希望访问中文页面? 请点此(简体中文版)  

Posted on: February 09, 2015

in Blog

4 Methods for Applying Analytics to Opposing Party Productions

Learn how to apply analytics to opposing productions and which metadata fields you should request in your stipulation to ensure the best analytics results.


Traditional linear review of documents for production is beginning to be replaced by review strategies predicated upon the use of analytics. Specifically, the use of email threading and near-duplication is now as common at D4 as traditional linear review. While not as common, other analytics tools, like Relativity’s clusters and categorization, are also finding a footing in document review. And, of course, there is growth in the application of predictive coding workflows. We have seen growth in the use of advanced analytics to enhance and expedite document review for large cases, and to significantly reduce—or completely eliminate—the review process in small cases. However, we have yet to see the same level of growth in the application of these tools to opposing party productions.

This white paper shows how one legal team applied predictive coding to avoid reviewing 70% of opposing counsel’s production and saved $1.4 million.

Here are a few methods for applying analytics to opposing productions and what metadata fields you should request in your stipulation to ensure the best analytics results.

4 Methods for Applying Analytics to Opposing Party Productions

There are several analytics methods that lend themselves to opposing party productions. Those methods include near-duplicate comparison and review, email threading inclusive review, cluster or thematic review, and categorized review. In each of these scenarios, we leverage coding, applied during your document review, to documents produced by the opposition. Doing so, allows you to quickly and efficiently go through the new documents to classify the production into two sets: those you already know about, and those that are unique to the opposing production.

Here is a breakdown of these four methods:

1. Near-Duplicate Comparison

All that is required for this strategy is document text. To execute, have your hosting provider run near-duplication across both your documents and the opposition’s. The vendor can then segregate the documents into three distinct sets: documents that are unique to your collection; documents that are unique to the opposing production; and those that are mixed. Once the separate populations are identified, you can prioritize the review to look at the unique opposition documents first, and then review the mixed.

2. Email Threading Inclusive Review

Depending on the tool used to create email threads, you may need some extra metadata fields to optimize the results. Those fields are: Family identifier (Attach Begin); Parent ID; To; From; CC; BCC; Subject; Date Sent; Conversation Index; and Attachment Names. Once the email threading has been run, your vendor can then segregate the documents into three distinct sets: documents that are unique to your collection; documents that are unique to the opposing production; and those that are mixed. Once the separate populations are identified, you can prioritize the review to look at the unique opposition emails first, and then review the mixed emails. Finally, you can leverage email threading so that you review only the inclusive emails. This means reviewing only the last email in any email conversation, reducing the number of redundant emails to review.

3. Cluster or Thematic Review

All that is required for this strategy is document text. Clustering or themes is an automated way for an analytics engine to group documents together based on conceptual theme. Once the clusters have been created, you can prioritize the production review by looking at the opposition’s documents that reside within the clusters containing the highest number of responsive documents from your internal document review.

4. Categorization Review

All that is required for this strategy is document text. Categorization is a way for the analytics engine to group documents together by pre-determined issues based on document examples. So, from your document review, you can identify up to five distinct issues and provide 5-20 example documents for each issue. The system will then categorize the opposing production based on those issues. This method allows you to find the important documents in the opposing production and focus your review on those documents first.

With these four methods in hand, it is easy to see how analytics can be used, not only as a means of organizing, enhancing, and expediting your document review for production, but also as a way to quickly comb through the opposing productions and locate the important documents within.

Metadata Fields to Request in Your Production Stipulation

To gain the most from analytics you’ll want to make sure that in your production stipulation, you are asking for the needed metadata fields. Those fields are: Family identifier (Attach Begin); Parent ID; To; From; CC; BCC; Subject; Date Sent; Conversation Index; and Attachment Names.

Whether opposing is dumping one million documents or just a few thousand, these methods will save time and money on review.

Discover More:

Discover More Categories

D4 Weekly eDiscovery Outlook

Power your eDiscovery intellect with our weekly newsletter.