Form of Production Battles: TIFF vs MSG

Pick your battles is a truism in form of production disputes.


In Feist v. Paxfire, Inc., the Plaintiff alleged the Defendant made multiple discovery abuses, from producing email in .msg format and not as TIFF, to making an “intentionally burdensome production.” Feist v. Paxfire, Inc., 2015 U.S. Dist. LEXIS 145024 (S.D.N.Y. Oct. 26, 2015). These arguments failed.

The Plaintiffs requested ESI to be produced as TIFFs and argued that the Defendant did not properly object to the form of production. Feist, at *2. The Court held that the Plaintiff’s argument was meritless, because 1) there were email communicates between counsel regarding the form of production for email in .pst format; 2) the Court downplayed the difference between .pst and .msg format; and 3) that “msg format contains more metadata than TIFF format.” Feist, at *4. The Court further stated that there was “no evidence” that the .msg form imposed “considerable costs” for the Plaintiffs. Id.

It is worth noting that TIFFs are static images that contain no metadata at all. Any metadata would need to be produced in a load file. While it is wise to always review in a review application, I personally would prefer reviewing native files in a near-native view in a hosted application.

The dispute further included the Defendant’s 2012 production being shared to the Plaintiff in a Dropbox folder. An attorney for the Plaintiff deleted some of the information on Dropbox, resulting in the Defendants re-producing the deleted ESI with a supplemental production. Feist, at *5.

The Court held the Plaintiff responsible for the production of any duplicative ESI, citing an email from Plaintiff’s counsel stating, “that Plaintiff’s counsel was attempting to restore files unsuccessfully, and that she understood ‘you are sending us a hard drive with the materials, so we don’t need to worry about drop box [sic]…’” Feist, at *5-6.

Judge Ronald Ellis ended the opinion with these final words to the litigants: On balance, both Parties have caused unnecessary delay in discovery and have exhibited a lack of communication regarding document production. Feist, at *6.

Bow Tie Thoughts

Attorneys often get lost in the “fog of war” in eDiscovery disputes. There are times when bringing a motion to compel is the absolute right answer; there are times when it is the absolute wrong course of action.

Is it really worth fighting over producing ESI with redactions as TIFFS instead of PDFs? Most review applications give options for either format. If the requesting party wants one over the other, the fight is just not cost effective.

History has many examples of miscommunications resulting in disaster. Lawyers who want to fight on every issue run the risk of ignoring what the opposing party is saying, creating a situation that can end in expensive motion practice. In the end, this results with money and time being lost with nothing gained.

Proportionality in Asymmetrical Discovery

The Plaintiffs and Defendants in a SEC case highlight the importance of proportionality between asymmetrical parties. In such cases, one side has all of the electronically stored information for discovery requests; the other side does all the requesting. However, the smaller party can have an extremely high burden reviewing any produced ESI, especially if searchable features have been removed.

Magistrate Judge Leda Dunn Wettre in City of Sterling Heights Gen. Emples. Ret. Sys. v. Prudential Fin., (an opinion not for publication) did a great job balancing the proportionality interests between a motion to add additional search terms and custodians to the dispute.

Balancing Custodians

The Plaintiffs sough to add between 22 to 45 additional custodians for the Defendants to add to their discovery search. The Requesting Party made a strong argument for the additional custodians, including a chart of the custodians with the factual basis for expanding the scope of discovery. The parties had already agreed to 66 custodians. City of Sterling Heights Gen. Emples. Ret. Sys. v. Prudential Fin., 2015 U.S. Dist. LEXIS 110712, *4-5; 8.

The Court denied adding the majority of the additional custodians, explaining, Neither plaintiffs nor the Court can know with certainty, of course, whether searches of the additional custodians’ ESI will yield unique, noncumulative documents.” Sterling Heights, at *9. The Court further explained that it was satisfied that the vast majority of the custodians likely had duplicate information. Id.

Judge Wettre drove home her proportionality analysis with the following: the Court is cognizant of the sizeable costs to Prudential involved in the searching, review and production of information from each additional custodian. Although Prudential is a large corporation with substantial resources, the Court should not be – and is not-insensitive to these costs.” Id.

The Court found that the Plaintiffs had provided enough information that there was a “fair inference” that there could be more custodians with additional information. Id. Judge Wettre explained in terms of proportionality that:

Although the number of agreed custodians is already substantial, the resources and personnel at Prudential devoted to the Verus audit and related issues also seem to have been immense. Therefore, it is not surprising that more than 66 Prudential employees may have been heavily involved in the issues relating to this case and may thus have relevant, noncumulative information. Moreover, allowing plaintiffs a moderate number of additional custodians does not seem disproportionate to the size and scale of this action. The Court understands that there are electronic de-duping tools that may be utilized to limit defendants’ review and production of duplicative documents, reducing some of the burden on Prudential of producing information from additional custodians.

Sterling Heights, at *10.

The Court explained that permitting the plaintiffs to select an additional 10 custodians “would balance fairly plaintiffs’ rights to relevant discovery against the costs and burden to defendants of providing that discovery.” Sterling Heights, at *10-11.

Search Terms

The question of adding four new search terms was decided swiftly. The Defendants challenged adding more search terms, claiming they had already produced 1.5 million pages of discovery. Sterling Heights, at *11. The Plaintiffs responded that over half of the 1.5 million pages were “completely unusable redacted pdfs of Excel spreadsheets.” Id.

The Court held that the four search terms appeared designed to target relevant information. Moreover, the Plaintiff noted that if the Defendant had produced a hit count that showed the terms had an “egregiously large” number of “hits,” the Plaintiffs would have considered narrowing the terms. Id.

Judge Wettre allowed the additional terms and explained:

The Court does not have before it information on which it is persuaded that it should deny these four additional terms because they would produce an unduly large number of results likely to be irrelevant to this case. While the Court does recognize that defendant has agreed to a large number of search terms, that is not sufficient basis in and of itself to deny plaintiffs the four additional search terms they seek.

Sterling Heights, at *12.

Bow Tie Thoughts

If proportionality cases were rock concerts, this case has a few “gavel drop” moments. It is great to see a Judge who incorporated proportionality throughout the entire opinion.

Proportionality arguments should not be made out of thin air. The Plaintiffs made a noble effort providing a chart with each additional custodian explaining the factual basis for expanding the scope of discovery for each individual. While they only got 10 additional custodians, this was an excellent way to explain to a Court the value of adding custodians to decide the merits of a case.

The search term arguments for both parties could have been stronger if the proposed search terms were supported by affidavits from expert witnesses. To be fair, the arguments might have been, but the tone of the opinion sounds like the arguments did not include expert affidavits. There are many lawyers who think that because they can conduct case law research that “search terms” in discovery are the same thing. That is a dangerous assumption.

I strongly encourage lawyers to work with eDiscovery consultants to help identify the concepts to identify electronically stored information. Advanced analytics from clustering of similar files to email threading, to visual analytics, to predictive coding all can help identify responsive files. Lawyers should think beyond “search terms” to concepts in order to search discovery.

Finally, I feel greatly for the Plaintiffs who had to review non-searchable PDF’s of Excel files. There are ways to redact Excel files that would require agreements between the parties, but it can be done. This would keep Excel files in native format and avoid spreadsheets exploding into multiple page nightmares.

In closing, a hat tip to the Judge and both parties on their well argued positions and the opinion.


Always Explain How Requested Information Moves a Case Forward for Proportionality

A requesting party issued a third party subpoena on the ATF for email messages created between 2000 to 2009. The Court denied the production of the messages on the grounds the emails were not reasonably accessible because of being stored on back-up tapes. The moving party also failed to demonstrate good cause to order the production of the requested information. The Court also addressed the proportionality of the request under Federal Rule of Civil Procedure Rule Rule 26(b)(2)(C). Baranski v. United States, 2015 U.S. Dist. LEXIS 71584, *52-54 (E.D. Mo. June 3, 2015).


The Court began the proportionality analysis highlighting the emails were between six to eight years old and stored on back-up tapes for disaster recovery. Baranski, at *52.

The Court further noted that the production of the emails would “involve substantial costs and time and the active intervention of computer specialists,” which had been explained by expert affidavit. Id.

An added factor included that the emails would have to be searched by specific users and the requesting party had not identified any users to be searched. Id. This made the discovery request not sufficiently specific. Baranski, at *53.

The Court further drove home the proportionality analysis that the cost could not be justified for the requested information without some showing that the requested emails contained information to resolve the issues in the lawsuit. Id. As such, the Court denied the requesting party’s subpoena.

Bow Tie Thoughts

Proportionality is the balancing of two factors: 1) What must be done to make the data accessible (including cost, time on the project, what technology must be used) and 2) How is the information going to be beneficial to the case.

Cost is irrelevant if the information has not objective value to a lawsuit. A party seeking information that especially requires translation into a reasonably useable form should be ready to state why it supports the causes of action in a case, with specific custodians, time frames, and any other information to assist the Judge in making her ruling. Without knowing how the information will move the case forward, a Court will be hard pressed to order expensive data recovery procedures.

You Don’t Want Discovery Overdraft Charges

Discovery deadlines matter. Wells Fargo learned that the hard way with producing a relevant email 8 months after the close of discovery. Given the nature of the relevancy to the lawsuit, limited additional discovery was reopened.

Opps_iStock_Here is the short overview of the case: Plaintiff’s asked Wells Fargo if the other Defendants (now dismissed) were a legitimate business engaged in securities sales. The bank said yes and the Plaintiff transferred $80,000 to the Defendant’s bank account in order to purchase securities that would yield a return of $280,000.

The second transaction involved reinvesting $250,000 of the promised $280,000 to fund a $500,000 loan to renovate an office building, plus an additional $50,000 transaction fee. Gazian v. Wells Fargo Bank Na, 2015 U.S. Dist. LEXIS 69701, *2-3.

The Plaintiffs attempted to withdraw $30,000 and transfer the $250,000 to the Defendants. Wells Fargo informed the Plaintiffs that the dismissed Defendants accounts had been emptied and no money would be transferred to the Plaintiffs. Id.

The Plaintiff’s sued Wells Fargo on the theory the bank “knowingly or negligently made false representations to Plaintiffs” about the dismissed Defendants. Gazian, at *3.

Enter the late-produced email message:

From: BROWN, Patrick [Wells Fargo]

Sent: Tue 8/2/2011 5:36:42 PM

Re: HSBC Bank Guarantee Registration Number BH5843[.]

I have conducted a review on the signor, Craig Cason, for this account, [redacted]7443 — Increase Capital Investments LLC and found several items of concern . . .

This individual has been investigated by the SEC for securities violations and accusations of fraud.

The address on the account is a virtual office that can be rented for $50/month, used frequently by shell companies to give the appearance of legitimacy even though no actual business is conducted there.

The client has filed multiple bankruptcies and has several outstanding judgments (some in excess of $100K), which is not consistent with someone purporting to have $250mil in assets.

Please do not process the receipt of this security. We will be restricting the account and referring the matter to our Security Fraud group. Additionally, please DO NOT disclose this information to the client or the outcome of our review. Please advise the client that we cannot assist him with his request.

Gazian, at *4.

The Defendants claimed this messaged pertained to another securities deal and was not relevant. Gazian, at *5. The Plaintiff and the Court did not agree.

The Court reopened discovery so the Plaintiffs could conduct 5 additional depositions of no more than 20 hours of depo time; 10 additional requests for production; and 10 additional interrogatories. The Court went further to say that the additional depositions and written discovery could inquire into the preservation of emails sent to and from one of the Wells Fargo custodians. Gazian, at *7-8.

Bow Tie Thoughts

A lawsuit can hinge on one smoking gun email that was produced late. This sort of situation is one that needs to be avoided by litigants. The problem when this happens is often one of not having an effective information governance solution, or not issuing a litigation hold correctly, or not collecting ESI, or not knowing how to leverage early case assessment tools to find potentially relevant information. To put it mildly, disaster can happen for many reasons.

How can parties avoid these situations? Use sound technologies to manage data that can issue a litigation hold and preserve relevant ESI. This does require determining what is relevant, but these are problems that can be solved by knowing what actions to take and the tools to properly litigate a case.

Straightening Out the Form of Production

Magistrate Judge Stanley Boone had to straighten out a form of production dispute in a consumer protection case over curling irons. As the parties in this case learned, sometimes the form of production needs a detangler.

iStock_CurlingHairThe Plaintiffs requested ESI to be produced in native file format or TIFF with associated metadata. The Defendant produced ESI as PDFs. Wilson v. Conair Corp., 2015 U.S. Dist. LEXIS 57654, 4-5 (E.D. Cal. Apr. 30, 2015).

Judge Boone began his analysis with a summary of the form of production rules under Rule 34:

(E) Producing the Documents or Electronically Stored Information. Unless otherwise stipulated or ordered by the court, these procedures apply to producing documents or electronically stored information:

(i) A party must produce documents as they are kept in the usual course of business or must organize and label them to correspond to the categories in the request;

(ii) If a request does not specify a form for producing electronically stored information, a party must produce it in a form or forms in which it is ordinarily maintained or in a reasonably usable form or forms; and

(iii) A party need not produce the same electronically stored information in more than one form.

Wilson, at *6-7.

The Plaintiff’s request for ESI to be produced in native format was very standard. However, the data requested was produced from a proprietary third-party “STARS” database. Wilson, at *8. The Plaintiffs would not be able to access or review this data as it is ordinarily maintained because of its proprietary nature.

The Defendants produced the proprietary ESI as PDFs. The Plaintiff challenged this static image form of production in favor of TIFFs with metadata. Wilson, at *9. However, the Defendants were willing to produce future ESI as TIFFs. Id.

Excel files were also produced as PDF’s in order to redact information. Id. The Plaintiffs sought the Excel files to be produced in native file format. Id.

The Plaintiffs argued in favor of a TIFF production over PDF because the “format is more efficient, cost effective, and better suited for use inside a database application and it will require additional work to get the data produced in PDF format into a usable state.” Wilson, at *9-10.

The Plaintiffs further demanded the ESI from the STARS database be produced in Excel format. The Defendants ultimately agreed to this production format, but did not explain how the issue of redaction would be addressed in the opinion. Wilson, at *10.

The Court stated, the “Rules do not require a party to produce ESI in the form most helpful to the opposing party.” Wilson, at *10, citing U.S. ex rel. Carter v. Bridgepoint Educ., Inc.,     F.R.D.    , 2015 U.S. Dist. LEXIS 26424, 2015 WL 818032, at *15 (S.D. Cal. Feb. 20, 2015). As such, the Court ultimately held that 1) the STARS data could not be produced in its native format; 2) the Defendant would produced additional discovery in TIFF format; and 3) the Defendant would produce associated metadata for its prior production if it had not already done so. Wilson, at *11.

Bow Tie Thoughts

As any good hair stylist can tell you, a good product can help detangle knotted up hair. The same can be said for virtually any of the review applications on the market today. Most pride themselves on being able to review native file format, near-native, and static images such as TIFF and PDF.

I think it is odd to have a fight over which static image to produce. Both TIFF and PDF work well in today’s modern review applications. This was not always the case, as PDFs can be both a native file and static image in older review applications. It has been awhile since I have seen this be an issue in document review. That being said, if a requesting party asks for a specific static image format, I recommend honoring the request.

There are horror stories where producing parties have produced batches of native files as massive PDF’s that are several hundred, or thousand, of pages. In those situations, the requesting party has a very strong argument that the production was not in a reasonably useable form.