The benefits realized with Predictive Coding and Technology Assisted Review are well documented. Most of the focus has been on applying the technology for outbound document review. What about applying predictive analytics to incoming production from the opposition? What would be different in the workflow? What are the benefits? What has experience taught us?
Finding What’s “Important” vs “Relevant”
Consider this. You’re involved in a matter where both sides have agreed to a broad set of search terms as a means to limit the scope of production. You’ve applied search terms to your set of documents, reviewed for privileged (and perhaps some level for relevance ) and made your production.
There is a possibility your opponent did the same. The production from opposing counsel could be, in fact, a “data dump,” and contains a large percentage of false positive documents. (False positives are documents where at least one of the key terms was present, but the document has absolutely no relevance to the matter.)
Predictive Analytics: Thinking Outside of the Box
The typical use-case for predictive analytics is to implement it into your litigation review workflow, which determines a prioritization strategy for reviewing the relevant and potentially relevant documents from discovery. Using predictive analytics for an incoming production is different – it is all about identifying the documents that are important to support your side of the case, as quickly as possible.
Real Life Story
Recently, D4 was involved in a IP case where the opposing counsel produced in excess of 800,000 documents to our customer’s in-house counsel. The collection produced was based on keyword searches that both parties had agreed upon, but our customer suspected the production wasn’t reviewed by opposing for relevancy. With depositions starting in only a few days, our customers counsel needed to find out which documents from the opposition were most important to support our side of the suit, and fast!
Upon receipt, our customer used Predictive Coding to prioritize the production. It took the Subject Matter Expert (SME) four days (33 man hours) to train the system to the point it reached stability (had learned all it could from the human). The system then ranked the entire production of 800,000 documents in priority order, according to what it learned from the SME. The results showed:
• 30% of the production with the highest scores contained more than half of the most important material to support the case and would be set for “eyes on review”;
• 30% of the production with the lowest scores contained very little (to no) important material and would not be reviewed;
• It was decided the remaining 40% of the production in the middle would not be reviewed unless the material in the top 30% containing over half of the important documents to support their case wasn’t sufficient.
Results from Using Predictive Analytics on Opposing Production
In the end, receiving counsel was able to avoid unnecessarily reviewing approximately 70% (560K documents) of what opposing counsel had submitted, and with very little risk. Keep in mind, this was not a “relevance” review for production, it an “importance” review against opposing counsel’s production. Sufficient documents to support the case were found in the top 30%, with most of that in the top 10%. The system provided this important set of material to counsel in less than a week! They avoided approximately $1.4M in review costs, focused their attention on the most important documents earlier in the process, and most importantly, won the case. This customer now uses Predictive Coding as a standard practice for incoming productions.
What’s the Risk?
For inbound production, you’re not worried about privileged decisions or reviewing for relevancy. For inbound production, the only question you should be asking yourself is, “what did they produce that will help my side of the case.” Simple.
So, the next time you anticipate your opposing party is likely to provide a large production (over 50,000 documents is a good metric), and since there is such little risk involved, use predictive analytics to prioritize upon receipt. You (and your client) will be glad you did.
Are you anticipating a large production from opposing counsel? Don’t be fooled into thinking your opposing counsel will make your job easier by checking for relevancy before sending you their production. Get strategic tips now.
Predictive Coding in Practice: Four Success Stories [WHITEPAPER]
Predictive Coding/Technology Assisted Review Resource Center
Four Examples of Predictive Coding Success | LawTechnologyNews.com
Technology Assisted Review: What You Need to Know [INFOGRAPHIC]
Technology Assisted Review – What Does TAR Really Mean? [WHITEPAPER]
Tags: cost-effective, discovery phase, document production, document review, e-discovery, ediscovery, electronic discovery, Predictive Coding, Proportionality, review tools, Technology Assisted Review, Tom Groom