The Keys to Analytics: Supercharge Your Document Review: Analytics and TAR are highly beneficial, but there’s still work to be done

Wednesday, September 28, 2016 - 15:24
Levin, Papantonio, Thomas, Mitchell, Rafferty & Proctor, P.A.
Louis Martin
Page Poerschke

Louis Martin

As technology assisted review (TAR) and analytics become increasingly commonplace in the discovery process, it is more important than ever for counsel to understand the benefits, drawbacks and planning required. Here, two well-known e-discovery experts, Louis Martin of iDiscovery Solutions and Page A. Poerschke of Levin Papantonio, discuss the current landscape and what may be coming down the pike. Their remarks have been edited for length and style.

MCC: Let’s start with some basics. Please describe analytics and how it differs from technology assisted review (TAR).

Martin: While TAR is popular and you hear the most about it, it is a solely a specific tool. Analytics is a broader tool kit or package that encompasses a suite of technologies including TAR, email threading, textually duplicate document detection and a few others. We use these to locate specific items and assist with the document review process.

Poerschke: TAR is useful for finding responsive and nonresponsive documents. We also often use other types of analytics Louis mentioned, like duplicate detection, to help eliminate the review of a lot of irrelevant documents, and email threading, which can help you focus on particular email chains.

MCC: How is today’s innovative technology changing the playing field in terms of TAR?

Poerschke: Linear review is really a thing of the past. It’s inefficient and inaccurate. Technology and analytics, such as concept searches and email threading, can level the playing field and allow smaller and midsized firms, which once couldn’t handle the vast number of documents produced in today’s litigation, tackle this volume of material. The cost of technology has come down, so the solutions are more affordable. To be relevant in the marketplace, firms are going to have to use these types of tools.

Martin: In a world of ever-growing data, it’s hard to reduce the volume of documents that must be reviewed, but analytics makes a big impact. If you can reduce the size of your population, it helps move the production forward and definitely lowers legal costs. You can better organize your documents for review, as well as find and highlight the documents you want to see first. Being able to quickly make such determinations is vitally important, allowing you to efficiently move in the direction that you need to.

MCC: Are these technologies available for use off the shelf without customization?

Martin: Without the proper experts, who have myriad experiences with these tools, it would be very difficult. If you don’t know what you’re doing, you can make a mistake that causes major problems with the production and costs significant time and money. An error can be magnified across an entire population, or an entire case, because of the way that analytics works. It is best to have someone you can rely on to help you set up your strategy, respond to your opposition, get your analytics going, and provide overall assistance throughout the process.

Poerschke: It’s important to understand that analytics and TAR don’t replace attorney review. You still need someone with subject matter knowledge to train the predictive coding engine. What analytics can do is substantially accelerate the document review process. The technologies themselves are becoming more and more attorney friendly. If you can find a vendor that really understands your needs and can work with you to streamline the use of technology, analytics – which does require some amount of customization – can be quite useful.

MCC: Going back to analytics, what are the greatest benefits?

Martin: Reduced time and cost. It also allows you to be more investigative, thorough, direct and precise. You find the documents you’re looking for faster, and it allows you to narrow your search parameters or an entire set of documents quickly. You will be able to collect less broadly, and review less broadly, than you would have without analytics.

Poerschke: Analytics allows you to maximize your resources. You can find the needle in the haystack, or the smoking gun–type document quickly, so that you can get back to focusing on practicing law. It also helps to minimize inconsistencies in the review process. Often, two people might code the same document, or a nearly identical document, in different ways. Analytics removes that variable, making both production and review more consistent.

MCC: What are the biggest pitfalls to using analytics?

Martin: There are still a lot of questions about analytics – how to use it, how to set it up, the best tools and technologies to use in your tool kit. You need to plan ahead and make sure you have people to help build a good strategy. That person or team needs to be knowledgeable about TAR, email threading and other analytic tools so they can speak to different pieces of the puzzle – what you’re going to turn over, how you’re going handle the entire process. If not, you could see some issues along the way – including the possibility that the opposition does not agree to your plan. They really need to be on board with the steps.

Poerschke: Particularly with regard to TAR, it’s only going to be as good as your subject matter expert. If you take the time to set it up correctly on the front end and review the sample sets or “seed” sets to train the technology appropriately and make sure that it’s running as expected, you should get very good results that will save enormous amounts of time and money. If you don’t put in the work on the front end, the problems in your sample set are going to be magnified exponentially throughout the document review process.

MCC: What are common mistakes made in running analytics?

Martin: Not paying attention to what is in your document set and just thinking you can push the big easy button and everything’s going to come out right on the other side. You need to be very clear and understand what it is you’re looking at, making sure you have everything in place, thinking ahead to what you’re going to be doing with your actual document set when you run analytics. For TAR, are there spreadsheets in the set and do they have heavy numbers involved? Sequences of numbers don’t mean anything to a TAR workflow. For email threading, check to make sure you have the entire thread so you can put them together. The most common mistakes are not paying attention and not checking and making sure that you have everything. To use analytics, you need to make sure that you have clear extracted text and other metadata. An analytics professional can help with all of that and more.

Poerschke: Another common mistake is poor communication with vendors. It’s really important that the vendor helping you with these technology solutions has a clear picture of what you want to accomplish. If the vendor understands what you’re looking for, they can supply solutions that attorneys might not have considered. If they don’t understand what you’re looking for, the outcome might not be what you want. You need a vendor who understands what you’re looking for and can convert that into tech talk and get back results that you want. You need a team you can trust.

MCC: What are two or three key takeaways for counsel surrounding analytics and TAR?

Poerschke: Planning, planning, planning. I can’t stress that enough. Having good communication and making sure you are fully explaining your needs. You need a good subject matter expert to train predictive coding. If you do those things, the results will substantially reduce your costs and make document review much more pleasurable. It won’t be nearly the expense and time burden that it’s been in the past.

Martin: I agree that planning and communication are vital. Counsel should always be communicative and open about what it is they want to accomplish. You should always begin with the end in mind. You want to talk early and often about what it is you want to get out of your analytics process or your TAR process. Talk to your subject matter experts, your trusted advisors. Also talk to opposing counsel, making sure they are on the same page as you are. Great communication with your technical team will enable them to provide the things you want so that you can succeed.

Poerschke: It’s also important to understand that these technologies are not mutually exclusive. You can use many different types of analytics tools in conjunction with one another. You don’t have to think of it as “one and done.” You might use four or five different types of tools in one document production.

Martin: Another takeaway is that while TAR can make things easier – and it does most of the time – you need to have a process in place and you need to plan for that.

MCC:  From your position on the front lines of the discovery process, have you seen any changes in the wake of the Federal Rules of Civil Procedure (FRCP) amendments that went into effect late last year?

Poerschke: The biggest changes I’ve seen, and I’ve been doing this for about 20 years, are the volume of documents produced and the frequency of electronic productions. Electronic discovery is now pretty much a given in every case. Everyone, including smaller and midsized firms, will have to understand and use these types of technologies, because without them today’s litigation is simply not doable. It’s something that everyone’s going to have to embrace.

Martin: I echo Page’s comments. We’ve seen analytics improve steadily over the past few years. You hear people talking about continuous active learning, sentiment analysis, semantics, disambiguation, and different types of things that can be done using analytics versus the standard stuff. In the next few years, that’s going to be as commonplace as TAR is now – and as keyword searching was a few years ago. It’s also important to understand that while analytics is extremely helpful and beneficial, it’s not an easy button. You need to plan for it and take it into account when budgeting a project. The best approach is to work with someone you know and trust, and make sure that you have an action plan – one that you can incorporate and use in a repeatable process across different cases. If you are dealing with a lot of litigation, having that in place allows you to move documents through the process as quickly as possible. Analytics is definitely a piece of that.

Poerschke: One of the top benefits of the technology is the defensibility and consistency of the results. When you have to go back into court and explain how you got your production set, having a planned workflow and process takes out a lot the human error component and can be very helpful in showing how you got to your results and the accuracy of those results.

Louis Martin is a Sr. Managing Consultant at iDiscovery Solutions. He can be reached at lmartin@idiscoverysolutions.comPage A. Poerschke is Of Counsel in the Securities and Business Litigation practice group of Levin Papantonio. She can be reached at