How electronic discovery can go wrong in practice and ways to avoid problems

1. E-Discovery: How to Do it Wrong (and Right!)
Avoiding Sanctions through Competent Management of the e-Discovery Process
By: Gene ~ l b e r t '
Presented at the 22quot;d Annual Technology Conference
University of Texas School of Law CLE, Four Seasons Hotel, Austin, TX, May 22,2009
quot;[O]ne or more of the retained lawyers chose not to look in the correct locations for the correct docunzents, to accept
the unsubstantiated assurances of an inzportant client that its search was sufficient, to ignore the warning signs that the
document search and production were inadequate. . . These choices enabled Qualcomnz to witlzhold hundreds oJ
thousands ofpages of relevant discovery and to assert numerous false and nzisleading arguments to tlze court andjuty.
This conduct wat,rarzts the inzposition of sanctions. 'quot;
I. Introduction
The number of sanctions in e-Discovery has increased in recent years as judges have
become more knowledgeable with the responsibilities of e-Discovery and the new federal
rules and comfortable holding lawyers and their clients before them accountable. E-
Discovery has become an increasing important part of many cases because so much of
what is important in business and personal interactions is now recorded
contemporaneously in emails and other electronic documents, aka electronic stored
information (ESI). The continued exponential increase in the amount of electronic
documents inevitably means more e-Discovery to review each year for any particular
type of case. Effective use of new technology and tools is essential to be able to review
and analyze potentially relevant documents in an efficient manner.
Sanctions involving e-Discovery are levied for a variety of reasons, including
misrepresenting ESI, falsely certifying discovery, late discovery and document
destruction. But many sanctions are applied not because attorneys involved did not know
the law, or care about the case, or even spend substantial effort and time trying to make a
legally adequate production. Instead, sanctions often occur because attorneys and their
clients simply do not understand the scope of needed discovery in a case and do not apply
proper management techniques to address their discovery responsibilities.
This paper will address some ways that attorneys get into trouble in e-Discovery and
discuss approaches, tools and techniques to avoid problems that can lead to sanctions.
I Gene Albert (JD, Southern Methodist University; MBA, University of Texas at Austin) is a founder of
Lexbe.com, which offers a web-based litigation management application, as well as other e-Discovery
services. He can be contacted at gene@Lexbe.com.
2
Qualcomm Inc. v. Broadcom Corp., No. 0 5 ~ ~ 1 9 5 8 - B
(BLM) (S.D.CA Jan. 7,2008), vacated in
part, No. 05CV1958-RMB BLM) (S.D. CA. Mar. 5,2008). Qualcomm Inc. v. Broadcom Corp., No.
05cv 1958-RMB (BLM), 2008 WL 66932 (S.D. Cal. Jan. 7,2008).
E-Discovery: How to do it Wrong (and Rlght!) Gene Albert, Lexbe.com
22quot;d Annual Technology Conference, University of Texas at Austln Page 1 of 25

2. 11. The 6 Million Dollar Subpoena
Before getting to suggestions as to how to do e-Discovery right, this paper will first look
at an example of how it can go very wrong. In re Fannie Mae Securities ~ i t i a a t i o n ~
upheld a contempt order against a non-party for failing to comply with a stipulated
discovery order. The Office of Federal Housing Enterprise Oversight (OFHEO) is a
government agency that supervised the Federal National Mortgage Association (quot;Fannie
Maequot;). In 2003, OFHEO investigated Fannie Mae's accounting and financial practices
and found that Fannie Mae quot;had departed from generally accepted accounting principles
in order to manipulate its reported earnings and inflate executive compensation.quot;
OFHEO's report prompted several private civil actions against Fannie Mae and its
executives, which were consolidated into multidistrict litigation in the United States
District Court for the District of Columbia.
During discovery three individual defendants who were senior executives at Fannie Mae
subpoenaed OFHEO, a non-party to the litigation, seeking records OFHEO had collected
during its investigation. On November 6, 2006, the district court denied OFHE07s
motion to quash, and directed it to comply during the following four months with the
subpoena. OFHEO requested and received two one-month extensions, and in the summer
of 2007 reported to the district court that it had produced all of the requested documents
in the production. The individual defendants thought that more documents existed, and in
a designated representative deposition OFHEO disclosed that it had failed to search for
off-site disaster recovery backup tapes.
The individual defendants then moved to hold OFHEO in contempt, and at the hearing,
OFHEO and the individual defendants quot;entered into a stipulated order that held the
contempt motions in abeyance and required OFHEO to conduct searches of its disaster-
recovery backup tapes and provide all responsive documents and privilege logs by
January 4,2008.quot; The stipulated order further stated that quot;OFHEO will work with the
[individual defendants] to provide the necessary information (without individual
document review) to develop appropriate search terms. By October 19,2007, the
Jindividual defendants] will specify the search terms to be used.quot; (emphasis added)
The individual defendants then specified over 400 search terms, which eventually
resulted in approximately 660,000 responsive documents to the keyword searches (80%
of total emails). OFHEO objected but the district court found that the stipulated order
quot;gave the [individual defendants] sole discretion to specify search terms and imposed no
limits on permissible terms.quot;
To try to comply, OFHEO then hired fifty contract attorneys for document and privilege
review, spending over $6 million on the discovery request, which was quot;more than 9% of
the agency's entire annual budget.quot; On November 29,2007, the day before an interim
deadline for producing various categories of documents, OFHEO requested another
extension until December 2 1, quot;assuring the district court that it could meet that extended
deadline.quot; The district court granted that extension as well, but on December 19, 2007,
' No. 08-5014 (D.C.D.C 2009).
E-Discovery: How to do i t Wrong (and Right!) Gene Albert, Lexbe.com
22ndAnnual Technology Conference, University of Texas at Austin Page 2 of 25

3. OFHEO told the district court that its prior assurances were quot;based on insufficient dataquot;,
and it could produce all non-privileged documents by the January 4,2008 deadline, but it
could not produce the required privilege logs until two months later.
The individual defendants then renewed their motions to hold OFHEO in contempt. The
district court so held and characterized OFHEO's efforts quot;too little too late,quot; and found
that quot;OFHEO ha[d] treated its court-ordered deadlines as movable goal posts and ha[d]
repeatedly miscalculated the efforts required for compliance and sought thereafter to
move them.quot; As a sanction, the court ordered production of all privileged documents
immediately on an attorney's eyes only basis.
OFHEO appealed the contempt finding and the sanction to the U.S. Court of Appeals for
the District of Columbia. While recognizing the quot;extensive effortsquot; made by OFHEO to
comply with the order, the Court of Appeals concluded that OFHEO had not established
a reasonable basis for concluding that the district court abused its discretion. The Court
found itself quot;ill-positioned to second-guessquot; a district court's assessment of whether strict
enforcement of its deadline was necessary. Noting that an abuse of discretion in such a
case quot;would risk undermining the authority of district courts to enforce the deadlines they
impose.quot;
Clearly the result of OFHEO spending so much for a 3rdparty subpoena is shocking, but
the review standard for abuse of discretion is high. Here are several things that OFHEO
did or did not do (with 20-20 hindsight) that may have contributed to the poor outcome
from OFHEO's perspective:
OFHEO abandoned objections under FRCP 45 to cost shifting and narrowing of
scope by agreeing to the stipulated order.
There is no evidence that OFHEO sought to establish that the backup tapes were
not reasonably accessible.
OFHEO did not offer expert testimony as to the unreasonableness of search terms
or of the discovery request in general.
Counsel for OFHEO agreed to an open-ended search term procedure without
apparently understanding the implications or potential costs.
In re Fannie Mae Securities Litigation illustrates the problems resulting from litigants and
attorneys not understanding early on in the case the amount of ESI at issue. This is
evidenced by the fact that OFHEO missed at least three discovery deadlines, and at the
end of 1 % years after the initial subpoena was served did not know how many
documents were responsive and had not completed privilege review. This in spite of the
fact it had spent over $6 million and had over 50 lawyers working on the request.
E-Discovery: How t o do ~t
Wrong (and Right!) Gene Albert, Lexbe.com
22quot;* Annual Technology Conference, University of Texas at Austin Page 3 of 25

4. Current Issue: In re Fannte Mae Securrtcrs Ldt~~&m:
E??
Select Facts:.A' .ll?ri ..........--
. . . .. . .. ..
.. . . , .-
Totd Facts: 14
Dl$rict court rules for the ddendarjks and &&& C % O t
%E o
durn-rg the mxt f ~ months,
w
OFEO requests that e-Dtscovefyhe bged to file5 m
OFHEO mr1w; several extenam5 Frem the c w t
OFHEi)reparts to thecaurt that & h a d p a d . a % h m
OFEO c d ~ t mn Rule 3U(bX&j depo thal B had F&
s tci wh &a#
6 mflifw
tts off-sikedaaster-recovery b& tapes
responstve Jwumer~k.and wwhqe
If*3fmjantrto specify search terms,
D&enMs 5pecIf.y W wrch tam
In re Fannie Mae Securities Litigation offers a couple of take-aways. From a legal
perspective it may say no more than that if you misrepresent yourself repeatedly to
the trial court, miss a lot of deadlines, and then agree to something inadvisable in a
stipulated order, good luck getting the trial court or an appellate court to let you off
the hook.
But from a practical perspective it provides a powerful warning:
If attorneys do not get a handle on e-Discovery and ESI issues early in the case, or do
not manage e-Discovery processes well, the consequences can be disastrous to the
client and the attorneys.
E-Discovery: How to do it Wrong (and R~ght!) Gene Albert, Lexbe.com
22ndAnnual Technology Conference, University of Texas at Austin Page 4 of 25

5. 111. e-Discovery Growth Drivers and Processes
Growth of ESI Drives e-Discovery
There has been much
discussion of the FRCP More E = More e-Discovery
e-Discovery
amendments in recent
years, but these merely
Digital Information Created, Captured, Replicated Worldw~de
reflect the new reality
of the amount and
importance of ESI.
The increasing amount
of ESI has been the
Online Storag~
primary driver of e- Digital Camera
Discovery growth. Facebook Skyp~
Datacenter Application
Both the types of Satellitelmages Scanner
available data storage Messaging Vtdeoconferenc~n~
media and amount of
storage available
I n Ctrstomer 5erbrice Record~ng
ptups MetbooksServers torens~cs Saa
continue to expand
unabated under Moore's
law.
It would be natural to think that the growth is about to level off, but nothing could be
further from the truth. Predictions from equipment manufacturers suggest that the
amount of digital information will continue to grow at an exponential rate for as long as
we can imagine.
Plus, digital storage is increasingly where the action is. In particular, email of
participants in a commercial litigation matter captures the contemporaneous facts of what
happened and why in a way that was unimaginable back in the days of typewriters and
phones. An attorney can hardly begin to understand what actually happened in a modern
commercial case without a thorough analysis of email and other ESI.
Limited Timeline in e-Discovery
Under the FRCP there is a fairly limited time span between when a complaint is served
and counsel needs to be prepared to attend the meet and confer meeting and the
scheduling conference with the judge. During the initial days of the case counsel for both
sides must identify sources of data repository, assess their contents and accessibility and
be prepared to 'meet and confer' and enter into discovery agreements. This can be a lot
to do in a short period of time for cases that involve much ESI.
E-Discovery: How t o do ~t
Wrong (and R ~ g h t l ) Gene Albert, Lexbe.com
22ndAnnual Technology Conference, Unrverslty of Texas at Aust~n Page 5 of 25

6. Meet and Confer
1
SmpeDataandttesams Mnr 3?iadlor Meet and ConA?r
Understand the ESI Process
e-Discovery production has a lifecycle that must be understood to be managed. E-
Discovery sanctions usually involve failure to deal with one of the major lifecycle
components properly. It can be helpful to break the e-Discovery process into the
following steps as defined by the Electronic Discovery Reference Model (EDRM)
Project, a leading industry working group:
Information management
refers to the evaluation of
how records management
programs impact electronic
records and e-Discovery.
Identfication refers to the
process of learning the
location of all data which
is subject to the duty to preserve and potentially disclose in a pending or prospective legal
proceeding.
The duty to preserve electronically stored information (or any potentially relevant
evidence) arises when a party reasonably foresees that the information may be relevant to
pending or anticipated litigation.
Duta collection is the acquisition of electronic information (data) marked as potentially
relevant as part of identification. Electronic information should be collected in a manner
that is comprehensive, maintains its content integrity and preserves its form.
E-Discovery: How to do ~t
Wrong (and Right!) Gene Albert, Lexbe.com
22ndAnnual Technology Conference, University of Texas at Austin Page 6 of 25

7. Electronic discovery processing is designed to accommodate a wide variety of
unstructured data, handle each form in a manner appropriate to its file type, and generate
output that is structured in accordance with review requirements that often vary with law
firm practices, client needs and review technology provider specifications.
Document review is used to identify responsive documents to produce and privileged
documents to withhold. Improvements in data storage, database and search technology,
and online review functionality are providing increasingly efficient options for handling
the volume of data and streamlining the review process.
Analysis is the process of evaluating a collection of electronic discovery materials to
determine relevant summary information, such as key topics of the case, important
people, specific vocabulary and jargon (which can vary significantly between and even
within companies), and important individual documents. Analysis is performed
throughout the remainder of the process as new information is uncovered and issues of
the case evolve. Good technology and techniques are essential to effective analysis.
Production refers to the delivery of responsive, non-privileged documents to opposing
counsel. Rule 26(f) sets an expectation that the method and format by which electronic
data are to be produced should be considered and negotiated by the parties early in the
discovery process.
Source: EDRM project4
I V . How to Do it Right
The remainder of this paper will address the e-Discovery process and tools and methods
to allow a production to be handled competently and efficiently. This requires both an
understanding of the process, skills to manage it properly and tools to be efficient.
Accept (Even if you can't Embrace) the New Discovery Obligation from Zubulake
In the leading e-Discovery case ~ u b u l a k eJudge Scheindlin outlined a list of
,~
responsibilities that counsel would be well-advised to follow when e-Discovery is
involved:
As soon as litigation is reasonably expected, attorneys should issue a litigation
hold directing all employees to preserve documents and data that may be relevant
to the litigation.
Once a hold is issued, counsel should determine all sources of potentially relevant
information and speak with the company's IT department about its document
http://www.edrm.net
Zubulake IV, 220 FRD 2 16 (SDNY 2003)
E-Discovery: How t o do i t Wrong (and Right!) Gene Albert, Lexbe.com
22quot;d Annual Technology Conference, University of Texas at Austin Page 7 of 25

8. retention policies, data retention architecture, backup procedures and how
recycling policies are actually implemented.
Lawyers should go further and interview all of the key employees involved in the
case to determine how they store electronic data such as emails and evaluate their
compliance with company document retention policies.
Employees should be instructed to produce electronic copies of their relevant
active files and ensure that all backup data that the company is required to retain
is safely stored to avoid the risk that backup tapes will be recycled.
Lawyers should take affirmative steps to monitor compliance with the litigation
hold and periodically re-issue the hold so that new employees are aware of it and
that it remains fresh in the minds of other employees.
These responsibilities assume a very interactive and involved role of the attroney with his
or her client and its employees and agents. It also assumes that the attorney has the skill
and resources to manage the e-Discovery process, as well as the attention and cooperation
of potentially many diverse employees of the client with control over electronic data.
Related to the Zubulake duty is FRCP 26(g), which requires an attorney to sign each
discovery disclosure, request, response or objection and to quot;certifly] that to the best of the
person's knowledge, information, and belief formed after a reasonable inquiry, the
disclosure is complete and correct.quot; The duty to make a quot;reasonable inquiryquot; is satisfied
when the investigation undertaken by the attorney leads to conclusions that are
reasonable under the circumstances. It is an objective standard similar to Rule 11. The
requirement of a certification based on reasonable inquiry is not new, but the recent
increase in discovery ESI has renewed focus on FRCP 26(g).
So when an attorney responds to a discovery request in error because some ESI was
undiscovered, he or she faces possible sanction exposure if there was not a reasonable
inquiry into the truthfulness of the filing under FRCP 26(g).
Identifying, preserving and collecting ESI can be a sizable project; one with an initially
unknown scope. Despite this, attorneys work under set deadlines to complete it.
Repeatedly missing deadlines and asking for extensions can materially undermine the
court's confidence in counsel.
E-Discovery: How to do i t Wrong (and Right!) Gene Albert, Lexbe.com
22ndAnnual Technology Conference, University of Texas at Austin Page 8 of 25

9. Developing Your Case and the Data Early; Early Case Assessment
Of course it has always been a good idea to get an essential grasp of the strengths and
weaknesses of one's case early as early as possible. However, this becomes more
important and more challenging with a case that involves substantial ESI.
Assessing one's case early is more important now because keyword selection used for
discovery requests and Rule 26 Meet and Confer agreements depends on accurate
identification of the important case issues. If the critical issues of a case are not
identified until later, then it will become necessary to request additional documents based
on new keywords, or from different custodians. This adds cost to discovery collections
and tries the patience of opponents and potentially the trial court. At some point in the
discovery process the cost of additional collections will shift to the requestor, or it simply
will not be possible to get for issues indentified too late in the process. A related
important issue involves identifying potential claims and defenses early enough in the
discovery process to be able to include them in the pleadings.
Assessing one's case is also more difficult with ESI intensive cases because getting
information about the case early from the data can be expensive and is dependent on
knowing the issues to generate keywords: a 'catch-22'. If one is trying to get the
important facts early in the case from the ESI (as opposed to witness interviews and case
participants, etc.) then the review of ESI should be an iterative process of trying proposed
keyword searches, manually reviewing a sample of the documents returned by the search
results, and then revising search terms.
The process of getting a handle on a case early is sometimes called 'early case
assessment'. As part of the early case assessment process, counsel will identify what
discovery is needed, identify key witnesses and data custodians, construct factual
timelines and determine the strengths and weaknesses of the case. This assessment is
done in the first 60-120 days of a case and communicated in various forms with the
client, opposing counsel and the judge. The goal of this proactive approach is to develop
an early strategy for the case, realistically evaluate the case, and determine the cost and
budget to prosecute or defend, and identify business practices that might be modified to
minimize litigation exposure in the future.
This sounds like 'mom and apple pie' and is not a new idea6. Why isn't this done on every
case? There are several real life barriers to doing a comprehensive early case assessment.
One is the need for client support. Doing an early case assessment will cost money and a
client may view it as 'over-working' a case if the client does not understand the process
and benefits. These benefits include a better understanding of the case early that can lead
to earlier opportunities for settlement and reducing unnecessary e-Discovery costs.
A more subtle objection is the concern with being the bearer of bad news. A legitimate
worry is that the client may not be ready to hear bad news about the case and may be
6
See. ex., Eric L. Barnum, A n Introduction To Early Case Assessment, The Practical Litigator, Nov. 2006,
at21.
E-Discovery: How to do ~t
Wrong (and Rlghtl) Gene Albert, Lexbe.com
22quot;* Annual Technology Conference, Unlverslty of Texas at Austln Page 9 of 25

10. inclined to 'shoot the messenger'. Some clients think they need a very gung ho 'jailhouse
dog' type of litigator and might interpret an early balanced analysis of case strengths and
weaknesses as lack of confidence by their attorney in their cause. This problem is
compounded when the litigator is recently hired and does not have experience with the
client. The client at this point has not yet developed confidence in the judgment of the
lawyer and this may make it harder for the lawyer to communicate bad news.
For these types of clients some lawyers believe it best to let the adverse facts of their case
sink in to the client over time. Depositions are particularly helpful in focusing a client on
the potential or actual weaknesses of their story. Eventually the client begins to accept
the shortcomings in their case, but not so much directly from their attorney or too soon,
and are eventually more inclined to settle. This is why so many cases settle on the
courthouse steps.
This approach can be unnecessarily expensive as clients eventually become ready to
settle a case on the courthouse steps for an outcome that might easily have resolved the
matter early on with much less cost and pain to the client. For this reason, the naYve
approach of not being able to hear bad news about one's case early is more common with
smaller companies, and executives and individuals who do not have enough experience in
the vagaries of litigation.
The cost of the 'wait and see' approach of not doing early case assessment is greater
today, and will continue to grow, because of the increase of ESI. Some cases may have
so much potentially relevant ESI that the cost to review it may outweigh the benefits of
litigating. In reality a case in which the opposition has enough merit in their case to
allow extensive e-Discovery justifies an early evaluation of that cost and raises the case's
settlement value. Also, not understanding one's case early can greatly increase the cost of
conducting e-Discovery by requiring repeated requests and reviews as key issues and
keywords are not understood, and custodians and data repositories not identified, until
relatively late in the case.
'Early case assessment' is also used in a different sense by some vendors offering e-
Discovery services and products. Here they are instead referring to tools to quickly
gather, deduplicate and search ESI as a first pass to provide and early analysis of the
amount and content of ESI. This can be important too in some cases, but is a different
idea than the process of interviewing witnesses and constructing timelines, etc., discussed
here as 'early case assessment'.
What types of things might one do as part of an early case assessment? One key step is
to identify all the key case participants. These include the parties, executives, material
witnesses, potential deponents, subjects of potential third part subpoenas, data custodians
of ESI, etc. Case participants can be gathered as a list and discussed with the client and
individuals familiar with the underlying facts of the dispute.
E-Discovery: How to do it Wrong (and Right!) Gene Albert, Lexbe.com
22ndAnnual Technology Conference, University of Texas at Austin Page 10 of 25

11. Rented bu~ld~ng
l e t Hobt frum
and 7120 Plne Forest Poad
Pensacda FL 32520.5923
m
x
Phone
Fax
745-792-0324
745-732-0325
Defendant 1
Phone: 850-873-6821
&& 'Ihippe~
Prnpert~es,
Inc. Owned 3nd 2510 H Pace Blvd
Fa*:
rnamtalnedJet Ho~st Pensacula, FL 32505 Defendant
Cell:
sut~ect acodent
of
Emall:
Phone: 850-433-5330
Hoist H~miJlmg 2lJ0 E Gregurl St Fa*:
Repared jet hact Fensacula, FL 32501
6 1 Jones
11 Phone: 480-123-5487
Witness to acclder,t 807 W Garden St, Pensacola, FL Fax: 480-123-5488
&$ Pres~dent
F h Cell: 480-123-5489 witness
Started a year agn
Emall: b~ll,iones@vahoo.corn
Developing this list is an iterative process, as clients and other persons familiar with the
case can be shown drafts of the list and it will jog their memories about other persons or
companies who should be included. quot;I don't see Judy F. on the list. She worked as
contractor for us two years ago and did a lot of work on that project. She keeps her
computer at home and will have a lot of files there.quot; Finding potential data custodians
early through this process can avoid problems and increased costs down the road.
Another early case assessment tool is to construct timelines at the outset. Certainly all
lawyers construct timelines eventually as a case progresses towards trial, but there are
advantages in doing it very early as part of an early case assessment. When doing
chronologies very early in the case, they are of course constructed from limited sources
available at that time: facts from initial client interviews, review of key documents
forwarded by the client, and alleged facts from the opponents pleadings. However, even
these limited sources can begin to give a good picture of where the strengths and
weaknesses of the case lie.
Current Issue: 05 franl~rs : L X ~ S
5s
/ I &.st the *&era
?he 4 . d went h e r&?r thdntn the b pa
r t] 1'i'Zquot;4 ?&a Fr&-g&&x yB'= -*dhd
/ W W e Cmwterny W o L!ld the b k k r w w d t e r ths?dc&W
Fr g;i2g M F a * a & & W
/ tm Kmt to U h w k e l t b L u
- b s ah* th* acclden IYtW we F t y ~ L~ & s ~
s & s I8 L a #X*
!L+
Early case timelines can also include facts that are needed to support the claim or defense
based on anticipated jury instructions for the claims involved. These facts remain
'orphaned' until associated with one or more documents or portions of deposition
testimony. The existence of these orphaned facts serve as red flags for evidence that
needs to be developed during the discovery stage. Similarly tracking facts that one's
opponent needs to prove for his or her case serves as focal points for developing
E-D~scovery:How to do it Wrong (and Right!) Gene Albert, Lexbe.com
22quot;dAnnual Technology Conference, Un~versity Texas a t Austln
of Page 1 of 25
1

12. opposing evidence and summary judgment motions for evidence that has not been
successfully developed during discovery.
Timelines also serve as a good way to begin educating clients about the potential holes or
weaknesses in their case. A client can begin to digest a truer picture of his or her case
through periodic review of timelines showing positive and negative facts backed up with
documentary evidence or deposition testimony. Such chronologies can also help the
client to develop supporting evidence, both documentary and oral, by jogging the client's
memory for missing or contrary evidence. As the case develops and more evidence
comes in, the issue timelines can be supplemented more easily. With functional case
timelines, an attorney is also in a better position to quickly develop and defend sumillary
judgment motions and to identify specific facts backed with evidence.
Map the Data, Identify the Custodians
At the Rule 26 Meet and Confer, the attorneys for the parties are expected to understand
their own IT infrastructure, how much potentially responsive ESI they might have, where
it is located and what custodians have control over it. The allowed 120 days (before any
extensions) may not be sufficient for many e-Discovery projects. A legal hold can only
be effective to custodians who are identified and receive it.
Many e-Discovery sanctions involve attorneys not knowing where their client's ESI is
until too late in the process. At best this involves embarrassment and loss of credibility
with opposition and the court. But failure to identify ESI early on can also lead to
spoliation as ESI is regularly destroyed in the usual course of business. So if ESI is not
identified promptly in the discovery process and preserved, it is possible that it will be
irretrievably lost. This is what happened in the notorious Qualcomm Inc. v. Broadcom
Corp case, in which a federal judge admonished Qualcomm and its counsel for failing
to discovery and turn over email evidence. Thousands of key emails were discovered
only after trial in what the district court referred to as a quot;monumental discovery blunderquot;.
At issue were emails and other electronic documents suggesting that, contrary to
Qualcomm's claims, that Qualcomm had participated the development of an industry standard
which would lead to infringement of Qualcomm's patents.
The court ordered Qualcomm to pay $8.5 million for Broadcom's litigation costs and
referred counsel for possible state bar sanctions. The court identified quot;numerous warning
flagsquot; that should have lead to discovery and production of the withheld documents,
including:
Broadcom's production of an email reflector list supporting that email regarding
standard setting was sent to Qualcomm representatives early on during the standard
setting process.
Broadcom's impeachment of multiple Qualcomm Rule 30(b)(6) witnesses showing
Qualcomm's involvement in early stages of the standard setting activity.
7
Qualcomm Inc. v. Broadcom Corp., supra.
E-Discovery: How t o do it Wrong (and Right!) Gene Albert, Lexbe.com
22quot;* Annual Technology Conference, University of Texas at Austin Page 12 of 25

13. Qualcomm's eve of trial discovery of nearly two dozen emails showing Qualcornm's
involvement early in the standard drafting process.
The court also noted Qualcomm's specific failures to act, including failure to search the
computers of its Rule 30(b)(6) witnesses for responsive email and other documents; failure to
use search terms that could have been derived from the materials produced by
Broadcom; and failure to otherwise access and review the computers, databases and emails of
its employees and consultant involved in the standard-setting process.
Failing to find data repositories can be disastrous. How does one go about finding them?
When an outside attorney first approaches the IT department of a client asking for a data
map, he or she is likely to get a 'network map', which is a graphical representation of
computers and other devices and networks. IT professionals maintain these maps as part
of their jobs, but they can sometimes get out-of-date due to the highly dynamic nature of
IT systems and the rush to do more pressing work. . .
This level of information,
however, hardly begins to I
address what is needed
for e-Discovery. First, it I
usually refers to
computers and network
nodes only, and not to
backup tapes, software as
a service, remote storage,
laptops and personal
devices, etc. Second, it
does not specifically
identify custodians,
retention policies and
other information needed
to assess the importance
of any particular data store.
E-Discovery: How t o do it Wrong (and Right!) Gene Albert, Lexbe.com
22ndAnnual Technology Conference, University of Texas at Austin Page 13 of 25

14. When the outside attorney asks for more information, the next stop may be the
accounting department. Companies routinely keep asset depreciation schedules of all
fixed assets and these records will give hints at to potential data repositories.
But these records are insufficient for e-Discovery purposes as well. Accounting records
seldom if ever will provide the kind of detailed use information needed. Also, companies
may carry assets on the books that
are not really in use for e-
Discovery purposes. Moreover,
asset schedules refer to fixed assets
only, and will typically ignore
non-capitalized leased equipment,
rented equipment, hosted services
and personal assets under the
Table PZ:
companies influence or control 7 L l s t of flxed .issets w r t l ~ l a s s e s
c D e p r e c ~ a t ~ ~ nl e s
o
(e.g., home computers, laptops, Cla?~ Deacrlpt~or cust
phones, personal data managers,
11 Equipment
etc.) 12 lEqu~pment
Equ~pment
Equ~pment
#23
#33
6.31 2 10
6,509 97
13 ,Equ~pment Equipment #26 6,562 33
1 4 Equ~pment Equlpmerd # I 0 7,070 14
IT and accounting records kept in 15 Equ~pment Equ~pment #34 7.619 63
the usual course of business are
really only the starting point for
building the type of database
needed to manage e-Discovery. Instead, counsel needs to construct a content map
designed to include the type of information needed to properly manage e-Discovery.
E-Discovery: How t o do i t Wrong (and Right!) Gene Albert, Lexbe.com
22quot;* Annual Technology Conference, University of Texas at Austin Page 14 of 25

15. - - --- -- - - - -
Repos~tory
Name .-a ~ n - O f ~ ~--- ----Server --.
~ Exchange - --
--
c e -- ID ?1€Q-i?136 7
--- --
vj Locatron
?-
--
Ma~n
L-
Offlce -
-_I_II
7
Content Descr~pt~on
-- -- -- - -- - - - - -- -- - - --
Exchange ernall server. MS Exchange 2007. Appr, 20 GB of Bus~ness
Owner
emad on server (more In local PST fdes on cl~ent
CPUs).
- - -
Governance Pol~c~es
- -- - - - - - - - When D~scont~nued
Users generally l~rn~ted 200MB on server, but excepbons
to
allowed on case-by-case baas. Local PSTs allowed
L - -- -
-- -
- - - --- I Not Reasonably Accessible
r- - - - - - - --- --- -- --- -
Record Last Updated = I -
- - ;
Dally full backup, wTh tapes retamed For 30 days, Monthly
backup tape reta~ned 6 month.
For
An ESI Content Map is a report of an organization's electronic data sources for use as a
litigation management tool. It will include the names of custodians, the type of
information included, backup policy, location and other information helpful in
determining if a repository is reasonably accessible for legal purposes. It will not include
irrelevant IT structure or accounting information. Data sources include e-mail archives,
network shares, document management systems, financial systems, proprietary databases,
external databases, etc. The map will include understandable information that counsel
needs to competently identify, preserve and collect ESI required for discovery related to
litigation or a governmental investigation.
Additionally, a well-maintained ESI content map can help show compliance with the safe
harbor in FRCP Rule 37 for a party that demonstrate with quot;good faithquot; that its ESI
practices are sound. This can help counsel demonstrate that a company has been
proactive regarding its discovery obligations.
E-Discovery: H o w to d o i t W r o n g ( a n d Rtght!) G e n e Albert, Lexbe.com
22ndAnnual Technology Conference, University o f Texas a t A u s t ~ n Page 15 o f 25

16. Backup tapes deserve special mention as they are often the subject of sanction cases.
Typically in sanction cases attorneys will represent to the court that all ESI in a case has
been found and then be embarrassed later as more backup tapes appear. How does this
happen?
Some cases involve intentional malfeasance. But more often the problem may simply be
one of lack of communication. One common problem is the lack of accurate knowledge
among senior management as to how backups are actually done in their company. The
cost of backup tapes is Q: Now many backup tapes do you maintain?
cheap, but the depends on & c you ask
3!
consequences of not
Oeponent Answer
having one when
CI0 quot;None. We back upeverything online now.quot;
needed are great. So
quot;Wealso still follow a poltcy of keeplng monthly backup tapes going
operators who make Director of IT back three months and 10 daily backup tapes that are rotated dally
backup tapes can and wer-writethe ones that are 10 daysold. 13 total.quot;
sometimes retain extras.
It is one thing to have IT Supervisor
quot;We keep daily tapes going back 30 daysand six months of monthly,
an IT procedure (e.g., to and there are 2 tapes for each backup, so about 70.quot;
rotate and overwrite
backup tapes) and it is haveone. Most old backup tapes are in our offsite warehouse.
another to have it Night-shift
There must be over a thousand there going back over 10 years. I
IT Operatot
followed consistently. think Fred and Bill have some at home too. Maybe others too, you'd
Few senior managers
(who most often
communicate with outside lawyers) and inside counsel actually are involved with the
details of how backup tapes are actually made. It is important, therefore, as part of
building a content map to identify and interview IT staff down to the operator level to
find all potential data repository early.
E-Discovery: How to do it Wrong (and Rlght!) Gene Albert, Lexbe.com
22ndAnnual Technology Conference, Universtty of Texas at Austin Page 16 of 25

17. Figure Out How Much You Have to Review
An important part of early case assessment is to determine how much potentially
discoverable ESI exists, how much is reasonably assessable, and the volume of page
equivalents to deal with once a review begins. Attorney review costs are primarily page
and document driven. A surprising number of page equivalents can exist on drives and
storage media like CDs, DVDs and flash drives.
A distinction needs to be made between raw ESI before and after culling and
deduplication. As part of I Pages in a MBIGB e-Discovery Calculator
these processes7 'ystem 71-11,sfree O I I I I I ~ ~ca~l:Liritul III~IJ~;~ e g a i ~ ~ ~ t e s , rjt~vali~u:3 vpes p g , t:~l~ttc~nk
ctr ~>igrib>yte,s file e!,la~~
files and other clearly F S i e , Word, b c e i and other rlative flies, P C J F ~TIFF^, etc ) estimate- ~qutualent
, pages and b.:zse?
irrelevant files are
removed, duplicate files
logged and only one copy
111s
1:tl~dyct':.
- I
raiiuiritor 1 5 ~ ~ s e ffijrl i;iiculatitlg e-cuscoveru' riocun~erit
I u revlev.; rosts a n d preparing Ilt~gatlon
of duplicate files left in
the collection for review.
Culling and deduplication
can reduce data stores by
20%-70% depending on
how data are collected
and other factors.
TI$%
Even at an early stage
before e-Discovery
processing, an estimate of
culling and deduplication Grantf Total
reduction in data size can [Compute] *=,
g&&{%-: , <
--y-%k-+ac .
be made and a page
http //www lexbe corn/hplPages-Megabyte-Gigabyte aspx
estimate generated.
Different file types convert into pages at very different rates, so it is not possible to get a
very accurate estimate of page equivalents by taking the overall size of data stores (in
MB or GB) and converting based on a single conversion factor (e.g., pages1GB). There is
just too much variation between different data collections. Instead, an estimate based on
the relative prevalence of different file types in the collection is more accurate. These
early page estimates are critical in determining what kind of staffing will be needed to
review docun~ents, how long the review will take and how much discovery review will
cost.
Consider a Quick ESI Analysis
Collection can be done in several ways and the method of collection may vary with the
type of case. In very routine cases, identified custodians can self-collect files and
forward to counsel. Independent collection from custodians is becoming more common
-
as it avoids the problem of a custodian intentionally or inadvertently under-disclosing
responsive EST. The IT department can collect email repository on the server, file shares,
E-Discovery: How t o do it Wrong (and Right!) Gene Albert, Lexbe.com
22quot;dAnnual Technology Conference, University of Texas at Austin Page 17 of 25

18. backup tapes, and other shared user data sources. More contentious cases may call for
collection by forensic experts from certain custodians. Forensic collection can preserve
all of the system metadata and forensic investigators can certify to the chain of custody of
the evidence in case of challenges to its authenticity.
Sometimes actual drive images are made of the computers of key custodians. One case
may involve ESI gathered from by under different methodologies, and as a result may be
a hodgepodge drive images, email stores, individual files, etc.
An increasingly popular technique is to do a quick analysis of what is in an ESI
collection. As mentioned above, it is critical to understand early on in an e-Discovery
project how much work must be done to be able to properly staff and construct accurate
project timelines. Clients also require budget estimates and need to be able to make
review decisions, and determine whether and how to proceed with litigation, with overall
costs in mind.
Quick ESI Analysis
Amount of Data in Analysis Review Statistics Potential Custodians Metadata Extr.
S~ze(GB) 51 3 Est Paaes Name #Flies %
No of F ~ l e s 3.728 Emall 492,480 29% Br~an Budw~n 349 13%
Images 89,775 5% Glor~a McM~nn 237 11%
File Type Distribution PDFs 112 860 7% Shannon Norberg 252 9%
No % S~ze[GB] % Presentat~ons 20,520 1% Dav~d Bahler 214 8%
Emall 1.274 34% 12 312 24% Spreadsheets 76,372 4% Ann Forster 182 7%
PDFs 528 14% 5643 11% Text 513,000 30% Ranjeev S~ng 155 6%
Sytern Flies 515 14% 11 799 23% Word Process~ng 184,680 11% Chad Huston 132 5%
Other 336 9% 3078 6% Other F ~ l e
Types 230.850 13% Reese Range 112 4%
Tert 295 8% 2 052 4% Total Pages 1 720,537 Allson Kneeland 95 4%
Word Processing 271 7 % 3078 6% Kr~sten Vale 81 3%
Audlo 217 6 % 6 156 12% PagestDocument 8 Rob Hernandez 69 3%
Images 183 5 % 3531 7% Est Documents 2,745,967 Mason Gross 58 2%
Presentat~ons 54 1% 1 026 2% Brlan McKn~ght 50 2%
Spreadsheets 37 1% 0 509 1% Review Time D~ane Dav~s 42 2%
V~deo 18
- 0% 2052 4% Doc Rev /Hr Man Hours Ann Volgt 36 1%
3.728 51 296 25 109,839 Joshua Szymczak 30 1%
50 54,919 Fred Klng 26 1%
75 36613 Lei Sun 22 1%
100 27,460 Berry Shelton 19
- 1 %
2.684 83%
Raw or partially processed ESI can be analyzed relatively inexpensively (compared with
full ESI processing) to obtain the following information early in the process as a 'Quick
ESI Analysisquot;:
What is the approximate distribution of file types in the data collection?
What is the approximate distribution of file dates in the collection? Discovery
productions are often limited by file creation dates and this can show how the data
collected will be reduced by date limitations.
What persons are identified as authors in the files (from email and file metadata)?
This shows if the collection identifies custodians that have not previously been
E-D~scovery:How t o do it Wrong (and R~ght!) Gene Albert, Lexbe.com
22quot;* Annual Technology Conference, Unlverstty of Texas at Aust~n Page 18 of 25

19. identified, of if previously suggests custodians are represented in the data
collected to date.
How many pages are likely to be converted from the data collected?
Construct Good Keyword Searches
We are in the middle of a paradigm shift in electronic discovery from mainly manual
review to mainly electronic review. Prior to e-Discovery, attorneys believed that they
should manually review every document (if not every page) of a document review set,
and anything short of that standard was deficient. As electronic data repositories have
exploded in size this approach has become untenable in an increasing number of cases
that involve so much ESI so as to be uneconomical (or even impossible) for attorneys to
try to manually review all potentially responsive documents. As a result attorneys have
become increasingly reliant on keyword searches to identify potentially responsive
documents.
But a study conducted the 1980s casts doubt on the ability of litigation teams to
accurately determine a set of comprehensive search terms that will return all or even most
potentially relevant documents. Attorneys and paralegals involved in a subway accident
case used a keyword methodology to search a discovery database consisting of 350,000
pages in 40,000 documents. The litigation team believed that they had located at least
75% of the relevant documents. A separate manual review of documents was conducted
and found that the litigation team had only identified 20% of potentially discoverable
documents through keyword searching alone.' Many of the missed documents included
vocabulary that the litigation team had not anticipated. Participants working for the
subway in internal communications referred to quot;the unfortunate incident,quot; while victims
referred to it as the quot;disasterquot;. Relevant documents also included oblique references to
the quot;event,quot; quot;incident,quot; quot;situation,quot; quot;problem,quot; and quot;difficulty.quot; Many documents were
missed that might have been identified with a thorough manual search or more inclusive
searching approaches, like concept searching. This study illustrates that keyword
searches should be designed with these types of potential problems in mind.
While keyword searching alone can miss many documents, so can manual review. As
anyone who has done document reviews knows, it is grueling work and concentration
lags after even a short period of time. Plus different reviewers will apply disparate
standards even working from a well-defined protocol and will produce inconsistent
results.
The influential Sedona Conference supports the conclusion that manual review alone is
not the answer. In Best Practices Commentary on the Use of Search and Information
Retrieval Methods in E-Discovery (August 2007),~ authors note that while most
the
Blair & Maron, An Evuluation of Retrieval Eflecliveness for u Full-Text Document Retrieval Sj~stem,
28
Com. A.C.M. 289 ( I 985) (publication o f the Association for Computing Machinery).
9
E-Discovery: How to do ~t
Wrong (and R~ght!) Gene Albert, Lexbe.com
22quot;* Annual Technology Conference, Unlverslty of Texas at Austln Page 19 of 25

20. lawyers consider quot;manual review [to be] the gold standard by which all searches should
be measured,quot; that in fact a manual review methodology inevitably results in its own
errors of missed documents. quot;Human review of documents in discovery is expensive,
time consuming, and error-prone.quot; Instead, a mixed approach is best with construction of
superior search keywords through an iterative methodology of keyword searches on a
representative sample and then manual review of returned documents from the sample to
develop an expanded list of keywords.
As keyword searches become more important, not surprisingly we are seeing opposing
parties construct very different sets of proposed search terms. We saw this in the
Fannie Mae Securities Litigation case, discussed above, and it is also the subject of
William A. Gross Constr. Assocs., Inc. v. American Mfrs. Mut. Ins. CO.,quot; in which a
U.S. magistrate judge recently issued a self-styled quot;wake-up callquot; to lawyers designing
keywords without adequate information quot;by the seat of their pants.quot; This case involved
claims concerning alleged defects and delays in the construction of the Bronx County
Hall of Justice. The Dormitory Authority of the State of New York (quot;DASNY quot;) directed
the project. At the time of the discovery dispute, non-party Hill International served as
DASNY's construction manager and agreed to produce Hill's DASNY-related ESI to the
other parties in the action. DASNY proposed a handful of search terms including
quot;DASNY,quot; quot;Dormitory Authority,quot; quot;Authorityquot;, quot;Hall of Justicequot; and quot;Bronx but not
Zooquot;. The other parties requested thousands of additional search terms, including
quot;sidewalk,quot; quot;delay,quot; quot;budget,quot; quot;elevator,quot; quot;claimquot;, etc.
The magistrate judge complained that he was in the quot;uncomfortable positionquot; of having
to construct a search term methodology without sufficient input from the parties or the
relevant custodian. He ruled that in addition to DASNY's proposed terms, the search
should incorporate the names of the parties' personnel involved in the courthouse project,
but rejected a much expanded keyword list without more justification. He criticized the
parties search term methodology as quot;just the latest example of lawyers designing
keyword searches in the dark.quot;
All keyword searches are not of like quality. Judge Paul Grimm recently wrote quot;while it
is universally acknowledged that keyword searches are useful tools for search and
retrieval of ESI, all keyword searches are not created equal; and there is a growing body
of literature that highlights the risks associated with conducting an unreliable or
inadequate keyword search or relying exclusively on such searches for privilege
review.quot; ''
Keywords are often negotiated very early in the case. But as the case evolves new issues
and keywords inevitably will be identified. As this will involve more expense for the
responding party, going back and negotiating more keywords may be difficult. Before
allowing this without consent a court may want to know why proposed keyword
additions were not identified earlier. This supports the importance of the iterative process
'O No. 07 Civ. 10639,2009 WL 724954 (S.D.N.Y. March 19,2009).
'I Victor Stanley, Inc. v. Creative Pipe, Inc., 250 F.R.D. at 25 1, 256-57(D. Md. 2008).
E-Discovery: How to do it Wrong (and Right!) Gene Albert, Lexbe.com
22ndAnnual Technology Conference, University of Texas at Austin Page 20 of 25

21. of sampling and an early case assessment to understand the case issues and potential case
keywords as early as possible.
Proper keyword construction is harder than it looks. Many lawyers underestimate the
complexities involved because they think it is just like running searches in LexisTMor
WestlawTM, which they have been doing as long as they have been practicing. But legal
research searches are much different in character because they address only structured
text databases designed to provide for easy and accurate results. Searching unstructured
e-Discovery repositories is a much different task and involves many more potential
dificulties.
There is never a 'perfect' set of keywords, and keyword construction will always involve
a trade-off between being overly-inclusive (false positive or Type I error) and being
overly-exclusive (false negative or Type I1 error). Keywords can have a number of other
problems, including:
Keywords created by brainstorming only, not sampling and testing
Failure to identify jargon of an industry or internal vocabulary of a company
Failing to account for misspellings and abbreviations, particularly from phones
and PDAs
General search terms that bring up too many irrelevant documents
Some now suggest that Boolean keyword searches be supplemented with other
information retrieval methodologies, like grouping and concept search12. This can
increase search results, but may increase the false negative rate as well. Also, these
approaches are often proprietary and not as transparent as simple Boolean keyword
searching.
Courts are beginning to call into question whether attorneys are competent to
develop keyword searches in the first place: quot;[Flor lawyers and judges to dare opine that
a certain search term or terms would be more likely to produce information than the
terms that were used is truly to go where angels fear to tread. This topic is clearly beyond
the ken of a layman and requires that any such conclusion be based on evidence that, for
example, meets the criteria of Rule 702 of the Federal Rules of ~ v i d e n c e . quot; 'This
~
thinking suggests we may soon see more use of experts in the area of keyword validation
and the resulting Daubert challenges as part of cases involving disputed e-Discovery
methodology.
12 http://www.lexbe.com/hp/lititech-concept-searching-litigation.aspx
I' United States v. O'Keefe, 537 F. Supp. 2d 14,24 (D.D.C. 2008).
E-Discovery: How t o do it Wrong (and Rlght!) Gene Albert, Lexbe.com
22ndAnnual Technology Conference, University of Texas at Austin Page 21 of 25

22. How can a party go about doing better than just brainstorming keyword terms? A good
approach is to sample and test returns for different proposed search terms. This provides
a quantitative approach to estimate how many documents will be returned with different
a
Quick Keyword Analysis
Plaintiffs Keywords Defendant's Keywords
Searc h Tern-, Docs Returned quot;10 Search Term Docs Returned %
access 247 796 DASNY (company name) 202 5%
alarrn 42 1% D~rn~dbry AuIhonty B 1%
budget 6!32 19% Authority AN0 party names 139 4%
burld ?,&4 78% Cal~rtf AND Bronx 94 3%
chanya order @32 24% Hall of Juottce 21 1 %
clam 529 14% 'F3rorbx but not Zoo 124 3%
delay 243 7%
drnm~af 107 3%
ele$ator 35 1%
elecIr~cal 3 9 8%
s~dewalk 1 d 5%
9
bPtB THClilSANDS MORE
iI
I
Total Search Return 3.~04 94%
I Total Search Return 561 16% I
Arnount o f Data in Analysis
Size (C43) 51 3
N o of F~les 3.728
keywords. The next step is to take a sample of returned disputed keywords to determine
through manual review the percentage of relevant documents brought back with the
search. An approach like this helps to give magistrates and judges a better basis to
include or exclude disputed search terms.
Decide on the Best Review Format
There is a lot of debate now as to the best format of ESI to request in discovery. The
choices are to convert to PDF or TIFF or keep the documents in native format for review.
The decision depends in large part on capacity of the document review software to be
used and how the process is managed.
TIFF is the older Review Format Options
standard and is still I 1 Advantages Disadvantages
most used in large I I >Busmess standard
Searchable 1
productions. A I >Can be searchable >Redact~on tr~ckler
IS
disadvantage is that , >Can Bates stamp on page
TIFF files are raster
I Image
I >Best for some software
>Redact~on error
less
1 >NOOCH in document
images (made up of 1 TIFF
bawd prone /
1 zRequ~res
speclal software to read
wfth searchable text
pixels like a ' >Can Bates-stamp on page ,
photograph). The
commonly used TIFF 1 e.g , Word,
>Includes metadata >Requires original applicat~on
or
formats are not Native 1 1 >Some files don't convert
1
web viewing software
1 to PDF/TIFF well >Cannot Bates stamp at page level
searchable. This is
j >No conversion charge
handled by creating a
E-Discovery: How t o do i t Wrong (and Right!) Gene Albert, Lexbe.com
22ndAnnual Technology Conference, University of Texas at Austin Page 22 of 25

23. separate text file for each TlFF file using OCR (optical character recognition). Special
TIFF-based litigation support software associates the image TIFF and matching text file
for searching and viewing. But these documents are not very portable and anyone who
wants to work with the files must have the required viewing software, usually older
litigation support software systems first developed before PDF became commonplace in
litigation.
PDF files are becoming more popular for e-Discovery productions. PDFs have the
advantage of retaining the raster image but also can embed searchable text in the file,
known as 'text under image PDF' or 'searchable PDF'. The text in this type of PDF can be
either transferred from a text-based document as part of the conversion process to PDF
(e.g., from a Word document), or created by using OCR with scanned documents. PDFs
are portable in that they can be transferred and viewed in a variety of contexts without
special software, other than the free Adobe AcrobatTM viewer or competitive PDF
viewers.
Native review refers to conducting a document review in the original format maintained
by the custodian. While still a minority approach, it is gaining proponents.'4 The chief
advantage of native review is avoiding the processing cost of converting ESI to PDF or
TIFF. One disadvantage is that the document review may be more time-consuming and
difficult than if the documents were converted first to PDF or TIFF. This problem has
been lessened in recent years with the newer generation of litigation document review
software that is designed to work well with documents in native format. These programs
allow for automatic conversion of common native file types into viewable HTML
renditions that can be reviewed without needing to open the reviewed documents one at a
time in the applicable native application.
Even if the decision to convert ESI to PDF or TIFF is made, certain document types do
not convert well and should also be available in native format. Excel XLS is the most
common example of a file type that often does not convert well for document review.
Possible problems include pages not formatted for printing, the need to see formulas as
well as values, filtered data, etc. But the problem is not limited to spreadsheets and many
other file types may need intermediate processing to view in PDF or TIFF or not be
viewable at all except in the applicable native application.
Special Considerations for Email in Document Review
As discussed above, attorneys are relying more on text searching and less on manual
review. This is especially true for email, which often makes up a majority of ESI, and
may comprise the most important EST in business disputes. Litigants often cull down
email repository to a more manageable size through keyword searches. When doing
productions that rely on text searching to identify responsive and privileged documents, it
may be important to ensure that the entire data set is searchable. Searching in native
format (no conversion to PDF or TlFF plus associated text files) works well with the
14
Ralph Losey, Should You Go Native?, http:llralphlosey.wordpress.com/2009/02/17/should-you-go-
native/
E-Discovery: How t o d o it Wrong (and Right!) Gene Albert, Lexbe.com
22quot;d Annual Technology Conference, University of Texas at Austin Page 23 of 25

24. body of an email as it is text-based and easily searchable with a variety of tools. Email
attachments, however, are more complicated. It is not uncommon for 10% or more of
attachment files in native email collections to be non-searchable for a variety of reasons.
If these attachments are not converted to a searchable format they will be missed in a
keyword search.
Reasons why email attachments may not be searchable include:
Scanned files and faxes included as attachments
Image based PDFs, JPEGs
Compressed files (ZIP, RAR)
Password protected files
Unrecognized file types, other non-searchable native files
This is a knotty problem that deserves some consideration at the outset of a case before e-
Discovery commitments are made. There are three choices as to how to handle email
repositories in native format (e.g., Outlook PST or MSG files):
Review all email and attached documents manually for responsiveness and
privilege. This can be very expensive in attorney time and still miss documents
that would be found in a keyword search.
Convert all emails and attachments to a searchable format (searchable PDF or
TIFF plus associated text files) with OCR and use keyword searching. The
attachments can be joined to the original email or saved as separate files and
associated with the original email. Conversion to PDF or TIFF with OCR will
provide the best result from a searchability perspective but introduces a higher
processing cost.
Keyword search in native format without converting email attachments to a fully
searchable file type with OCR. This is less expensive than the other choices, but
may miss a substantial percentage of responsive or privileged documents in the
form of non-searchable attachments.
In any case the approach should be discussed and agreed upon early on in the process so
that the attorney is not criticized later for missing discoverable evidence.
Conduct the Review
Once ESI has been collected, culled and processed, it is available for attorney review for
substance and privilege. For cases involving large collections of electronic data, the
attorney will want to use a well-designed review tool to efficiently work through the
document collection. The review process involves looking at documents responsive to
searches, reviewing for key issues and responsiveness, and identifying or confirming
determinations of privilege and confidentiality.
E-Discovery: How t o do ~t
Wrong (and Right!) Gene Albert, Lexbe.com
2Znd Annual Technology Conference, Un~versity Texas at Austin
of Page 24 of 25

25. 1 / I F * 125% - d, EJ9
_-
I_
vsx
X
DoFaapb7 OStdnddrd
--
LNVOICE Ortxt lzGZizl
*+
Wemqd0rldl.RwNext
22.
Dr.cc.111kk.r 2004
,hiquot;, lkwatenng
124; SW Orange A*r.nuc
a+
@ Dot Disc - -
Nstcs Fa?
5e-e % :
?- n-
o ,2~& p ~ t
Plantatton Counlr. BeFlora oiscmay
429-430-q605
Ropz/17na)9
~ ~ ~ p a n u v e b
I Kumber ,5565
D Claw- & e
d w
om- Wd.Robrt
LFJconirdentral
-quot;-----------i f //
~ Dcseription - ---f-quot;=--1 3~u.m~~
*E
,
~ ~ ~ - ~ ~ p t B c e h o oiisqemhly l- 3~ton jei ------
n . b k hoist I
L --.-
L- -
-] ]
{
c
-
' f ~ ] f b r e i g h~t a t 1-rj T* I~~~~ HOL.~ &
l?-'?-'W r't two arhn~chlans ] i I we. ~ 2 ,
zow
,[12-22-2004 1; .Manltfter -1-
1 '
DDLSOVLe
m ar
hr
-L
R Green
M
1: l DUE' I
TOTAL , S 1219 00
szE zef;&&*;;
TW mv 2
ew
n
I -.- - -
Chsrvtathr* W t q R m O e a l
Please remit thc [oral amoiint due wthm 50 day$
*- E -- --
c*c-F=+4z --____.-.
1 hank Sou. =-w* --
.- ---
--F=wq L -. -- - .-
CugmmC4id5
AtW-vc- .--..-.-_______
Conduct the Privilege Review and Prepare the Privilege Log
In the In re Fannie Mae Securities Litigation case, privilege review consumed most of the
$6 Million spent. It is not unusual that privilege review and log generation will constitute
a large portion of a discovery budget. While claw-back agreements are available under
FRCP 26(b), attorneys are rarely comfortable with letting a discovery production out the
door without some additional attorney searching or manual review for privileged
materials. The consequences of disclosing privileged materials may be too great, even if
protected from inadvertent waiver. Computerized review tools help here by making the
review more efficient and then automatically generating the privilege log.
V. Conclusion
Failure to understand and maintain control of e-Discovery processes can lead to a number
of bad consequences. At minimum, the attorney can be embarrassed and lose credibility
with his or her client and the court. But shortcomings can also lead to sanctions as even
inadvertent e-Discovery mistakes can have serious negative ramifications as they may
violate e-Discovery related representations and make it appear (even if not true) that the
attorney and/or client has something to hide. E-Discovery certainly can be managed
properly, but it requires a solid understanding of the underlying processes and technology
and diligent project management throughout the case.
This paper is ava~lable Lexbe.com
at
to
http:l/w.lexbe.com/hd litilawhostedleDiswverv-How Do it Wrong (and Riahthdf
E-Discovery: How to do it Wrong (and Right!) Gene Albert, Lexbe.com
22ndAnnual Technology Conference, University of Texas at Austin Page 25 of 25