This site uses cookies to provide you with a more responsive and personalized service. By using this site you agree to our use of cookies. Please read our cookie notice for more information on the cookies we use and how to delete or block them.

Bookmark Email Print this page

Predictive coding streamlines document review by reducing the number of materials necessary for review

Learn how predictive coding led to a 60% reduction in the materials necessary for review, saving the company both time and money



A global entertainment company needed to comply with a Foreign Corrupt Practices Act investigation.  But using keyword search terms that counsel negotiated with the Department of Justice resulted in more than 140,000 documents for review.

The Deloitte Discovery project team collaborated with the client's eDiscovery team to educate them about Dynamic Review, Deloitte's proprietary analytics process that can increase the efficiency of document review by prioritizing the documents most likely to be responsive. After a thorough vetting by counsel of both process and technology during the first set of review documents, Dynamic Review was applied to a second set - the result: A sixty percent (60%) reduction in the materials necessary for review.

The Challenge

The client's counsel was familiar with the concept of “predictive coding”, but had reservations about using the process on the project. After all, the project was rather small and counsel expressed hesitation about the value of using technology to assist in the review process. Therefore, the client's counsel needed to be convinced on the effectiveness of the technology and process, as well as how it would add value to the end results. 

How We Helped

Deloitte applied Dynamic Review in parallel to the documents already undergoing a linear review. As the review continued, Deloitte utilized documents coded during the linear review to create an analytical model and then scored the remaining review population for the likelihood of responsiveness. Careful analysis of the predictive model against the manual coding indicated a strong correlation. Based on these results, counsel approved prioritizing the review of the remaining documents utilizing Dynamic Review's scoring. As the review moved from the highest to the lowest scored documents, counsel monitored the steady drop in document responsiveness rates. Finally, upon completion of the review, the responsiveness rate of the group of documents indicated by Deloitte as most likely to be responsive was 73.17%, compared to the 10% overall responsiveness rate of the linear review. Armed with this information and the positive experience with the Dynamic Review process, counsel agreed to leverage the process on the next set of documents slated for linear review.

The next set of documents resulting from a keyword search were scored and reviewed using the same approach - moving from the highest to lowest scores. As the team moved into the lower scored documents, the same drop in responsiveness rate was observed. Once the top eight of 10 groups of scores, or “bands”, had been reviewed, counsel performed a statistical analysis of the remaining documents. Based on that analysis, counsel determined that no review of the two remaining bands was required due to the low responsiveness rate. Fully sixty percent (60%) of the documents that would otherwise have been reviewed in a linear fashion were excluded from review based on prioritization using Deloitte's Dynamic Review.

Four key factors contributed to the success of the project: 

  • Collaborative consultation with counsel's eDiscovery team provided transparency and confidence in both Deloitte's process and analytical modeling
  • Targeted review based upon Dynamic Review clearly demonstrated an improvement over linear review
  • Statistical analysis validated the choice to not review sixty percent (60%) of the universe that hit search terms
  • Counsel's eDiscovery team was able to monitor the entire process and was assured that Deloitte's processes were efficient and effective.  


The client and its counsel were able to achieve the objectives of understanding how predictive coding could aid their review workflow, prioritizing documents based on responsiveness, and ultimately driving down costs by avoiding unnecessary review of documents that were statistically not responsive.


Steve Shebest
Deloitte Transactions and Business Analytics LLP
+1 202 220 2631

Derrick C. Lee
Deloitte Transactions and Business Analytics LLP
+1 212 937 8231

As used in this document, “Deloitte” means Deloitte Transactions and Business Analytics LLP, an affiliate of Deloitte Financial Advisory Services LLP. Deloitte Transactions and Business Analytics LLP is not a certified public accounting firm. Please see for a detailed description of the legal structure of Deloitte LLP and its subsidiaries. Certain services may not be available to attest clients under the rules and regulations of public accounting.


Related links

Share this page

Email this Send to LinkedIn Send to Facebook Tweet this More sharing options

Stay connected