One of the things I do regularly is help clients get appropriate ESI orders in place for a matter.

What is an ESI order, you ask?  An ESI order is an agreement between the parties, and then signed off on by the judge, as to how the parties will handle ESI.  It’s like a Scheduling Order, but instead lays out how the parties will deal with ESI. The order has two main purposes:  1) to define a process that the parties agree on for identifying relevant data, and 2) to lay out how that data will be produced to the other side (i.e. format of production), how that data will be handled during the litigation and what the accompanying privilege log will look like.

Many federal district courts have sample ESI orders that the parties can use, like this one for the Northern District of California.  We’ve included all of the orders for the various district courts in the eDiscovery Assistant app, so if you have that handy look under the Rules for your jurisdiction and you’ll see them there.  I generally adapt the language of the federal orders for use in state courts, as most do not have proposed sample orders.  Once you do one, you’ll have a template that you can modify going forward.

Why You Need an ESI Order

Why bother with an ESI order?  Can’t you just do it as you go along?   If both sides have two competent eDiscovery lawyers who are communicating regularly and tracking the issues, then you may be able to skip an order out of the gate (although  you’ll want something in writing later that you can lean on in the event of non-compliance).  But if not, sitting down and addressing these issues at the outset of the litigation is the only way you can make good, economical decisions about how to conduct eDiscovery for a case.

The order is a must have in large commercial, mass tort, or class action matters, but in smaller cases it’s a means to an end.  And the “end” is to force the parties to sit down and think about discovery at the outset of the case.   Thinking about discovery at the outset of the case, or when you are planning to file one, saves LOADS OF TIME AND MONEY.

What to Include in an ESI Order

There are generally five main substantive issues that should be included in an ESI order:  the scope of preservation, the process for culling information from initial collection (i.e. search or TAR), the format of production (i.e. native vs. TIFF), how a phased approach will work (if applicable, but it’s ALWAYS the best idea unless your case is very small), and how parties will handle privileged documents.  Each of these issues is a separate chapter in a book, so we’ll just hit the highlights here.  A sixth piece to always include is that the parties will COOPERATE and continue to meet and confer on all issues.

Scope of Preservation

For clients with lots of data, identifying an appropriate scope of preservation saves considerable time and money because it clearly defines the party’s obligations and allows them to keep purging data outside of the scope under normal retention policies without risk.  There are six aspects to preservation to think about:

• The date range for keeping data, e.g.  June 15, 2010 to January 1, 2012.
• The number or names of custodians that will be preserved for each side.  If you use names, or exchange them with the other side, include job titles and a description of that person’s involvement.  Otherwise, a list of names is useless to the other side.  Avoid picking arbitrarily — instead if the parties are cooperating, exchange names of who you believe is the most important and agree to sample data and use analytics to see who the other custodians are that need to be captured/produced.  With eDiscovery, sampling and an iterative process of gathering information is the best, most cost-effective and informative approach to take.  Remember to include third parties in this list as well if that’s applicable.
• The types of ESI that will be preserved, searched and produced.  This list might include email, instant messaging, text messages, databases, cloud based services, medical records, social media, voicemail, etc.
• The sources of ESI that will be preserved, searched and produced.  This list might include email archives, email servers, mobile devices, specific databases (e.g. SAP, Oracle, CRM software), specific social media accounts (make sure to include the list of sites WITH account information, i.e. Kelly Twigger’s Facebook page, ESI Attorneys’ Twitter account, etc.), etc.
• What information, if any, will NOT be searched and that the parties agree is inaccessible (the burden/cost of searching it outweighs it’s likely value).  Make sure your jurisdiction’s rules allow for this distinction — some states, like California, have done away with it.
• An agreement that if new sources or types of ESI are discovered during litigation, the discovering party will immediately disclose that to opposing counsel and the parties will meet and confer on how to address preservation as it relates to the new source of information.

Search or Process for Culling Initial Data Collection

No matter how refined your date range and custodian set, you’ll always preserve and collect more data than you want to review for production (and conversely, you’ll receive more than you want to review).  So, you need a plan for how to get your 100 GB of data (just making that number up, it can be much higher or lower) down to the 5 GB or less that you really need to review and then produce.

Search terms are the most often used solution — and also one fraught with difficulty.  Search terms are difficult because lawyers think they can decide on them without either 1) getting input from the custodians who created or received the ESI or 2) looking at a sample of the data to see what the best search terms are.   #2 is my go-to strategy, and then I use #1 (talking to custodians) to further refine what I might use for search terms once I see what the data is telling me. The complicating factor is that you can’t use a review platform to do that — you need a tool that helps you out earlier.  I like [focus-color]ntella[/focus_color] for that step, and Nuix is another good ECA tool (but more expensive).

Technology Assisted Review (TAR) has been a much hyped technology over the past 18 months in eDiscovery, and it certainly has it’s place.  Right now, the technology is still new and novel and is very expensive, so it’s been relegated to use with very, very large cases with millions or billions of documents, or for firms that have bought a solution and brought it in house for use with all appropriate cases.  TAR works by having a knowledgeable group of attorneys (just 1-3) code a seed set of documents to teach the computer what’s responsive and what is not.  Once the computer learns, it’s off to the races on it’s own, and the review cost savings are substantial.  There are also multiple reports showing the results of TAR are much better than humans reviewing documents — meaning that the computer does a better job percentage wise at finding the relevant documents.

Regardless of what path you choose, your order needs to set out the process that the parties agree to follow.

Format for Production of Data

Format for production means what types of files you want to receive for productions.

The default for many lawyers receiving ESI seems to be asking for TIFF images with extracted text, or pdf files.  Both make my insides want to explode.  Why?  Well, and they’ll be more on this soon on our blog, dealing with native data considerably less expensive for both parties.  You need to think about two things when deciding on format of production:  how much data you’ll be dealing with (whether producing or receiving) and how you want to handle the data.  How you want to handle the data means how you want to be able to use tools to analyze it, or whether you just want to toss it all in a review platform and conduct a linear (document by document) review.  You need to know what format the review platform takes.  If you are working with a vendor or lit support group, ask.  If not, if you’re using a self-serve platform like Logikcull, or Lexbe, check with them.  They’ll guide you through choosing a format.

The volume of the data is important because if it’s a higher volume than what will allow you to do a liner review (again, reviewing doc by doc with a review team), then you need to use some more sophisticated tools to help you learn what you have and how to organize and group those docs for review.  Those analytical tools may require a different format for loading.

TIP:  The FRCP and many state rules have a specific rule (see e.g. FRCP 34(b)(1)(C)) allowing the receiving party to request the form of production in it’s RFP’s.  Make sure  you do it there AND include it in your order.  If you don’t, the other side gets to produce ESI in a reasonably useable form, and courts vary on what forms are reasonable.  You may lose your opportunity to save money and work with the data how you want to.

Phased Approach

As I said at the outset, and many, many judges have advocated, sampling and using a phased approach in eDiscovery is the most cost-effective what to go to get what you need.  Why?  Because there are always just a few people that were at the heart of the matter, and they were either the sender, recipient or cc’d on most all correspondence because that’s what people do.  So use those folks and decide on who the first custodians will be, what the search plan is and then say we’ll do this in phase 1, and by x date we’ll agree and advise the other side on what we’ll do in phase 2.  Set it out in your order so that neither party can argue that their obligations are met by simply providing phase 1 info.

There’s no magic to HOW you do your phased approach, so get creative based on what you have.  The judge wants to know the parties are working together and being reasonable, so chances are good you won’t have any roadblocks from the court if you are.

Withholding Privileged Data

We all know how expensive and laborious the traditional privilege log is, especially in a corporate matter, mass tort or class action case.  Review platforms and other software today help handle this nightmare by creating a log based on metadata fields.  That can be very useful, but you need to make sure that you show the other side what you are contemplating — give them a dummy log so they know what they can expect in an automated privilege log and then devise a process for how to work through the issues.

For example, if messages from a list of outside counsel will automatically be marked privilege, set that out in the order.  Then come up with terms to use for the automated log — but make expectations clear.  I’ve had cases where the defense wanted to shift their obligation for creating a privilege log completely to the plaintiffs by saying we won’t tell you the basis for privilege, but if you think it’s an issue, ask.  That won’t get the job done.

The order needs to include the various steps that you’ll take to provide a privilege log and that meets the obligations under the controlling law.

Summing it up  . . .

The chief complaints about eDiscovery are the volumes of information and the costs associated with dealing with it all.  The solution isn’t that difficult:  sit down and think about discovery for your case early on, cooperate and work with the other side to save you both time and money, and get an ESI order in place that provides a framework for the case.  This process also helps you focus your thoughts early on what information is really important AND it helps identify third parties who might need to be contacted for ESI.

Follow the list above and you’ll be miles ahead of where you usually are in your cases when you don’t think about what  you really need in discovery until 6-8 months after it’s been filed.