There are no Magic Words for Taxation of Processing Costs, but there is the Law

Taxation of cost cases can be dull. However, when the Judge shoots down an argument with a YouTube video in the opinion, you know it’s a party.

The prevailing party in a trademark infringement case between companies that make portable electronic fitness tracking devices sought $88,888.86 in costs and the clerk awarded $54,089.15. The Court ultimately granted $63,660.94 in costs. Fitbug Ltd. v. Fitbit, Inc., 2015 U.S. Dist. LEXIS 62879, *2 (N.D. Cal. May 13, 2015).

The opposing party made a procedural argument against awarding costs based on the fact the declaration filed in support of the Bill of Costs with statutory citations did not include the three words “allowable by law.” The Court rejected this argument, stating:

… the requirement a party say the “three little words,” “allowable by law,” is merely a reminder that the Court expects them to submit costs they believe are taxable, not a set of magic words necessary to receive any costs. Cf. Sarah Vaughan, Three Little Words, on Live at the London House (Mercury Records 1958), available at:

Fitbug Ltd., at *4-5.

With that, Judge Samuel Conti simultaneously invoked one of the greatest jazz singers of the 20th Century and reminded us that legal arguments do not come out of book of spells. Well played. Fitbug Ltd., at *6.

Taxation of eDiscovery Costs

The Court noted that Section 1920 was enacted in 1853 and did not speak on electronic discovery costs (largely because “new” technology at that time included typewriters and telegraphs). Fitbug Ltd., at *6.

Judge Conti summarized that taxation for “exemplification and the costs of making copies of any materials where the copies are necessarily obtained for use in the case . . .” applies to some electronic discovery expenses. Fitbug Ltd., at *7.

Granting Data Extraction and Processing Costs

The Court allowed $32,282.05 in processing and data extraction costs to be taxed. Fitbug Ltd., at *11. There are many cases where processing costs are denied, so this is a significant award.

The supporting declarations explained that the costs related “to collection, scanning and conversion of documents, and related processes necessary to the eDiscovery process,” and these entries specifically “concern document collection, including scanning and related processes.” Fitbug Ltd., at *8. In a supplemental declaration, the prevailing party explained the costs were necessary to prepare discovery documents for production to the opposing party in the format to which the parties agreed to produce ESI. Fitbug Ltd., at *8-9.

The Court held the processing expenses were properly taxed for complying with the parties’ ESI agreement. Furthermore, the Court also allowed the costs for extracting metadata, because “these too are necessarily incurred, allowable exemplification costs because they were incurred not for the convenience of counsel, but to comply with the parties’ agreement.” Fitbug Ltd., at *10-11.

The Court also held that the cost of converting document formats as required per the parties’ agreement were also taxable. The Court explained the costs were “expressly contemplated by the parties’ agreement and are necessarily incurred exemplification costs.” Fitbug Ltd., at *11-12.

Bow Tie Thoughts

ESI Protocols determined at the Rule 26(f) Conference and codified in a Court Order at the Rule 16(b) Hearing matter. This case highlights that production formats agreed to by the parties can be taxable. However, if the parties had agreed that discovery costs were to be carried by the parties, this case could have turned out very differently.

A lawyer cannot cry “I Cast Ye Costs Out!” as if these were magic words that would prohibit a prevailing party from recovering agreed to costs for the production of ESI. However, a well prepared ESI Protocol that addresses costs could have that very effect.


WHOA! A Prevailing Party Recovered $57,873.61 in eDiscovery Costs!

thumb-456698_1280My God, is it true? Did a Prevailing Party recover virtually all of its eDiscovery costs?

The answer is yes, thanks to a case in Colorado.

United States District Judge Christine M. Arguello opened her order denying the Plaintiff’s motion to review the clerk’s taxation of costs with the following:

Because Defendants’ costs related to the electronically stored information (“ESI”) are expenses enumerated in 28 U.S.C. § 1920(4), and Plaintiffs were aware that Defendants would have to retain an outside consultant to retrieve and convert the ESI into a retrievable format, Plaintiffs’ Motion is denied.

Comprehensive Addiction Treatment Ctr. v. Leslea, 2015 U.S. Dist. LEXIS 17878, 1.

Rock on. Let’s review the Court’s reasoning.

The Plaintiff took the position that the Defendants’ eDiscovery cost award be reduced from $57,873.61 to $2,387.03, striking the work of a third-party eDiscovery service provider who performed the “retrieving, restoring, and converting data,” on the grounds the work did not constitute “copying.” Leslea, at *2.

The Court explained the Defendants hired their eDiscovery service provider to retrieve and restore ESI in order to respond to the Plaintiff’s Interrogatories and Requests for Productions. The requested discovery included “correspondence, summaries, emails, reports, and memos” relating to specific subject matter. Leslea, at *4-5. The Court noted that the work was complex and time-intensive, requiring three consecutive tolling agreements. Id.

The Defendants communicated with the Plaintiffs three times on the challenges over ESI, including providing detailed information on the scope of the data, archiving, and retention periods on multiple sources of data (hard drives, back-up tapes, etc). Leslea, at *5. In the second communication, the Defendants explained how the service provider restored 83 back-up tapes; and in the third the service provider’s forensic investigator detailed the difficulties in restoring the subject ESI. Id

The Court noted that the Plaintiffs were aware of the ESI challenges, did not recommend any changes to the scope of discovery, and even filed a new complaint with additional allegations. Leslea, at *5-6.


The Court held that the ESI expenses were “reasonably necessary for use in the case” and not done for the mere convenience of the parties. Leslea, at *6. The Court concluded the order as follows:

Indeed, Plaintiffs were aware of the monumental effort to retrieve and convert the data into a retrievable format in response to their Interrogatories and Requests for Production. The costs incurred by Defendants, the prevailing party, in responding to Plaintiffs’ requests are expenses that are shifted to Plaintiffs, the losing party. Indeed, Plaintiffs own litigation choices and aggressive course of discovery necessarily resulted in “heightened” defense costs. Plaintiffs have not demonstrated that these costs are improper. Accordingly, Defendants are entitled to recover their costs in full measure as determined by the Clerk, which it has identified as $57,873.61.

Leslea, at *6-7, citing In re Williams Sec. Litig-WCG Subclass, 558 F.3d at 1150.

Bow Tie Thoughts

Thank you Judge Arguello for understanding a simple truth: eDiscovery requires technology to retrieve information and translate it into reasonable useable forms that are necessary for the case. This technology and expertise costs money. Yes, this case had an expert who explained what was being done during the litigation. Not every case has such powerful facts explaining the why and how of restoring ESI to make it reasonably useable, but this is an epic victory for taxation of eDiscovery costs.

Always Document Services to Explain eDiscovery Services for Taxation of Costs

A prevailing party sought $61,548.65 in costs. They got $7,106.65. Kwan Software Eng’G v. Foray Techs., 2014 U.S. Dist. LEXIS 63933, 10-19 (N.D. Cal. May 8, 2014).

United States District Judge Susan Illston is very good at eDiscovery disputes. This case highlights the importance of not just knowing how to conduct eDiscovery services, but how to invoice for them.

The non-prevailing party challenged eight eDiscovery invoices on the grounds the prevailing party sought 1) eDiscovery costs that went beyond the costs associated with the actual production of the documents; and 2) failed to provide sufficient detail of its e-discovery costs to allow the Court to determine what are actual copying costs and what are non-taxable intellectual efforts. Kwan Software Eng’G, at *13.

Embed from Getty Images

The “Document Processing” invoice included fees for conversion from native files to TIFFs; “Bates stamping” each page (which is down by the processing engine, not someone actually stamping); and exporting the associated metadata. Kwan Software Eng’G, at *14. These fees are generally recoverable, but the Court stated that there was insufficient information to determine whether the costs were taxable. Id. Moreover, the prevailing party produced only 229,000 pages of discovery (assuming each TIFF is being counted as a page), but the invoices showed 344,445 pages had been processed (again, assuming TIFFs as pages and not files). Id.

The Court stated that the charges included documents that were not produced, thus not used in the litigation. As such, the Court only awarded costs of $6,870, which represented a charge of $0.03 per document for bates stamping and TIFF conversion for 229,000 documents. Kwan Software Eng’G, at *14-15.

The Court further denied costs for data hosting and management, because those costs are not recoverable. The Court also denied costs for production hard drive not actually produced to the opposing party (which otherwise would have been recoverable if actually produced). Kwan Software Eng’G, at *15-16.

Judge Illston also denied the producing party’s project management costs, because it was “unclear from the invoices whether the project management fees were actually related to document processing as opposed to other non-taxable activities.” Kwan Software Eng’G, at *16.

Bow Tie Thoughts

Judges cannot award costs to prevailing parties if there is not sufficient documentation to explain how the services were necessary for the litigation in producing discovery. This requires service providers to go into detail in their invoicing to explain each step of the workflow.

Consider the project management costs. Those could be recoverable, if there is sufficient information to explain how it was part of processing that went to actual copying, opposed to just intellectual efforts.

eDiscovery requires a specialized skill set and knowledge. Processing is not something done in a vacuum, but takes knowledge of how the application works and understanding what is needed in the case to be done effectively. A party cannot simply invoice “processed data for production” and hope a Judge will opine why those actions were necessary for the case. This takes documentation to educate the Judge on why the actions were done in order to produce the discovery in the litigation.

Hands-on eDiscovery: California Seminar on Responding to Discovery Requests

I had the good fortune to organize a seminar on responding to electronic discovery requests for the Santa Clara County Bar Association’s Civil Practice Committee on February 27, 2013. However, this seminar was different from other eDiscovery CLE’s, because the attendees spent a full hour conducting searches for responsive ESI to requests for production. The speakers included Santa Clara County Judge Socrates Manoukian (currently assigned to civil discovery), Tyler Atkinson of McManis Faulkner and Charlie Kaupp of Digital Statra.

Our seminar first focused for one hour on the California eDiscovery Act, California Rules of Court on eDiscovery, search and strategies for conducting document review.

Unfortunately, there is very little published California case law on eDiscovery. We have two main cases to explore, specifically Toshiba America Electronics Components v. Superior Court, 124 Cal. App. 4th 762, 764 (Cal. App. 6th Dist. 2004), which addressed mandatory cost-shifting for translation of back-up tapes into a reasonably useable form and Doppes v. Bentley Motors, Inc., 174 Cal. App. 4th 967 (Cal. App. 4th Dist. 2009), which does not address the eDiscovery Act, but unstated litigation hold issues and eDiscovery abuses resulting in an answer stricken and a default judgment entered.

However, at least one unpublished California opinion hints Courts want more than mere speculation that a discovery production was inadequate:

Following remand, Sukumar asked Nautilus to disclose its e-mails and all other electronically stored information concerning the Med-Fit order. After Nautilus responded that it had already disclosed all relevant documents, Sukumar filed a motion to compel. The trial court denied the motion, concluding that Nautilus’s response was sufficient and Sukumar “has offered only speculation that additional documents exist.” On appeal, Sukumar asserts that the trial court’s order denying his motion to compel should be reversed.

Sukumar v. Med-Fit Sys., 2012 Cal. App. Unpub. LEXIS 3309 (Cal. App. 4th Dist. May 2, 2012).

The appeal in the above case was denied, however it a glimpse at how California courts are handling production issues. Unfortunately, California discovery orders are not published and unpublished cases cannot be cited for any precedential value.  We literally have to watch for tentative orders to see how these issues are being presented to the courts to determine any trends. 

Review-TeamThe second hour required attendees to work as teams in a review database finding responsive discovery from the ENRON dataset, which was provided by Digital Strata in their InControl review platform.

Searching for responsive electronically stored information is a frequent topic at continuing legal education seminars, but infrequently a hands-on experience for attorneys. Our attendees were very engaged and diligently worked through the different hypothetical discovery requests.

We gave several case law examples of “bad” discovery requests, such as the following:

Produce any and all information related to email, including messages, from 1997 to 2006. 

Using the above as a reminder that production requests must be reasonably tailored to secure the production of documents relevant to the issues in a Federal lawsuit (See,Thompson v. Jiffy Lube Int’l, 2006 U.S. Dist. LEXIS 27837 (D. Kan. May 1, 2006)), we developed Requests for Production such as the following for the attendees:

Request for Production 3:

Any and all electronically stored information pertaining to the $56 million loss on Catalytica Energy Systems, sent between 12/01/2000 to 12/31/2001, in native file format, with the following extracted text or metadata:

From, To, CC, BCC, Date, Time, Subject, Document Author, Document Name, Custodian, Control Number, Folder (System File Pathway).

Request for Production 4: 

Any and all electronically stored information authored by Will Nolen, Sally Beck, Susie Ayala, Shona Wilson or David Port relating to project “jedi” sent between 1/01/2000 to 12/31/2001, in native file format with extracted text, substantive and embedded metadata.

One hour of conducting searches is only the beginning of how to respond to discovery requests. However, it is a very good first hour for attorneys who want to learn how to effectively search and respond to discovery requests.

Working with virtual screen

I would like to put together a future program focused on conducting privilege review, redaction, production and privilege log creation at a future seminar. I also think attorneys would benefit from a half to full day conference focusing on practical eDiscovery, such as issuing litigation holds, tracking hold compliance, document review strategies, developing search strings, testing different search tools (i.e., concept, complex Boolean, predictive coding), and production.

Conducting discovery is a skill. Like any skill, it is best to learn it by actually doing it. I believe our profession needs more hands-on eDiscovery events for attorneys to build their comfort level and confidence to competently represent their clients.

A Standard for Undue Burden: Excruciating, But Highly Educational and Useful, Detail

Discovery Balancing Acts in a War Zone

United States ex rel. McBride v. Halliburton Co., is a qui tam action over alleged fraudulent billing for services provided to the US military in Iraq.  The case involved inflated headcounts in Morale, Welfare and Recreation (“MWR”) facilities and fraudulent billing for those costs to the United States government.  United States ex rel. McBride v. Halliburton Co., 2011 U.S. Dist. LEXIS 6412, 1-2 (D.D.C. Jan. 24, 2011).

Balancing Search Terms and Undue Burden

The discovery disputes were numerous, and in his trademark detailed fashion, Magistrate Judge Facciola addressed each one in turn. 

One dispute centered on the collection and searching of additional custodians who were “cc-ed” on relevant emails was justified or unduly burdensome on the producing party. McBride, at *12-22. 

The Defendants initially searched 20 custodians for responsive electronically stored information.  McBride, at *12.  The Defendants voluntarily searched the data a second time with expanded keywords.  McBride, at *13-14. 

The Plaintiff demanded additional information, which promoted the Defendants to run another expanded search.  McBride, at *14-15.  The Defendants searched 230 custodians, which included all custodians in all litigation involving the Defendants.  McBride, at *15.  The uber-search of did not result any hits for the desired custodian the Plaintiff wanted in their request.  Id.  The Defendant took the position that since the emails were not found in the custodians the parties agreed to search; the Defendants did not need to conduct any more searches.  Id. 

The Plaintiff demanded the Defendants conduct additional searches of anyone who had been “cc-ed” on the already relevant email messages with the relevant reports.  McBride, at *15. 

The Defendants objected, claiming that search would add another 35 custodians.  Moreover, the Defendants estimated the search process would take months.  McBride, at *15-16. 

Judge Facciola stated a phrase that will cast a shadow of what must be demonstrated for an effective undue burden argument: “In excruciating, but highly educational and useful, detail.”

The Defendants’ manager for the e-Discovery group submitted a declaration that explained process of how data is collected from employees, which might include the following sources of ESI:

 (1) The employee’s hard disk drive, which can include the original disk drive, a “ghosted” image of the drive, and self-collected copies of the drive, of which [the Defendants] may possess none, one, or several collections for each identified custodian;

(2) Copies of current or former employees’ electronic data made at different points in time in response to various audits, inquiries, and litigation, which copies may include information no longer on the employees’ hard drive today (or was not on the hard drive at the time of his or her departure); and

(3) The employee’s mailbox or “homeshare” (personal network folder), with the possibility that there could be multiple existent copies of both the mailbox and homeshare if the employee was a custodian in other matters; indeed, there are multiple sources of such data, and it is not usual for more than a dozen to exist.

McBride, at *16-17. 

The Court explained that after the data was found, it was then copied in “a forensically appropriate manner to preserve its metadata and prevent its alteration.” McBride, at *17. 

The collection of data in the US involved data being sent to a service provider from storage; however, things would be a little trickier for the data in Iraq.  McBride, at *17. 

The overseas ESI would need to be shipped back to the United States or harvested over network connections.  Id.  The collection over a network would take days from just one custodian.  The Defendants’ e-Discovery manager estimated that each custodian had 15 to 20 gigabytes of data.  Collecting the data could take two to ten days for each custodian.  McBride, at *17-18. 

The workflow would then include processing the data and de-duplication for review.  McBride, at *18.   The data would then be searched for relevance and privileged.  Id.

Utility vs. Cost of Searches

The Court gave a very succinct summary of discovery cost analysis: 

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

McBride, at *18-19.   

The Court recognized a significant factor in favor of production: “Claims of fraud in providing services to military personnel raise important, vital issues of governmental supervision and public trust.” McBride, at *19.

However, the parties had spent $650,000 in discovery.  The “king’s ransom” in Judge Facciola’s words did not include attorneys’ fees. McBride, at *19.

The discovery already produced included two million paper documents, thousands of spreadsheets and half a million emails.  Id.

The Court denied the Plaintiff’s motion, because the Plaintiff had not demonstrated that the email messages not already produced were “critical to her proof.” McBride, at *19-20.  As such, the cost of the collection and search outweighed its utility. McBride, at *20.  In the words of the Court, “there is no showing whatsoever from what has been produced that those emails not produced will make the existence of some crucial fact more likely than not.”  Id.

The Federal Rules of Civil Procedure vs. Common Law Spoliation

The Plaintiff argued that the Defendants would violate a preservation order if they could not produce email messages from a specific custodian and should be sanctioned accordingly.  McBride, at *21.    The Court began its analysis with the following: 

It is important to keep separate, and not conflate, two distinct concepts: whether an additional search for certain information should be conducted, and whether or not the inability to find certain information should be sanctioned. The first is a question of interpreting the Federal Rules of Civil Procedure, while the second is a matter of the application of the common law principle of spoliation.

McBride, at *21.    

The Plaintiff’s sanction argument effectively would have a producing lose their right to argue against additional searches under Federal Rule of Civil Procedure Rule 26(b)(2)(C)(iii) if the producing party could not produce data they were asserted to preserve.  McBride, at *21-22.    

As the Court stated:

If that were true…then even an entirely innocent loss of information would deprive the party who occasioned the loss of reliance upon the balancing of the factors in Rule 26(b)(2)(C)(iii) to argue against an additional search. There is nothing in the Federal Rules that could possibly justify such an extraordinary sanction, and thereby rob Rule 26(b)(2)(C)(iii) of its intended universal application.

McBride, at *22.    

Bow Tie Thoughts

Proving undue burden requires specific documentation of the process used to collect data; where the data is location; time involved in collecting the ESI; and estimated cost.  While judges across the country will not be applying an “excruciating detail” test, parties need to demonstrate undue burden with evidence.

Load Files and Then Some…

Special Guest Blogger Pete Coons, VP of D4.

This is part two of a discussion that attempts to explain some of the often used and often misunderstood eDiscovery terms.

Last week I discussed DeNIST’ing and now we will tackle “Load Files”.  I am also going to squeeze in “Processing”.

Load Files

The Sedona Conference Glossary (great reference document) defines a Load File as:

“Load file:  A file that relates to a set of scanned images or electronically processed files, and indicates where individual pages or files belong together as documents, to include attachments, and where each document begins and ends. A load file may also contain data relevant to the individual documents, such as metadata, coded data, text, and the like. Load files must be obtained and provided in prearranged formats to ensure transfer of accurate and usable images and data.”

That pretty much covers it.  Now let’s take a step back and attempt to break down exactly where a Load File fits into a typical eDiscovery process.

Let’s say Company XYZ is being sued by a former employee for discrimination.  Company XYZ must now identify and preserve documents that may be relevant to the claim.   Data is identified and collected by a qualified individual within the organization or by a third party eDiscovery/Forensic service provider.  Typically, the collected data is then processed (another confusing term) so it can be placed into a database for review and eventual production to the opposing party. 


The Sedona Conference Glossary defines Processing as:

Image Processing: To capture an image or representation, usually from electronic data in native format, enter it in a computer system and process and manipulate it.”

Processing data usually involves ingesting the file or e-mail into eDiscovery software.  The software then catalogs the file and extracts all available text.  This text is usually placed into a separate text file and it’s associated with the native file or e-mail.  The processing process also extracts various metadata elements from the file and stores that information in a database. 

Let’s take a Word Document that contains the text “Hello World”.  After ingesting into the eDiscovery software a record is created in the database.  That record will contain metadata elements about the file such as: Author, Date Created, Date Modified, and Date Last Printed. 

As stated previously, the software will also create and store an accompanying text file that will have the text “Hello World” in addition to saving the original native file. 

We know that no case involves just one document so let’s pretend we have 10,000 Word documents and 10,000 e-mails.  It really doesn’t matter because the process is basically the same.  

This is a very simple explanation of the processing process and there are other steps that occur, like indexing or tiffing (imaging), but for all intents and purposes our data is now processed and it can now be prepared to be loaded into a review database.  This is where the Load File comes into play.

We have to get the data OUT so we can put it back IN somewhere.

Load files are usually simple text files (some are a bit more complex).  Meaning they can be opened and viewed with Notepad or WordPad in Windows.  To the uninitiated they may look daunting but after seeing a few they begin to look the same. 

Think of a load file as transport file that is used to facilitate the transfer of data and its associated metadata from one database to another.  We have to LOAD the data into one database from another. The load file contains information about each and every record (file) that was processed. 

That information can include the original file name, the author, date created, beginning ID number, the number of attachment that exist in an e-mail, the parent ID of the attachment, etc.  There are potentially dozens of metadata objects that can be provided in a load file (usually agreed upon by both parties prior to its creation).  The load file will also contain a link to the native file and the accompanying extracted text file. 

The load file along with the native file (or tiff) and its extracted text is the complete package for loading into another database. 

And that’s processing and load files!

The Plumbing of a Motion to Compel

In a product defect case about brass plumbing fittings, the Defendants fought a motion to compel electronically stored information relevant to class certification, because of undue burden and cost.  They lost…in large part. In re Zurn Pex Plumbing Prods. Liab. Litig., 2009 U.S. Dist. LEXIS 47636, 1 (D. Minn. June 5, 2009).

Paper Clogging the Drain

The Plumbing of a Motion to CompelThe Court initially required the parties at the beginning of the litigation to only focus on hard copy documents out of the concern the electronically stored information would drive up the cost.  2.  The Court stated:

“ESI may prove to be relevant to the first stage of discovery, we cannot meaningfully make that prediction now, and require the parties to engage in what could be vastly more expensive, and yet utterly futile, discovery.” In re Zurn Pex Plumbing Prods. Liab. Litig., 2.


The Court further stated on electronically stored information:

“[S]hould the parties uncover voids in the information disclosed in hard copy form, they are . . . at liberty to press for further discovery including electronically stored information.” In re Zurn Pex Plumbing Prods. Liab. Litig., 2.

The Defendants’ Electronically Stored Information

Deposition testimony revealed the Defendants’ electronically stored information includes employee email folders and files on the network system.  Additionally, the ESI of former employees’ were archived on DVD.  There was also a separate network drive that contained un-segregated folders from over 600 employees.  In re Zurn Pex Plumbing Prods. Liab. Litig., 2-3.

Enter the Litigation Hold and Search Terms

The Defendants produced a list of individuals who were issued a litigation hold. In re Zurn Pex Plumbing Prods. Liab. Litig., 3.

The Plaintiffs in turn requested searches across the different databases with 26 key words.

The Defendants opposed the request as being “not necessary for class certification and that the request was overly broad and would be extremely costly.”  In re Zurn Pex Plumbing Prods. Liab. Litig., 3. 

The Defendants’ Arguments Going Down the Drain

Arguments Down the Drain

The Defendants argued that 1) the first Court Order did not allow for any ESI productions and 2) producing ESI would be unduly burdensome and costly. 

The first Court order did NOT foreclose electronic discovery.  In re Zurn Pex Plumbing Prods. Liab. Litig., 4.  The Court stated the first Order allowed for electronic discovery if there were “voids in the information disclosed in hard copy form.” Id.  The Court noted email messages showed possible gaps over fittings failures and aggressive water compared to the paper documents.  Id. 

The purpose of civil discovery is to allow parties to “obtain the factual information needed to prepare a case for trial.” In re Zurn Pex Plumbing Prods. Liab. Litig., 5.  Despite the discovery being limited to class action certification, the Court stated there was a presumption in favor of production the Defendants’ electronically stored information.  In re Zurn Pex Plumbing Prods. Liab. Litig., 4-5.

Cost Counts

The Defendants were able to stop a complete rout with their cost arguments. 

The Defendants had approximately 61 gigabytes of data, which would have totaled around 27 million pages if printed.  The Defendants estimated searching all of the databases to cost around $1,150,000 and take seventeen weeks to perform.  That did not include collection and processing. In re Zurn Pex Plumbing Prods. Liab. Litig., 6-7.

The Court Order

The Court did not buy the Defendants’ burdensome arguments, especially considering a lawyer who was not an expert on searches made the claim. In re Zurn Pex Plumbing Prods. Liab. Litig., 6.

The Court did try to limit false hits from the search terms by limiting the number of search terms to 14 specific terms.  The Defendants’s searches were limited to the network drives, custodian emails and the DVDs of the former employees.  In re Zurn Pex Plumbing Prods. Liab. Litig., 6-7. 

Bow Tie Thoughts:

The parties would have been better served addressing electronically stored information at the beginning of the lawsuit, instead of delaying it for over a year.  Both sides would have been in a better position to judge class certification or settlement options if they knew the big picture.  Moreover, the Defendants had over a 600 person company.  Using email would be a fact of life.

The collection, processing, data reduction and review of 61GB would not be cheap, but $1,150,000.00 for the searches alone over 17 weeks sounds inflated. 

The perceived cost might be inflated by thinking of electronically stored information as the same as reviewing 27 million pieces of paper.  ESI can be reduced by targeted collections, pre-discovery tools to reduce the data set that are then fully searchable in a litigation support software.  With paper review, lawyers, contract attorneys and paralegals are digging through boxes of paper only searchable by the human eye. One can see where a seventeen week estimate comes from such a brut force search. 

Processing costs, which include data reduction to cull down electronically stored information, is approximately $600 to $1000 a gigabyte, depending on the vendor’s services.  Attorney review in a litigation support software is reduced by having to review less ESI, such as not reviewing any email newsletters, spam and focusing on specific date ranges.  With that said, estimating the cost to search 61GB that also did NOT include collection and processing to be over a million dollars seems high on the facts presented.

Discovery Production Workflow: Lessons from Magistrate Judges Facciola & Grimm

businesswomanclimbingIn my prior posting Playing with Fire: Producing ESI as Paper we looked at the dangers of producing ESI as paper.  In addition to the legal analysis, Judge Facciola outlined a workflow for the parties in Covad Communications Company v. Revonet, Inc. The Defendants represented to the Court that it would take a paralegal five to ten hours or work at a rate of $178.50 for a maximum cost of $1,785.00 to do a privilege review of the documents to be produced.  I think using litigation support technology would have expedited the process and taken less time, thus lowering costs.  Covad, 7.

 As part of the unfolding discovery production dispute, the Defendant offered to produce the emails in TIFF format, but insisted that the Plaintiff pay for a paralegal to “manually delete the non-responsive and privileged TIFF’s that have already been redacted from the paper document production.”  Covad, 7.  As stated before, this would take up to ten hours for a paralegal to complete. 

 The Court suggested the workflow entail the paralegal removing the paper documents that were irrelevant or privileged and then finding the corresponding “native” e-mails. The paralegal would then delete the irrelevant ones and move the privileged ones to another receptacle so they can be logged for a privilege log and made available for the Court’s review in case of a challenge.  Covad, 7.

 The workflow should not have focused on deleting TIFFs or Native Files, but reviewing the discovery production in a litigation support software. The reviewing paralegal could review the production and mark (tag) documents for production or privilege.  This could be done with a brut force document review literally reviewing each document. Additionally, running searches for responsive or privileged documents allows lawyers to reduce time reviewing irrelevant documents.

 The different litigation support solutions may vary in degree, but all allow for data entry to log a document as “responsive” and “privileged.”  This can be accomplished with data entry in a database field, or “tagging” features while may include checking a box such as “Attorney-Client Privilege.”  Many products also allow data to be organized in virtual folders the review can create, such as “Plaintiff Discovery Request 17.”   Production tools also vary, but most litigation support solutions allow users to “produce” ESI marked for production in an exportable format to their opposing counsel.  Conversely, many litigation support vendors are always happy to provide production services. 

 Using litigation support software would avoid the discussion of “deleting” non-responsive or privilege ESI.  Instead, the focus is marking documents for production.  ESI that is non-responsive can be marked as such.  Additionally, privileged ESI can be logged in the litigation support database.  Many litigation support solutions can now generate a privilege and redaction log.  This empowers reviewers to focus on analyzing the ESI and maximize the use of the litigation support solution. 

 There are many ways to push through a automated document review, but the two most basic options are 1) having an associate or paralegal sit in front of the computer and review the ESI like they were going through a box of paper 2) run searches for names, dates, email subject lines or other key words for responsive or privileged documents.

 Brute force ESI review can work, but might drive up costs in treating a large volume of ESI like paper. Developing a search protocol is imperative if your review is to have any nuance beyond a brut force review of your database.  However, lawyers and paralegals have to know how their litigation support solutions work in order to competently search for responsive ESI to production requests. 

 Peskoff v. Faber had multiple opinions by Judge Facciola over several discovery disputes, including a lawyer who performed searches for responsive email on hard drives and representing to the Court he had produced all responsive email.  The attorney was ordered to describe in detail the nature of each search for responsive email.  Judge Facciola later issued an evidentiary order in Peskoff v Faber, 2007 WL 530096 (D.D.C.) regarding whether the lawyer’s search did in fact produce all email messages:

 “I must insist that the person performing the search have the competence and skill to do so comprehensively. An evidentiary hearing will then be held, at which I expect the person who made the attestation to testify and explain how he or she conducted the search, his or her qualifications to conduct the search, and why I should find the search was adequate.”  Peskoff v Faber, 2007 WL 530096 (D.D.C.).

 Personally, I think Peskoff sounded more like a collection and processing situation, but the opinion is not clear what the attorney was using to search the hard drives for responsive email. 

 Developing quality assurance protocols to ensure responsive ESI is produced and privileges are protected is advisable for competent representation of a client.  Victor Stanley, Inc. v Creative Pipe, 2008 U.S. Dist. LEXIS 42025  is a case study in a law firm failing to perform any sort of quality assurance testing in their searches, resulting in 165 inadvertently produced documents.  The producing party claimed the 165 documents were privileged. 

 Magistrate Judge Grimm found the producing party waived their privileges.  The Court outlined the following failures of the Defendant leading to the waiver:

  • Failure to prove the keyword search was reasonable
  • Failure to identify keywords
  • Failure to show qualifications of person who came up with search terms
  • Failure to show quality assurance testing
  • Failure to explain what they did and how it was sufficient

 Factoring these opinions together we can see a list if “to do’s” to run defensible searches and protect any privilege information.  The first being those using litigation support software must be knowledgeable in how to effective use the solution. 

computerexpert I do not think courts will require an expert witness for statistical analysis of search terms in every case to show a search was defensible in traditional litigation support solutions for document review.  For example, no judge requires an expert to explain how a DVD player works or requires affidavits that a paralegal knows how to push “play” or “fast forward.”  Many of these litigation support products have existed for 20 years and courts accept lawyers know how to use them. 

 There would be higher standards for an expert witness in situations including collection of ESI off a hard drive or processing of ESI over those search terms, but that standard would be extreme for someone using common solutions such as CT Summation iBlaze, LexisNexis Concordance or West Case Notebook.  However, a party using a litigation support solution may need to have affidavits explaining what qualified persons with knowledge of a litigation support solution did for a privilege search.  It is possible an Attorney Client Privilege search might just be searching for the domain name of the law firm in “To” and “From” email.  In other cases it may be far more advanced.  What is “reasonable” and what a court will expect from a party will likely vary with the facts of each case. 

 “Quality assurance” could be running a test search and confirming the results match with known privileged documents.  For example, if you have known emails from between the lawyer and client, and your privilege search does not include those emails in the search results, you might have a problem with your search terms.

 Using litigation support solutions can reduce the time to review ESI, create production sets and privilege/redaction logs.  When it comes to litigation support technology, lawyers cannot treat their paralegals like a 14 year old son who is asked to hook up a Wii or program an XM Radio.  Understanding litigation support solutions is necessary for effective document review, reducing costs and demonstrating to a court your searches did in fact produce all responsive ESI.

Be Careful What You Ask For, Because Someone Just Might Produce It

There were three big “search term” cases in 2008 by Magistrate Judges John Facciola and Paul Grimm (both of whom also wear bow ties).[1] In United States v. O’Keefe, Judge Facciola wrote, “…for 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.

The following is not one of the “Big Three,” but illustrates the point lawyers need to be very careful what search terms they use and demand in discovery.


demand-lawyerIn Whitlow v. Martin, 2008 U.S. Dist. LEXIS 46111 ( C.D. Ill. June 12, 2008 ), the Plaintiff supplied search terms to the Defendant in a request for production. The Defendant in turned produced all the information that was found by the Plaintiff’s search terms.


The Plaintiff brought a motion to compel and claimed the production incomplete, because, “[t]he only search was one using the terms Plaintiffs supplied.


The Court denied the motion to compel, finding that the Plaintiffs failed to explain how a search with search terms provided by Plaintiffs’ counsel was insufficient.

Lawyers truly need to be careful what they request with search terms. This area is truly “where angels fear to tread” and attorneys will serve themselves well by seeking out experts to assist with determining search terms for collection, processing, meet & confers and requests for production.

[1] The “Big Three” include United States v. O’Keefe, No. 06-CR-249, 2008 WL 44972 (D.D.C. Feb. 18, 2008; Equity Analytics, LLC v Lundin, 248 F.R.D. 331 (D.D.C. 2008 ); Victor Stanley, Inc. v Creative Pipe, 2008 U.S. Dist. LEXIS 42025. There were many other search term cases that also warrant review.


The e-Discovery Chase: Strategies to Reduce Electronically Stored Information in Discovery Requests

All of our tools and toys that generate electronically stored information can frustrate law firms with high e-Discovery costs.  Lawyers who choose to have their discovery productions printed can have nearly 10,000 times more paper than 10 years ago.[1]

It is difficult to visualize how “big” ESI can be. WIRED magazine recently gave examples of data size most people can comprehend[2]:

1 Terabyte: a Hard Drive with 260, 000 songs

20 Terabytes: All the photos uploaded to Facebook each month

120 Terabytes: All the data from the Hubble Space Telescope

530 Terabytes: All the videos on Youtube

1 Petabyte: Data processed by Google’s servers every 72 minutes

12 Exabytes: All human produced information[3]

The data explosion can drive up e-Discovery costs for collection, processing and review to rival that of fielding an America’s Cup team.  And like the America’s Cup, if only billionaires can afford access to Federal Courts, then justice is not being served.

Danger of Costs from Overly Broad Requests for Production

Given the sheer volume of ESI, discovery requests need to be focused to avoid opening a Pandora’s Box of data.  For example, a third party request for production demanded the following:

The content of any and all electronic files, e-mail messages (with attachments), Instant Message communications and/or other communication created any time between August 20, 2001 to July 20, 2007 and maintained by Yahoo! related to account holder Jacqueline Hone’s subscription with Yahoo!, Yahoo! mail and/or Yahoo! Messenger.[4]

Even if the court had allowed production of this electronically stored information, how is it a victory to review SIX YEARS worth of email and instant messages?  The cost to collect the information, process it and review time would make the venture questionable.

Lawyers need to employ strategies to reduce the volume of ESI to control costs.  These strategies including requesting what you need, utilizing trusted vendors for targeted collection, and using technology for pre-processing analysis, effective processing and intelligent review.

The Discovery Request

There is no discovery production to review without first a discovery request.  Lawyers should avoid discovery requests such as, “All email to or from the Plaintiff from 1999 to 2008.”  Such requests are rarely granted when challenged, cut “due to the fact” and use “because” due to the fact they are overbroad, unduly burdensome, and often amount to a mass fishing expedition.[5] Moreover, getting 800,000 email messages from one individual would be a Pyrrhic victory considering review costs.

Lawyers should consider timeframes, form of production, specific individuals and other factors to create a narrowly tailored request. This may reduce motion practice and help control production and review costs.

A party responding to such an overly broad request should object and force the demanding party to focus their requests for “all email” or “all information related to the website.”  The petabyte explosion of ESI has not eliminated the discovery requirement that requests be narrowly tailored and reasonably calculated.

Targeted Collection

Information defensibly collected off hard drives should be focused on relevant or responsive electronically stored information.  If representing a contractor in a construction defect case, targeting collection to the housing project at issue will collect less information to review than a mass copying of hard drives.  However, care should be given not to under collect ESI, creating a risk for re-collection if something is missed.  Moreover, this should not undercut any preservation duties.  One might have to mirror image a hard drive to preserve all the data if the facts call for it, but what is collected for review should be focused.

Pre-Processing Analysis

Software tools have been developed for “non-linear” review for “pre-processing.”[6] What this translates into is looking at email messages before they are processed[7] or prepared for a review system[8].

“Pre-Processing” allows a look at email strings and a determination of what needs to be processed for linear review in a product like CT Summation iBlaze or Lexis Nexis Concordance.  This stage can eliminate email messages that are spam or newsletters based on their domains, such as “” or “” to reduce what is ultimately loaded into a review platform.

Effective Processing

Processing of electronically stored information is the extraction of metadata and full text in a format readable in a review platform.  The number of native files ESI processing software can process are in excess of 500 to 700 native file formats and growing.

Processing can be “brute force” and literally just included everything that was collected.  However, just because preservations duties may require the mirror image of a hard drive, that does not require an entire hard drive be processed for production.  Processing can be focused with key words, such as names of parties or witnesses, date ranges and other criteria to narrow the information for review.  Such “smart processing” can reduce costs by shortening review time by narrowing the production to a focused production.

Making the Most of Review: Not just a box of paper

Reviewing ESI is not the same war room experience as digging through boxes of paper.  Conference rooms do not need to be commandeered for months with contract attorneys plowing through fields of paper for the smoking gun document.  Technology can help reduce such backbreaking work.

Law firms can avoid the haphazard approach to review with assigning reviewers “review sets.”  A “review set” is saved database search, that can be a DOCID, date range and keyword(s).  The litigation team generally knows some of the basic terms and dates relevant to the lawsuit, and searching based on these terms for review helps focus review.  Reviews can also be set for specific individuals, document types or almost any term appearing in the database.

Review tools allow for issue coding (and also vary by product).  Coding for causes of action, specifically by elements of causes of action, allows for thoughtful review and case preparation.  A reviewer can also consider admissibility issues during the coding stage, identifying any problems before trial.

These are just a few strategies for review.  There are many others.  The main point is not to treat fully searchable electronically stored information as a box paper.  Searching and organizing based on key words can help focus review into what is relevant.

Don’t Drown in Petabytes: Search to Reduce ESI

All of our cool iPhones, laptops and Wii’s are engines of electronically stored information.  Technology has created an excessive amount of ESI in litigation, but technology can solve the problem in reducing ESI so law firms can focus on practicing law, not drowning in data.  

[1] John Bringardner, WIRED, “Winning the Lawsuit,” page 112, July 2008

[2] “The Petabyte Age,” WIRED, July 2008, pp 106-107


[4] Hone v. Presidente U.S.A., Inc., 2008 U.S. Dist. LEXIS 55722 ( N.D. Cal. July 21, 2008 )

[5] See, Hone v. Presidente U.S.A., Inc., 2008 U.S. Dist. LEXIS 55722 (N.D. Cal. July 21, 2008 ), Quinby v WESTLB AG 2006 WL 59521, 1 (S.D.N.Y) (Jan. 11, 2006), and Thompson v Jiffy Lube International, Inc, 2006 WL 1174040, 3 (May 1, 2006).

[6] Companies such as Metalincs both make products for this purpose.

[7] Processing products include IPRO’s eCapture, Needle Finder, LAW Pre-Discovery from Lexis or CT Summation Discovery Cracker, to name a few.

[8] Review platforms as hosted solutions such as Concordance FYI, CT Summation CaseVantage or desktop solutions such as CT Summation iBlaze or Concordance