive commands of so-called “retention” policies have been well documented in recent corporate scandals. Nor is this phenomenon limited to private companies or to the United States. In December 2004, the British government announced a new “retention” policy that would require the destruction of massive amounts of information just before a new Freedom of Information law is to go into effect.

EEOC regulations require that records be preserved during the pendency of EEOC charges and any ensuing litigation, as well as to show the presence or absence of disparate impact. 29 C.F.R. §§ 1602.14 and 1607.4. Unjustified failure to comply should give rise to adverse inferences. EEOC v. American National Bank, 652 F.2d 1176, 1195 (4th Cir. 1981); Lewis v. Bloomsburg Mills, 773 F.2d 561 at 568. Contra: Anderson v. Douglas & Lomason Co., Inc., 26 F.3d 1277, 1287 n.13 (5th Cir. 1994), cert. denied513 U.S. 1149 (1995).

E.g., Lipsett v. Blanco, 975 F.2d 934, 939, 59 FEP Cases 1498 (1st Cir. 1992) (rejecting defendants’ attack on plaintiffs’ counsel’s time in a fee award, and stating: “This case was bitterly contested. Appellants mounted a Stalingrad defense, resisting Lipsett at every turn and forcing her to win her hard-earned victory from rock to rock and from tree to tree. Since a litigant’s staffing needs often vary in direct proportion to the ferocity of her adversaries’ handling of the case, this factor weighs heavily in the balance.”).