<b><i>Legal Tech: Case Study</b></i> Getting Creative with the Relativity Platform: How to Save a Multibillion-Dollar Deal

At Reed Smith, we represent leading companies in everything from complex litigation and cross-border transactions to crucial regulatory matters, and we've been using Relativity since 2012 to help us tackle these projects and more.

Many know the Relativity platform as a tool that assists with litigation, specifically moving data across the Electronic Discovery Reference Model (EDRM), from legal hold, collection, and processing all the way through document review and production. However, its capabilities extend far beyond that. Over the years, we've found that Relativity is flexible and customizable enough for us to do almost anything. We encourage others to start thinking outside the box of traditional document review, too.

It's easy to preach “think outside the box.” But in the world of the linear EDRM, how does one actually put this concept into motion? The answer depends on the project, but there are some basic guidelines that we've found go a long way: 1) leverage the power of text analytics technology whenever possible; 2) customize all objects (not just those associated with documents); and 3) don't be afraid to get creative.

If you keep those in mind, you'll be off to a good start.

To show you what I mean, consider a project our team tackled last summer. Our client, a global financial company, was negotiating a multibillion-dollar sale of holdings. Before moving forward, the company needed to determine what kinds of restrictions and obligations each of their holdings was bound to on more than 10,000 contracts with various third parties.

In other words, we needed to determine which business units were buried in red tape and which would be the easiest to move quickly. Additionally, we only had one month to do it.

Unfortunately, our client, as is typical in most organizations, did not make it a practice to organize its contracts in a meaningful way. Instead, all of the different contract types ' those with real estate companies, software vendors, service providers, etc. ' were stored together, and addendums, attachments, expired contracts and renewals were intermixed, making it impossible to quickly determine what the stipulations were, which contracts were current, and which contracts were out of date.

To complicate matters further, this was a global corporation, so there were contracts in several different languages that would need to be handled by native speakers.

From this mess, we needed to determine which contracts belonged to which business units, what the stipulations to those contracts were, and which ones would require consent ' and what kind of consent ' so that we could obtain proper consent and ultimately move forward with the deal.

Where would you start?

Leveraging the Power of Analytics Technology

As a first step, we needed to determine what resources ' foreign-language reviewers and subject matter experts ' we would need to employ to complete the project. We started by analyzing all 10,000+ contracts in two ways: 1) by language so we could send the non-English contracts to foreign-language reviewers at our global offices; and 2) by theme or contract type (e.g., real estate, service agreement, etc.) to determine how many and what kinds of subject matter experts we would need to engage. Two features in Relativity Analytics assisted with this: language identification and clustering.

Language identification does exactly what you might guess: It examines the extracted text of each document to determine in what language or languages that document is written. It does this by analyzing three features of language: 1) character sets (Greek versus English, for example); 2) letters and the presence of absence of accent marks; and 3) the spelling of words (for instance, words that end in “-ing” are typically English).

With clustering, the text analytics engine again groups documents based on their content, but this time on conceptual similarity; in other words, what the contracts are about. To perform clustering, analytics identifies the most logical grouping of documents, organizes them into clusters, and labels each cluster with a naming convention that matches the conceptual content within, which made it easy for our team to see at a glance which ' and how many ' contracts were aligned with one another. For example, one cluster contained contracts between one of our client's holdings and a software vendor, while another contained real estate contracts for a particular building the company owned.

By using clustering and language identification, we were able to group contracts by language and topic and, within the hour, calculate how many foreign-language reviewers and subject matter experts we needed to employ to complete the project efficiently and within our timeline. Then, using Relativity, we could easily batch documents to the appropriate reviewers, who could begin reviewing immediately.

For this particular project, we sent the non-English contracts to reviewers in several of our global offices (Japan, China and Russia), and we batched the English contracts into two separate groups: real estate contracts, which would go to our real estate practice, and all other English contracts.

Getting Creative with Relativity Dynamic Objects

Organizing the contracts and mobilizing a review team across our offices was only half of the story. We also needed a way for reviewers to quickly and easily record multiple decisions on each contract ' decisions that go beyond the typical “responsive” and “not responsive” ones we use in litigation.

That's where the creativity and flexibility of the platform come into play. While analytics organized the contracts, we created custom questionnaires that our review team could use to record decisions about the contracts. The questionnaires included items such as:

• Would consent be necessary?
• What kind of consent?
• How many days does the vendor need to respond?
• Does consent need to be done in writing, or is verbal consent acceptable?

Creating the questionnaire was quite simple, taking only 15 minutes to build using Relativity Dynamic Objects. Every Relativity workspace consists of interconnected objects that securely store different types of information. For example, a document object might contain fields for coding decisions, custodian, or the company's name.

With dynamic objects, we can create objects to store the information we need, then link them to other objects in the workspace.

In this particular matter, we created various Question objects and linked them to the appropriate Document objects for the different document clusters (e.g., real estate contracts and contracts with particular software vendors). Each review team had access to different questionnaires with questions targeted to their assigned batch of contracts.

We even took it a step further and created a custom consent tracker, identifying the parties from whom we requested permission, when, if and how they responded, and which contracts they replied to (as some vendors were associated with multiple contracts), thereby creating a customized virtual rolodex of vendors and their associated consent.

Saving the Deal

Within just a few days of initially receiving the contracts from our client, our review teams across the globe were fast at work on the project. In the end, we successfully helped the company prioritize business units and close the deal within the one-month deadline.

This is just one example of how combining innovation with a flexible and customizable solution helps us exceed the expectations of our clients. By “thinking outside the box,” we didn't just accomplish a standard review, we were able to quickly slice and dice the data in different ways and determine which business units our client should sell first and which would need further consideration.

Sharri Wilner is the East Coast litigation technology services manager at Reed Smith. She works with both clients and attorneys to develop strategies and implement technology to manage issues regarding electronically stored information.